commit 39cfa35bfd309683d72395211b627b054e4f64f5 Author: Martin Fouilleul Date: Sun Aug 14 18:19:40 2022 +0200 initial commit diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..72d1402 --- /dev/null +++ b/.gitignore @@ -0,0 +1,3 @@ +.DS_Store +bin/* +*.metallib \ No newline at end of file diff --git a/build.sh b/build.sh new file mode 100755 index 0000000..a13d04e --- /dev/null +++ b/build.sh @@ -0,0 +1,81 @@ +#!/bin/bash + +DEBUG_FLAGS="-g -DDEBUG -DLOG_COMPILE_DEBUG" +#DEBUG_FLAGS="-O3" + +#-------------------------------------------------------------- +# set target +#-------------------------------------------------------------- + +target="$1" +if [ -z $target ] ; then + target='lib' +fi + +shaderFlagParam="$2" +#-------------------------------------------------------------- +# Detect OS and set environment variables accordingly +#-------------------------------------------------------------- +OS=$(uname -s) + +if [ $OS = "Darwin" ] ; then + #echo "Target '$target' for macOS" + CC=clang + CXX=clang++ + DYLIB_SUFFIX='dylib' + SYS_LIBS='' + FLAGS="-mmacos-version-min=10.15.4 -DMG_IMPLEMENTS_BACKEND_METAL -maes" + CFLAGS="-std=c11" + +elif [ $OS = "Linux" ] ; then + echo "Error: Linux is not supported yet" + exit -1 +else + echo "Error: Unsupported OS $OS" + exit -1 +fi + +#-------------------------------------------------------------- +# Set paths +#-------------------------------------------------------------- +BINDIR="./bin" +SRCDIR="./src" +RESDIR="./resources" +INCLUDES="-I$SRCDIR -I$SRCDIR/util -I$SRCDIR/platform" + +#-------------------------------------------------------------- +# Build +#-------------------------------------------------------------- + +if [ ! \( -e bin \) ] ; then + mkdir ./bin +fi + +if [ $target = 'lib' ] ; then + + # compile metal shader + xcrun -sdk macosx metal $shaderFlagParam -c -o $BINDIR/metal_shader.air $SRCDIR/metal_shader.metal + xcrun -sdk macosx metallib -o $RESDIR/metal_shader.metallib $BINDIR/metal_shader.air + + # compile milepost. We use one compilation unit for all C code, and one compilation + # unit for all ObjectiveC code + $CC $DEBUG_FLAGS -c -o $BINDIR/milepost_c.o $CFLAGS $FLAGS $INCLUDES $SRCDIR/milepost.c + $CC $DEBUG_FLAGS -c -o $BINDIR/milepost_objc.o $FLAGS $INCLUDES $SRCDIR/milepost.mm + + # build the static library + libtool -static -o $BINDIR/libmilepost.a $BINDIR/milepost_c.o $BINDIR/milepost_objc.o + +else + # additional targets + if [ $target = 'test' ] ; then + pushd examples/test_app + ./build.sh + popd + + elif [ $target = 'clean' ] ; then + rm -r ./bin + else + echo "unrecognized target $target" + exit -1 + fi +fi diff --git a/examples/test_app/build.sh b/examples/test_app/build.sh new file mode 100755 index 0000000..e62fb81 --- /dev/null +++ b/examples/test_app/build.sh @@ -0,0 +1,11 @@ +#!/bin/bash + +BINDIR=../../bin +RESDIR=../../resources +SRCDIR=../../src + +INCLUDES="-I$SRCDIR -I$SRCDIR/util -I$SRCDIR/platform -I$SRCDIR/app" +LIBS="-L$BINDIR -lmilepost -framework Carbon -framework Cocoa -framework Metal -framework QuartzCore" +FLAGS="-mmacos-version-min=10.15.4 -DDEBUG -DLOG_COMPILE_DEBUG" + +clang -g $FLAGS $LIBS $INCLUDES -o $BINDIR/test_app main.c diff --git a/examples/test_app/main.c b/examples/test_app/main.c new file mode 100644 index 0000000..bd4f976 --- /dev/null +++ b/examples/test_app/main.c @@ -0,0 +1,279 @@ +/************************************************************//** +* +* @file: main.cpp +* @author: Martin Fouilleul +* @date: 30/07/2022 +* @revision: +* +*****************************************************************/ +#include +#include"milepost.h" + +#define LOG_SUBSYSTEM "Main" + +int main() +{ + LogLevel(LOG_LEVEL_DEBUG); + + mp_init(); + + mp_rect rect = {.x = 100, .y = 100, .w = 800, .h = 600}; + mp_window window = mp_window_create(rect, "test", 0); + + mg_init(); + ui_init(); +/* + mp_rect frame = {0, 0, 800, 600}; + mp_view view = mp_view_create(window, frame); + mg_surface surface = mg_surface_create_for_view(view, MG_BACKEND_METAL); +/*/ + mg_surface surface = mg_surface_create_for_window(window, MG_BACKEND_METAL); +//*/ + mg_canvas canvas = mg_canvas_create(surface, (mp_rect){0, 0, 800, 600}); + mg_image image = mg_image_create_from_file(canvas, str8_lit("Top512.png"), true); + + u8 colors[64]; + for(int i=0; i<64; i+=4) + { + colors[i] = 255; + colors[i+1] = 0; + colors[i+2] = 0; + colors[i+3] = 255; + } + mg_image image3 = mg_image_create_from_rgba8(canvas, 4, 4, colors); + + mg_image image2 = mg_image_create_from_file(canvas, str8_lit("triceratops.png"), true); + + //NOTE(martin): create font + char* fontPath = 0; + mp_app_get_resource_path("../resources/Andale Mono.ttf", &fontPath); + FILE* fontFile = fopen(fontPath, "r"); + free(fontPath); + if(!fontFile) + { + LOG_ERROR("Could not load font file '%s'\n", fontPath); + return(-1); + } + unsigned char* fontData = 0; + fseek(fontFile, 0, SEEK_END); + u32 fontDataSize = ftell(fontFile); + rewind(fontFile); + fontData = (unsigned char*)malloc(fontDataSize); + fread(fontData, 1, fontDataSize, fontFile); + fclose(fontFile); + + unicode_range ranges[5] = {UNICODE_RANGE_BASIC_LATIN, + UNICODE_RANGE_C1_CONTROLS_AND_LATIN_1_SUPPLEMENT, + UNICODE_RANGE_LATIN_EXTENDED_A, + UNICODE_RANGE_LATIN_EXTENDED_B, + UNICODE_RANGE_SPECIALS}; + + mg_font font = mg_font_create_from_memory(fontDataSize, fontData, 5, ranges); + free(fontData); + + mp_window_bring_to_front_and_focus(window); + + mp_pump_events(-1); + while(!mp_should_quit()) + { + mp_event event = {}; + mp_pump_events(0); + while(mp_next_event(&event)) + { + switch(event.type) + { + case MP_EVENT_KEYBOARD_CHAR: + { + printf("entered char %s\n", event.character.sequence); + } break; + + case MP_EVENT_WINDOW_CLOSE: + { + mp_do_quit(); + } break; + + case MP_EVENT_WINDOW_RESIZE: + { + /* + mp_rect frame = {0, 0, event.frame.rect.w, event.frame.rect.h}; + mp_view_set_frame(view, frame); + */ + mg_canvas_viewport(canvas, (mp_rect){0, 0, event.frame.rect.w, event.frame.rect.h}); + + } break; + + case MP_EVENT_FRAME: + { + } break; + + default: + break; + } + } + mg_surface_prepare(surface); + + mg_set_color_rgba(canvas, 1, 1, 1, 1); + mg_clear(canvas); + + mg_move_to(canvas, 0, 0); + mg_line_to(canvas, 400, 300); + mg_set_width(canvas, 5); + mg_set_color_rgba(canvas, 1, 0, 0, 1); + mg_stroke(canvas); + + mg_image_draw(canvas, image, (mp_rect){400, 300, 200, 200}); + + mg_rounded_image_draw(canvas, image2, (mp_rect){200, 20, 200, 200}, 50); + + mg_set_color_rgba(canvas, 0, 1, 0, 0.5); + mg_rectangle_fill(canvas, 800, 800, 400, 200); + + mg_set_font(canvas, font); + mg_set_font_size(canvas, 32); + mg_move_to(canvas, 500, 500); + mg_text_outlines(canvas, str8_lit("Hello, world!")); + mg_set_color_rgba(canvas, 0, 0, 0, 1); + mg_fill(canvas); + + mp_rect windowRect = mp_window_get_content_rect(window); + + ui_style defaultStyle = {.borderSize = 2, + .borderColor = {0, 0, 1, 1}, + .textColor = {0, 0, 0, 1}, + .font = font, + .fontSize = 32}; + + ui_style buttonStyle[UI_STYLE_SELECTOR_COUNT]; + for(int i=0; i + +#define STB_TRUETYPE_IMPLEMENTATION +#include"stb_truetype.h" + +#include"lists.h" +#include"memory.h" +#include"macro_helpers.h" +#include"graphics_internal.h" + +#define LOG_SUBSYSTEM "Graphics" + +//--------------------------------------------------------------- +// graphics canvas structs +//--------------------------------------------------------------- + +typedef enum { MG_PATH_MOVE, + MG_PATH_LINE, + MG_PATH_QUADRATIC, + MG_PATH_CUBIC } mg_path_elt_type; + +typedef struct mg_path_elt +{ + mg_path_elt_type type; + vec2 p[3]; + +} mg_path_elt; + +typedef struct mg_path_descriptor +{ + u32 startIndex; + u32 count; + vec2 startPoint; + +} mg_path_descriptor; + +typedef struct mg_attributes +{ + f32 width; + f32 tolerance; + mg_color color; + mg_joint_type joint; + f32 maxJointExcursion; + mg_cap_type cap; + + mg_font font; + f32 fontSize; + + mg_image image; + + mp_rect clip; + +} mg_attributes; + +typedef struct mg_rounded_rect +{ + f32 x; + f32 y; + f32 w; + f32 h; + f32 r; +} mg_rounded_rect; + +typedef enum { MG_CMD_CLEAR = 0, + MG_CMD_FILL, + MG_CMD_STROKE, + MG_CMD_RECT_FILL, + MG_CMD_RECT_STROKE, + MG_CMD_ROUND_RECT_FILL, + MG_CMD_ROUND_RECT_STROKE, + MG_CMD_ELLIPSE_FILL, + MG_CMD_ELLIPSE_STROKE, + MG_CMD_JUMP, + MG_CMD_MATRIX_PUSH, + MG_CMD_MATRIX_POP, + MG_CMD_CLIP_PUSH, + MG_CMD_CLIP_POP, + MG_CMD_IMAGE_DRAW, + MG_CMD_ROUNDED_IMAGE_DRAW, + } mg_primitive_cmd; + +typedef struct mg_primitive +{ + mg_primitive_cmd cmd; + mg_attributes attributes; + + union + { + mg_path_descriptor path; + mp_rect rect; + mg_rounded_rect roundedRect; + utf32 codePoint; + u32 jump; + mg_mat2x3 matrix; + }; + +} mg_primitive; + +typedef struct mg_glyph_map_entry +{ + unicode_range range; + u32 firstGlyphIndex; + +} mg_glyph_map_entry; + +typedef struct mg_glyph_info +{ + bool exists; + utf32 codePoint; + mg_path_descriptor pathDescriptor; + mg_text_extents extents; + //... + +} mg_glyph_info; + +const u32 MG_STREAM_MAX_COUNT = 128; +const u32 MG_IMAGE_MAX_COUNT = 128; + +typedef struct mg_stream_data +{ + list_elt freeListElt; + u32 generation; + u64 frame; + + u32 firstCommand; + u32 lastCommand; + u32 count; + + bool pendingJump; + +} mg_stream_data; + +typedef struct mg_image_data +{ + list_elt listElt; + u32 generation; + + mp_rect rect; + +} mg_image_data; + +const u32 MG_MATRIX_STACK_MAX_DEPTH = 64; +const u32 MG_CLIP_STACK_MAX_DEPTH = 64; +const u32 MG_MAX_PATH_ELEMENT_COUNT = 2<<20; +const u32 MG_MAX_PRIMITIVE_COUNT = 8<<10; + +typedef struct mg_canvas_data +{ + list_elt freeListElt; + u32 generation; + + mg_stream_data streams[MG_STREAM_MAX_COUNT]; + list_info streamsFreeList; + u64 frameCounter; + + mg_stream_data* currentStream; + + mg_mat2x3 transform; + mp_rect clip; + + mg_attributes attributes; + bool textFlip; + mg_path_elt pathElements[MG_MAX_PATH_ELEMENT_COUNT]; + mg_path_descriptor path; + vec2 subPathStartPoint; + vec2 subPathLastPoint; + + mg_mat2x3 matrixStack[MG_MATRIX_STACK_MAX_DEPTH]; + u32 matrixStackSize; + + mp_rect clipStack[MG_CLIP_STACK_MAX_DEPTH]; + u32 clipStackSize; + + u32 nextZIndex; + u32 primitiveCount; + mg_primitive primitives[MG_MAX_PRIMITIVE_COUNT]; + + u32 vertexCount; + u32 indexCount; + + mg_canvas_painter* painter; + + mg_image_data images[MG_IMAGE_MAX_COUNT]; + u32 imageNextIndex; + list_info imageFreeList; + + vec2 atlasPos; + u32 atlasLineHeight; + mg_image blankImage; + +} mg_canvas_data; + +typedef struct mg_font_info +{ + list_elt freeListElt; + u32 generation; + + u32 rangeCount; + u32 glyphCount; + u32 outlineCount; + mg_glyph_map_entry* glyphMap; + mg_glyph_info* glyphs; + mg_path_elt* outlines; + + f32 unitsPerEm; + mg_font_extents extents; + +} mg_font_info; + +//--------------------------------------------------------------- +// internal handle system +//--------------------------------------------------------------- + +const u32 MG_MAX_SURFACES = 256, + MG_MAX_CONTEXTS = 256, + MG_FONT_MAX_COUNT = 256; + +typedef struct mg_surface_slot +{ + list_elt freeListElt; + u32 generation; + mg_surface_info* surface; + +} mg_surface_slot; + +typedef struct mg_info +{ + bool init; + list_info surfaceFreeList; + list_info contextFreeList; + list_info fontFreeList; + u32 fontsNextIndex; + mg_surface_slot surfaceSlots[MG_MAX_SURFACES]; + mg_canvas_data contexts[MG_MAX_CONTEXTS]; + mg_font_info fonts[MG_FONT_MAX_COUNT]; + +} mg_info; + +static mg_info __mgInfo = {0}; + +void mg_init() +{ + if(!__mgInfo.init) + { + ListInit(&__mgInfo.surfaceFreeList); + for(int i=0; igeneration == UINT32_MAX) + { + LOG_ERROR("surface slot generation wrap around\n"); + } + #endif + slot->generation++; + ListPush(&__mgInfo.surfaceFreeList, &slot->freeListElt); +} + +mg_surface_slot* mg_surface_slot_from_handle(mg_surface surface) +{ + u32 index = surface.h>>32; + u32 generation = surface.h & 0xffffffff; + if(index >= MG_MAX_SURFACES) + { + return(0); + } + mg_surface_slot* slot = &__mgInfo.surfaceSlots[index]; + if(slot->generation != generation) + { + return(0); + } + else + { + return(slot); + } +} + +mg_surface_info* mg_surface_ptr_from_handle(mg_surface surface) +{ + mg_surface_slot* slot = mg_surface_slot_from_handle(surface); + if(slot) + { + return(slot->surface); + } + else + { + return(0); + } +} + +mg_surface mg_surface_handle_from_slot(mg_surface_slot* slot) +{ + DEBUG_ASSERT( (slot - __mgInfo.surfaceSlots) >= 0 + && (slot - __mgInfo.surfaceSlots) < MG_MAX_SURFACES); + + u64 h = ((u64)(slot - __mgInfo.surfaceSlots))<<32 + |((u64)(slot->generation)); + return((mg_surface){.h = h}); +} + +mg_surface mg_surface_alloc_handle(mg_surface_info* surface) +{ + mg_surface_slot* slot = mg_surface_slot_alloc(); + slot->surface = surface; + mg_surface handle = mg_surface_handle_from_slot(slot); + return(handle); +} + +mg_surface mg_surface_nil() +{ + return((mg_surface){.h = 0}); +} + +mg_canvas_data* mg_canvas_alloc() +{ + return(ListPopEntry(&__mgInfo.contextFreeList, mg_canvas_data, freeListElt)); +} + +void mg_canvas_recycle(mg_canvas_data* context) +{ + #ifdef DEBUG + if(context->generation == UINT32_MAX) + { + LOG_ERROR("graphics context generation wrap around\n"); + } + #endif + context->generation++; + ListPush(&__mgInfo.contextFreeList, &context->freeListElt); +} + +mg_canvas_data* mg_canvas_ptr_from_handle(mg_canvas handle) +{ + u32 index = handle.h>>32; + u32 generation = handle.h & 0xffffffff; + if(index >= MG_MAX_CONTEXTS) + { + return(0); + } + mg_canvas_data* context = &__mgInfo.contexts[index]; + if(context->generation != generation) + { + return(0); + } + else + { + return(context); + } +} + +mg_canvas mg_canvas_handle_from_ptr(mg_canvas_data* context) +{ + DEBUG_ASSERT( (context - __mgInfo.contexts) >= 0 + && (context - __mgInfo.contexts) < MG_MAX_CONTEXTS); + + u64 h = ((u64)(context - __mgInfo.contexts))<<32 + |((u64)(context->generation)); + return((mg_canvas){.h = h}); +} + +mg_canvas mg_canvas_nil() +{ + return((mg_canvas){.h = 0}); +} + +mg_image_data* mg_image_ptr_from_handle(mg_canvas_data* context, mg_image handle) +{ + u32 index = handle.h>>32; + u32 generation = handle.h & 0xffffffff; + + if(index >= MG_IMAGE_MAX_COUNT) + { + return(0); + } + mg_image_data* image = &context->images[index]; + if(image->generation != generation) + { + return(0); + } + else + { + return(image); + } +} + +//--------------------------------------------------------------- +// graphics surface API +//--------------------------------------------------------------- + +#ifdef MG_IMPLEMENTS_BACKEND_METAL +//NOTE: function is defined in metal_backend.mm +mg_surface mg_metal_surface_create_for_window(mp_window window); +mg_surface mg_metal_surface_create_for_view(mp_view view); +#endif //MG_IMPLEMENTS_BACKEND_METAL + +void mg_init(); + +mg_surface mg_surface_create_for_window(mp_window window, mg_backend_id backend) +{ + DEBUG_ASSERT(__mgInfo.init); + + mg_surface surface = mg_surface_nil(); + + switch(backend) + { + #ifdef MG_IMPLEMENTS_BACKEND_METAL + case MG_BACKEND_METAL: + surface = mg_metal_surface_create_for_window(window); + break; + #endif + + //... + + default: + break; + } + + return(surface); +} + +mg_surface mg_surface_create_for_view(mp_view view, mg_backend_id backend) +{ + DEBUG_ASSERT(__mgInfo.init); + + mg_surface surface = mg_surface_nil(); + + switch(backend) + { + #ifdef MG_IMPLEMENTS_BACKEND_METAL + case MG_BACKEND_METAL: + surface = mg_metal_surface_create_for_view(view); + break; + #endif + + //... + + default: + break; + } + + return(surface); +} + + +void mg_surface_destroy(mg_surface handle) +{ + DEBUG_ASSERT(__mgInfo.init); + + mg_surface_slot* slot = mg_surface_slot_from_handle(handle); + if(slot) + { + slot->surface->destroy(slot->surface); + mg_surface_slot_recycle(slot); + } +} + +void mg_surface_resize(mg_surface surface, int width, int height) +{ + DEBUG_ASSERT(__mgInfo.init); + + mg_surface_info* surfaceInfo = mg_surface_ptr_from_handle(surface); + if(surfaceInfo) + { + surfaceInfo->resize(surfaceInfo, width, height); + } +} + +void mg_surface_prepare(mg_surface surface) +{ + DEBUG_ASSERT(__mgInfo.init); + + mg_surface_info* surfaceInfo = mg_surface_ptr_from_handle(surface); + if(surfaceInfo) + { + surfaceInfo->prepare(surfaceInfo); + } +} + +void mg_surface_present(mg_surface surface) +{ + DEBUG_ASSERT(__mgInfo.init); + + mg_surface_info* surfaceInfo = mg_surface_ptr_from_handle(surface); + if(surfaceInfo) + { + surfaceInfo->present(surfaceInfo); + } +} + +vec2 mg_surface_size(mg_surface surface) +{ + DEBUG_ASSERT(__mgInfo.init); + vec2 res = {}; + mg_surface_info* surfaceInfo = mg_surface_ptr_from_handle(surface); + if(surfaceInfo) + { + res = surfaceInfo->getSize(surfaceInfo); + } + return(res); +} + + +//--------------------------------------------------------------- +// graphics stream handles +//--------------------------------------------------------------- +/* + Graphics command stream handles are invalidated when the command stream is appended to the current stream, + and at the end of the frame. + Thus command streams handles contain the index of the stream, the frame of the stream, and a generation + count inside that frame: + + 0 40 56 64 + +---------------------------------+------------------------+------------+ + | frameCounter | generation | index | + +---------------------------------+------------------------+------------+ + 40 bits 16 bits 8 bits + + (This gives use 2^40 frames, ie ~600 years at 60Hz, 65536 possible reuse per frames and max 256 simultaneous streams.) + + This way we can invalidate use of old handles in the same frame by incrementing the generation counter of the recycled stream, + and we can invalidate all previous handles at the end of a frame by just incrementing the frame counter of the graphics context. +*/ +mg_stream_data* mg_stream_alloc(mg_canvas_data* context) +{ + mg_stream_data* stream = ListPopEntry(&context->streamsFreeList, mg_stream_data, freeListElt); + if(stream) + { + stream->frame = context->frameCounter; + stream->pendingJump = false; + stream->count = 0; + stream->firstCommand = 0; + stream->lastCommand = 0; + } + return(stream); +} + +void mg_stream_recycle(mg_canvas_data* context, mg_stream_data* stream) +{ + #ifdef DEBUG + if(stream->generation == UINT32_MAX) + { + LOG_ERROR("graphics command stream generation wrap around\n"); + } + #endif + stream->generation++; + ListPush(&context->streamsFreeList, &stream->freeListElt); +} + +mg_stream_data* mg_stream_ptr_from_handle(mg_canvas_data* context, mg_stream handle) +{ + u32 index = handle.h>>56; + u32 generation = (handle.h>>40) & 0xffff; + u64 frame = handle.h & 0xffffffffff; + + if(index >= MG_STREAM_MAX_COUNT) + { + return(0); + } + mg_stream_data* stream = &context->streams[index]; + if( stream->generation != generation + ||stream->frame != context->frameCounter) + { + return(0); + } + else + { + return(stream); + } +} + +mg_stream mg_stream_handle_from_ptr(mg_canvas_data* context, mg_stream_data* stream) +{ + DEBUG_ASSERT( (stream - context->streams) >= 0 + && (stream - context->streams) < MG_STREAM_MAX_COUNT); + + u64 h = ((u64)(stream - context->streams))<<56 + |((u64)(stream->generation))<<40 + |((u64)(context->frameCounter)); + + return((mg_stream){.h = h}); +} + +mg_stream mg_stream_null_handle() +{ + return((mg_stream){.h = 0}); +} + +//--------------------------------------------------------------- +// internal graphics context functions +//--------------------------------------------------------------- + +mg_mat2x3 mg_mat2x3_mul_m(mg_mat2x3 lhs, mg_mat2x3 rhs) +{ + mg_mat2x3 res; + res.m[0] = lhs.m[0]*rhs.m[0] + lhs.m[1]*rhs.m[3]; + res.m[1] = lhs.m[0]*rhs.m[1] + lhs.m[1]*rhs.m[4]; + res.m[2] = lhs.m[0]*rhs.m[2] + lhs.m[1]*rhs.m[5] + lhs.m[2]; + res.m[3] = lhs.m[3]*rhs.m[0] + lhs.m[4]*rhs.m[3]; + res.m[4] = lhs.m[3]*rhs.m[1] + lhs.m[4]*rhs.m[4]; + res.m[5] = lhs.m[3]*rhs.m[2] + lhs.m[4]*rhs.m[5] + lhs.m[5]; + + return(res); +} + +vec2 mg_mat2x3_mul(mg_mat2x3 m, vec2 p) +{ + f32 x = p.x*m.m[0] + p.y*m.m[1] + m.m[2]; + f32 y = p.x*m.m[3] + p.y*m.m[4] + m.m[5]; + return((vec2){x, y}); +} + +void mg_reset_z_index(mg_canvas_data* context) +{ + context->nextZIndex = 1; +} + +u32 mg_get_next_z_index(mg_canvas_data* context) +{ + return(context->nextZIndex++); +} + +/////////////////////////////////////// WIP ///////////////////////////////////////////////////////////////////////// +u32 mg_vertices_base_index(mg_canvas_data* context) +{ + return(context->vertexCount); +} + +void mg_push_textured_vertex(mg_canvas_data* context, vec2 pos, vec4 cubic, vec2 uv, mg_color color, u64 zIndex) +{ + mgc_vertex_layout* layout = &context->painter->vertexLayout; + + DEBUG_ASSERT(context->vertexCount < layout->maxVertexCount); + + u32 offset = context->vertexCount; + + *(vec2*)(((char*)layout->posBuffer) + offset*layout->posStride) = pos; + *(vec4*)(((char*)layout->cubicBuffer) + offset*layout->cubicStride) = cubic; + *(vec2*)(((char*)layout->uvBuffer) + offset*layout->uvStride) = uv; + *(mg_color*)(((char*)layout->colorBuffer) + offset*layout->colorStride) = color; + *(u32*)(((char*)layout->zIndexBuffer) + offset*layout->zIndexStride) = zIndex; + *(mp_rect*)(((char*)layout->clipsBuffer) + offset*layout->clipsStride) = context->clip; + + context->vertexCount++; +} + +void mg_push_vertex(mg_canvas_data* context, vec2 pos, vec4 cubic, mg_color color, u64 zIndex) +{ + mg_push_textured_vertex(context, pos, cubic, (vec2){0, 0}, color, zIndex); +} + + +/////////////////////////////////////// WIP ///////////////////////////////////////////////////////////////////////// + +int* mg_reserve_indices(mg_canvas_data* context, u32 indexCount) +{ + mgc_vertex_layout* layout = &context->painter->vertexLayout; + + ASSERT(context->indexCount + indexCount < layout->maxIndexCount); + int* base = ((int*)layout->indexBuffer) + context->indexCount; + context->indexCount += indexCount; + return(base); +} + +//----------------------------------------------------------------------------------------------------------- +// Path Filling +//----------------------------------------------------------------------------------------------------------- +//NOTE(martin): forward declarations +void mg_render_fill_cubic(mg_canvas_data* context, vec2 p[4], u32 zIndex, mg_color color); + +//NOTE(martin): quadratics filling + +void mg_render_fill_quadratic(mg_canvas_data* context, vec2 p[3], u32 zIndex, mg_color color) +{ + u32 baseIndex = mg_vertices_base_index(context); + + i32* indices = mg_reserve_indices(context, 3); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[0].x, p[0].y}), + (vec4){0, 0, 0, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[1].x, p[1].y}), + (vec4){0.5, 0, 0.5, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[2].x, p[2].y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; +} + +//NOTE(martin): cubic filling + +void mg_split_and_fill_cubic(mg_canvas_data* context, vec2 p[4], f32 tSplit, u32 zIndex, mg_color color) +{ + int subVertexCount = 0; + int subIndexCount = 0; + + f32 OneMinusTSplit = 1-tSplit; + + vec2 q0 = {OneMinusTSplit*p[0].x + tSplit*p[1].x, + OneMinusTSplit*p[0].y + tSplit*p[1].y}; + + vec2 q1 = {OneMinusTSplit*p[1].x + tSplit*p[2].x, + OneMinusTSplit*p[1].y + tSplit*p[2].y}; + + vec2 q2 = {OneMinusTSplit*p[2].x + tSplit*p[3].x, + OneMinusTSplit*p[2].y + tSplit*p[3].y}; + + vec2 r0 = {OneMinusTSplit*q0.x + tSplit*q1.x, + OneMinusTSplit*q0.y + tSplit*q1.y}; + + vec2 r1 = {OneMinusTSplit*q1.x + tSplit*q2.x, + OneMinusTSplit*q1.y + tSplit*q2.y}; + + vec2 split = {OneMinusTSplit*r0.x + tSplit*r1.x, + OneMinusTSplit*r0.y + tSplit*r1.y};; + + vec2 subPointsLow[4] = {p[0], q0, r0, split}; + vec2 subPointsHigh[4] = {split, r1, q2, p[3]}; + + //NOTE(martin): add base triangle + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 3); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[0].x, p[0].y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){split.x, split.y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[3].x, p[3].y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + + mg_render_fill_cubic(context, subPointsLow, zIndex, color); + mg_render_fill_cubic(context, subPointsHigh, zIndex, color); + + return; +} + +void mg_render_fill_cubic(mg_canvas_data* context, vec2 p[4], u32 zIndex, mg_color color) +{ + LOG_DEBUG("graphics render fill cubic\n"); + + vec4 testCoords[4]; + + /*NOTE(martin): first convert the control points to power basis, multiplying by M3 + + | 1 0 0 0| + M3 = |-3 3 0 0| + | 3 -6 3 0| + |-1 3 -3 1| + ie: + c0 = p0 + c1 = -3*p0 + 3*p1 + c2 = 3*p0 - 6*p1 + 3*p2 + c3 = -p0 + 3*p1 - 3*p2 + p3 + */ + f32 c1x = 3.0*p[1].x - 3.0*p[0].x; + f32 c1y = 3.0*p[1].y - 3.0*p[0].y; + + f32 c2x = 3.0*p[0].x + 3.0*p[2].x - 6.0*p[1].x; + f32 c2y = 3.0*p[0].y + 3.0*p[2].y - 6.0*p[1].y; + + f32 c3x = 3.0*p[1].x - 3.0*p[2].x + p[3].x - p[0].x; + f32 c3y = 3.0*p[1].y - 3.0*p[2].y + p[3].y - p[0].y; + + //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + //TODO(martin): we should do the tex coords computations in f64 and scale them to avoid f32 precision/range glitches in shader + //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + c1x /= 10; + c1y /= 10; + c2x /= 10; + c2y /= 10; + c3x /= 10; + c3y /= 10; + + /*NOTE(martin): + now, compute determinants d0, d1, d2, d3, which gives the coefficients of the + inflection points polynomial: + + I(t, s) = d0*t^3 - 3*d1*t^2*s + 3*d2*t*s^2 - d3*s^3 + + The roots of this polynomial are the inflection points of the parametric curve, in homogeneous + coordinates (ie we can have an inflection point at inifinity with s=0). + + |x3 y3 w3| |x3 y3 w3| |x3 y3 w3| |x2 y2 w2| + d0 = det |x2 y2 w2| d1 = -det |x2 y2 w2| d2 = det |x1 y1 w1| d3 = -det |x1 y1 w1| + |x1 y1 w1| |x0 y0 w0| |x0 y0 w0| |x0 y0 w0| + + In our case, the pi.w equal 1 (no point at infinity), so _in_the_power_basis_, w1 = w2 = w3 = 0 and w0 = 1 + (which also means d0 = 0) + */ + + f32 d1 = c3y*c2x - c3x*c2y; + f32 d2 = c3x*c1y - c3y*c1x; + f32 d3 = c2y*c1x - c2x*c1y; + + //NOTE(martin): compute the second factor of the discriminant discr(I) = d1^2*(3*d2^2 - 4*d3*d1) + f32 discrFactor2 = 3.0*Square(d2) - 4.0*d3*d1; + + //NOTE(martin): each following case gives the number of roots, hence the category of the parametric curve + if(fabs(d1) < 0.1 && fabs(d2) < 0.1 && d3 != 0) + { + //NOTE(martin): quadratic degenerate case + LOG_DEBUG("quadratic curve\n"); + + //NOTE(martin): compute quadratic curve control point, which is at p0 + 1.5*(p1-p0) = 1.5*p1 - 0.5*p0 + vec2 quadControlPoints[3] = { p[0], + {1.5*p[1].x - 0.5*p[0].x, 1.5*p[1].y - 0.5*p[0].y}, + p[3]}; + + mg_render_fill_quadratic(context, quadControlPoints, zIndex, color); + return; + } + else if( (discrFactor2 > 0 && d1 != 0) + ||(discrFactor2 == 0 && d1 != 0)) + { + //NOTE(martin): serpentine curve or cusp with inflection at infinity + // (these two cases are handled the same way). + LOG_DEBUG("%s\n", (discrFactor2 > 0 && d1 != 0) ? "serpentine curve" : "cusp with inflection at infinity"); + + //NOTE(martin): compute the solutions (tl, sl), (tm, sm), and (tn, sn) of the inflection point equation + f32 tl = d2 + sqrt(discrFactor2/3); + f32 sl = 2*d1; + f32 tm = d2 - sqrt(discrFactor2/3); + f32 sm = sl; + + /*NOTE(martin): + the power basis coefficients of points k,l,m,n are collected into the rows of the 4x4 matrix F: + + | tl*tm tl^3 tm^3 1 | + | -sm*tl - sl*tm -3sl*tl^2 -3*sm*tm^2 0 | + | sl*sm 3*sl^2*tl 3*sm^2*tm 0 | + | 0 -sl^3 -sm^3 0 | + + This matrix is then multiplied by M3^(-1) on the left which yelds the bezier coefficients of k, l, m, n + which are assigned as a 4D texture coordinates to control points. + + + | 1 0 0 0 | + M3^(-1) = | 1 1/3 0 0 | + | 1 2/3 1/3 0 | + | 1 1 1 1 | + */ + testCoords[0].x = tl*tm; + testCoords[0].y = Cube(tl); + testCoords[0].z = Cube(tm); + + testCoords[1].x = tl*tm - (sm*tl + sl*tm)/3; + testCoords[1].y = Cube(tl) - sl*Square(tl); + testCoords[1].z = Cube(tm) - sm*Square(tm); + + testCoords[2].x = tl*tm - (sm*tl + sl*tm)*2/3 + sl*sm/3; + testCoords[2].y = Cube(tl) - 2*sl*Square(tl) + Square(sl)*tl; + testCoords[2].z = Cube(tm) - 2*sm*Square(tm) + Square(sm)*tm; + + testCoords[3].x = tl*tm - (sm*tl + sl*tm) + sl*sm; + testCoords[3].y = Cube(tl) - 3*sl*Square(tl) + 3*Square(sl)*tl - Cube(sl); + testCoords[3].z = Cube(tm) - 3*sm*Square(tm) + 3*Square(sm)*tm - Cube(sm); + } + else if(discrFactor2 < 0 && d1 != 0) + { + //NOTE(martin): loop curve + LOG_DEBUG("loop curve\n"); + + f32 td = d2 + sqrt(-discrFactor2); + f32 sd = 2*d1; + f32 te = d2 - sqrt(-discrFactor2); + f32 se = sd; + + //NOTE(martin): if one of the parameters (td/sd) or (te/se) is in the interval [0,1], the double point + // is inside the control points convex hull and would cause a shading anomaly. If this is + // the case, subdivide the curve at that point + + //TODO: study edge case where td/sd ~ 1 or 0 (which causes an infinite recursion in split and fill). + // quick fix for now is adding a little slop in the check... + + if(sd != 0 && td/sd < 0.99 && td/sd > 0.01) + { + LOG_DEBUG("split curve at first double point\n"); + mg_split_and_fill_cubic(context, p, td/sd, zIndex, color); + return; + } + if(se != 0 && te/se < 0.99 && te/se > 0.01) + { + LOG_DEBUG("split curve at second double point\n"); + mg_split_and_fill_cubic(context, p, te/se, zIndex, color); + return; + } + + /*NOTE(martin): + the power basis coefficients of points k,l,m,n are collected into the rows of the 4x4 matrix F: + + | td*te td^2*te td*te^2 1 | + | -se*td - sd*te -se*td^2 - 2sd*te*td -sd*te^2 - 2*se*td*te 0 | + | sd*se te*sd^2 + 2*se*td*sd td*se^2 + 2*sd*te*se 0 | + | 0 -sd^2*se -sd*se^2 0 | + + This matrix is then multiplied by M3^(-1) on the left which yelds the bezier coefficients of k, l, m, n + which are assigned as a 4D texture coordinates to control points. + + + | 1 0 0 0 | + M3^(-1) = | 1 1/3 0 0 | + | 1 2/3 1/3 0 | + | 1 1 1 1 | + */ + testCoords[0].x = td*te; + testCoords[0].y = Square(td)*te; + testCoords[0].z = td*Square(te); + + testCoords[1].x = td*te - (se*td + sd*te)/3.0; + testCoords[1].y = Square(td)*te - (se*Square(td) + 2.*sd*te*td)/3.0; + testCoords[1].z = td*Square(te) - (sd*Square(te) + 2*se*td*te)/3.0; + + testCoords[2].x = td*te - 2.0*(se*td + sd*te)/3.0 + sd*se/3.0; + testCoords[2].y = Square(td)*te - 2.0*(se*Square(td) + 2.0*sd*te*td)/3.0 + (te*Square(sd) + 2.0*se*td*sd)/3.0; + testCoords[2].z = td*Square(te) - 2.0*(sd*Square(te) + 2.0*se*td*te)/3.0 + (td*Square(se) + 2.0*sd*te*se)/3.0; + + testCoords[3].x = td*te - (se*td + sd*te) + sd*se; + testCoords[3].y = Square(td)*te - (se*Square(td) + 2.0*sd*te*td) + (te*Square(sd) + 2.0*se*td*sd) - Square(sd)*se; + testCoords[3].z = td*Square(te) - (sd*Square(te) + 2.0*se*td*te) + (td*Square(se) + 2.0*sd*te*se) - sd*Square(se); + } + else if(d1 == 0 && d2 != 0) + { + //NOTE(martin): cusp with cusp at infinity + LOG_DEBUG("cusp at infinity curve\n"); + + f32 tl = d3; + f32 sl = 3*d2; + + /*NOTE(martin): + the power basis coefficients of points k,l,m,n are collected into the rows of the 4x4 matrix F: + + | tl tl^3 1 1 | + | -sl -3sl*tl^2 0 0 | + | 0 3*sl^2*tl 0 0 | + | 0 -sl^3 0 0 | + + This matrix is then multiplied by M3^(-1) on the left which yelds the bezier coefficients of k, l, m, n + which are assigned as a 4D texture coordinates to control points. + + + | 1 0 0 0 | + M3^(-1) = | 1 1/3 0 0 | + | 1 2/3 1/3 0 | + | 1 1 1 1 | + */ + + testCoords[0].x = tl; + testCoords[0].y = Cube(tl); + testCoords[0].z = 1; + + testCoords[1].x = tl - sl/3; + testCoords[1].y = Cube(tl) - sl*Square(tl); + testCoords[1].z = 1; + + testCoords[2].x = tl - sl*2/3; + testCoords[2].y = Cube(tl) - 2*sl*Square(tl) + Square(sl)*tl; + testCoords[2].z = 1; + + testCoords[3].x = tl - sl; + testCoords[3].y = Cube(tl) - 3*sl*Square(tl) + 3*Square(sl)*tl - Cube(sl); + testCoords[3].z = 1; + } + else if(d1 == 0 && d2 == 0 && d3 == 0) + { + //NOTE(martin): line or point degenerate case, ignored + LOG_DEBUG("line or point curve (ignored)\n"); + return; + } + else + { + //TODO(martin): handle error ? put some epsilon slack on the conditions ? + LOG_DEBUG("none of the above...\n"); + ASSERT(0, "not implemented yet !"); + return; + } + + //NOTE(martin): compute convex hull indices using Gift wrapping / Jarvis' march algorithm + int convexHullIndices[4]; + int leftMostPointIndex = 0; + + for(int i=0; i<4; i++) + { + if(p[i].x < p[leftMostPointIndex].x) + { + leftMostPointIndex = i; + } + } + int currentPointIndex = leftMostPointIndex; + int i=0; + int convexHullCount = 0; + + do + { + convexHullIndices[i] = currentPointIndex; + convexHullCount++; + int bestGuessIndex = 0; + + for(int j=0; j<4; j++) + { + vec2 bestGuessEdge = {.x = p[bestGuessIndex].x - p[currentPointIndex].x, + .y = p[bestGuessIndex].y - p[currentPointIndex].y}; + + vec2 nextGuessEdge = {.x = p[j].x - p[currentPointIndex].x, + .y = p[j].y - p[currentPointIndex].y}; + + //NOTE(martin): if control point j is on the right of current edge, it is a best guess + // (in case of colinearity we choose the point which is farthest from the current point) + + f32 crossProduct = bestGuessEdge.x*nextGuessEdge.y - bestGuessEdge.y*nextGuessEdge.x; + + if( bestGuessIndex == currentPointIndex + || crossProduct < 0) + { + bestGuessIndex = j; + } + else if(crossProduct == 0) + { + + //NOTE(martin): if vectors v1, v2 are colinear and distinct, and ||v1|| > ||v2||, + // either abs(v1.x) > abs(v2.x) or abs(v1.y) > abs(v2.y) + // so we don't actually need to compute their norm to select the greatest + // (and if v1 and v2 are equal we don't have to update our best guess.) + + //TODO(martin): in case of colinearity we should rather select the edge that has the greatest dot product with last edge ?? + + if(fabs(nextGuessEdge.x) > fabs(bestGuessEdge.x) + || fabs(nextGuessEdge.y) > fabs(bestGuessEdge.y)) + { + bestGuessIndex = j; + } + } + } + i++; + currentPointIndex = bestGuessIndex; + + } while(currentPointIndex != leftMostPointIndex && i<4); + + //TODO: quick fix, maybe later cull degenerate hulls beforehand + if(convexHullCount <= 2) + { + //NOTE(martin): if convex hull has only two point, we have a degenerate cubic that displays nothing. + return; + } + + //NOTE(martin): rearrange convex hull to put p0 first + int startIndex = -1; + int orderedHullIndices[4]; + for(int i=0; i= 0) + { + orderedHullIndices[i-startIndex] = convexHullIndices[i]; + } + } + for(int i=0; i 0 + // ie the 4th coordinate s flips the inside/outside test. + // We affect s such that control points p1 and p2 are always outside the covered area. + + if(convexHullCount <= 3) + { + //NOTE(martin): the convex hull is a triangle + + //NOTE(martin): when we degenerate from 4 control points to a 3 points convex hull, this means one of the + // control points p1 or p2 could be inside the covered area. We want to compute the test for the + // control point which belongs to the convex hull, as we know it should be outside the covered area. + // + // Since there are 3 points in the hull and p0 is the first, and p3 belongs to the hull, this means we + // must select the point of the convex hull which is neither the first, nor p3. + + int testPointIndex = orderedHullIndices[1] == 3 ? orderedHullIndices[2] : orderedHullIndices[1]; + int outsideTest = 1; + if(Cube(testCoords[testPointIndex].x)-testCoords[testPointIndex].y*testCoords[testPointIndex].z < 0) + { + outsideTest = -1; + } + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 3); + + for(int i=0; i<3; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, p[orderedHullIndices[i]]); + vec4 cubic = testCoords[orderedHullIndices[i]]; + cubic.w = outsideTest; + mg_push_vertex(context, pos, cubic, color, zIndex); + + indices[i] = baseIndex + i; + } + } + else if(orderedHullIndices[2] == 3) + { + //NOTE(martin): p1 and p2 are not on the same side of (p0,p3). The outside test can be different for + // the two triangles + int outsideTest1 = 1; + int outsideTest2 = 1; + + int testIndex = orderedHullIndices[1]; + if(Cube(testCoords[testIndex].x)-testCoords[testIndex].y*testCoords[testIndex].z < 0) + { + outsideTest1 = -1; + } + + testIndex = orderedHullIndices[3]; + if(Cube(testCoords[testIndex].x)-testCoords[testIndex].y*testCoords[testIndex].z < 0) + { + outsideTest2 = -1; + } + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[0]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[0]]), outsideTest1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[1]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[1]]), outsideTest1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[2]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[2]]), outsideTest1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[0]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[0]]), outsideTest2}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[2]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[2]]), outsideTest2}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[3]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[3]]), outsideTest2}, + color, + zIndex); + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 3; + indices[4] = baseIndex + 4; + indices[5] = baseIndex + 5; + } + else + { + //NOTE(martin): if p1 and p2 are on the same side of (p0,p3), the outside test is the same for both triangles + int outsideTest = 1; + if(Cube(testCoords[1].x)-testCoords[1].y*testCoords[1].z < 0) + { + outsideTest = -1; + } + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + for(int i=0; i<4; i++) + { + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p[orderedHullIndices[i]]), + (vec4){vec4_expand_xyz(testCoords[orderedHullIndices[i]]), outsideTest}, + color, + zIndex); + } + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; + } +} + +//NOTE(martin): global path fill + +void mg_render_fill(mg_canvas_data* context, mg_path_elt* elements, mg_path_descriptor* path, u32 zIndex, mg_color color) +{ + u32 eltCount = path->count; + vec2 startPoint = path->startPoint; + vec2 endPoint = path->startPoint; + vec2 currentPoint = path->startPoint; + + for(int eltIndex=0; eltIndexp[0], elt->p[1], elt->p[2]}; + + switch(elt->type) + { + case MG_PATH_MOVE: + { + startPoint = elt->p[0]; + endPoint = elt->p[0]; + currentPoint = endPoint; + continue; + } break; + + case MG_PATH_LINE: + { + endPoint = controlPoints[1]; + } break; + + case MG_PATH_QUADRATIC: + { + mg_render_fill_quadratic(context, controlPoints, zIndex, color); + endPoint = controlPoints[2]; + + } break; + + case MG_PATH_CUBIC: + { + mg_render_fill_cubic(context, controlPoints, zIndex, color); + endPoint = controlPoints[3]; + } break; + } + + //NOTE(martin): now fill interior triangle + u32 baseIndex = mg_vertices_base_index(context); + int* indices = mg_reserve_indices(context, 3); + + vec2 pos[3]; + pos[0] = mg_mat2x3_mul(context->transform, startPoint); + pos[1] = mg_mat2x3_mul(context->transform, currentPoint); + pos[2] = mg_mat2x3_mul(context->transform, endPoint); + + vec4 cubic = {1, 1, 1, 1}; + + for(int i=0; i<3; i++) + { + mg_push_vertex(context, pos[i], cubic, color, zIndex); + indices[i] = baseIndex + i; + } + + currentPoint = endPoint; + } +} + +//----------------------------------------------------------------------------------------------------------- +// Path Stroking +//----------------------------------------------------------------------------------------------------------- + +void mg_render_stroke_line(mg_canvas_data* context, vec2 p[2], u32 zIndex, mg_attributes* attributes) +{ + //NOTE(martin): get normals multiplied by halfWidth + f32 halfW = attributes->width/2; + + vec2 n0 = {p[0].y - p[1].y, + p[1].x - p[0].x}; + f32 norm0 = sqrt(n0.x*n0.x + n0.y*n0.y); + n0.x *= halfW/norm0; + n0.y *= halfW/norm0; + + + mg_color color = attributes->color; + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[0].x + n0.x, p[0].y + n0.y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[1].x + n0.x, p[1].y + n0.y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[1].x - n0.x, p[1].y - n0.y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p[0].x - n0.x, p[0].y - n0.y}), + (vec4){1, 1, 1, 1}, + color, + zIndex); + + indices[0] = baseIndex; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; +} + +void mg_offset_hull(int count, vec2* p, vec2* result, f32 offset) +{ +////////////////////////////////////////////////////////////////////////////////////// +//WARN: quick fix for coincident middle control points + if(count == 4 && (p[1].x - p[2].x < 0.01) && (p[1].y - p[2].y < 0.01)) + { + vec2 hull3[3] = {p[0], p[1], p[3]}; + vec2 result3[3]; + mg_offset_hull(3, hull3, result3, offset); + result[0] = result3[0]; + result[1] = result3[1]; + result[2] = result3[1]; + result[3] = result3[2]; + return; + } +/////////////////////////////////////////////////////////////////////////////////////: + + //TODO(martin): review edge cases (coincident points ? colinear points ? control points pointing outward end point?) + //NOTE(martin): first offset control point is just the offset of first control point + vec2 n = {p[0].y - p[1].y, + p[1].x - p[0].x}; + f32 norm = sqrt(n.x*n.x + n.y*n.y); + n.x *= offset/norm; + n.y *= offset/norm; + + result[0].x = p[0].x + n.x; + result[0].y = p[0].y + n.y; + + //NOTE(martin): subsequent offset control points are the intersection of offset control lines + for(int i=1; iwidth); + mg_offset_hull(3, p, negativeOffsetHull, -0.5 * attributes->width); + + //NOTE(martin): the distance d between the offset curve and the path must be between w/2-tolerance and w/2+tolerance + // thus, by constraining tolerance to be at most, 0.5*width, we can rewrite this condition like this: + // + // (w/2-tolerance)^2 < d^2 < (w/2+tolerance)^2 + // + // we compute the maximum overshoot outside these bounds and split the curve at the corresponding parameter + + f32 tolerance = minimum(attributes->tolerance, 0.5 * attributes->width); + f32 d2LowBound = Square(0.5 * attributes->width - attributes->tolerance); + f32 d2HighBound = Square(0.5 * attributes->width + attributes->tolerance); + + f32 maxOvershoot = 0; + f32 maxOvershootParameter = 0; + + for(int i=0; i maxOvershoot) + { + maxOvershoot = overshoot; + maxOvershootParameter = t; + } + } + + if(maxOvershoot > 0) + { + //TODO(martin): split at maxErrorParameter and recurse + vec2 splitLeft[3]; + vec2 splitRight[3]; + mg_quadratic_split(p, maxOvershootParameter, splitLeft, splitRight); + mg_render_stroke_quadratic(context, splitLeft, zIndex, attributes); + mg_render_stroke_quadratic(context, splitRight, zIndex, attributes); + } + else + { + //NOTE(martin): push the actual fill commands for the offset contour + + u32 zIndex = mg_get_next_z_index(context); + + mg_render_fill_quadratic(context, positiveOffsetHull, zIndex, attributes->color); + mg_render_fill_quadratic(context, negativeOffsetHull, zIndex, attributes->color); + + //NOTE(martin): add base triangles + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, positiveOffsetHull[0]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, positiveOffsetHull[2]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, negativeOffsetHull[2]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, negativeOffsetHull[0]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; + } + #undef CHECK_SAMPLE_COUNT +} + +vec2 mg_cubic_get_point(vec2 p[4], f32 t) +{ + vec2 r; + + f32 oneMt = 1-t; + f32 oneMt2 = Square(oneMt); + f32 oneMt3 = oneMt2*oneMt; + f32 t2 = Square(t); + f32 t3 = t2*t; + + r.x = oneMt3*p[0].x + 3*oneMt2*t*p[1].x + 3*oneMt*t2*p[2].x + t3*p[3].x; + r.y = oneMt3*p[0].y + 3*oneMt2*t*p[1].y + 3*oneMt*t2*p[2].y + t3*p[3].y; + + return(r); +} + +void mg_cubic_split(vec2 p[4], f32 t, vec2 outLeft[4], vec2 outRight[4]) +{ + //NOTE(martin): split bezier curve p at parameter t, using De Casteljau's algorithm + // the q_n are the points along the hull's segments at parameter t + // the r_n are the points along the (q_n, q_n+1) segments at parameter t + // s is the split point. + + f32 oneMt = 1-t; + + vec2 q0 = {oneMt*p[0].x + t*p[1].x, + oneMt*p[0].y + t*p[1].y}; + + vec2 q1 = {oneMt*p[1].x + t*p[2].x, + oneMt*p[1].y + t*p[2].y}; + + vec2 q2 = {oneMt*p[2].x + t*p[3].x, + oneMt*p[2].y + t*p[3].y}; + + vec2 r0 = {oneMt*q0.x + t*q1.x, + oneMt*q0.y + t*q1.y}; + + vec2 r1 = {oneMt*q1.x + t*q2.x, + oneMt*q1.y + t*q2.y}; + + vec2 s = {oneMt*r0.x + t*r1.x, + oneMt*r0.y + t*r1.y};; + + outLeft[0] = p[0]; + outLeft[1] = q0; + outLeft[2] = r0; + outLeft[3] = s; + + outRight[0] = s; + outRight[1] = r1; + outRight[2] = q2; + outRight[3] = p[3]; +} + +void mg_render_stroke_cubic(mg_canvas_data* context, vec2 p[4], u32 zIndex, mg_attributes* attributes) +{ + #define CHECK_SAMPLE_COUNT 5 + f32 checkSamples[CHECK_SAMPLE_COUNT] = {1./6, 2./6, 3./6, 4./6, 5./6}; + + vec2 positiveOffsetHull[4]; + vec2 negativeOffsetHull[4]; + + mg_offset_hull(4, p, positiveOffsetHull, 0.5 * attributes->width); + mg_offset_hull(4, p, negativeOffsetHull, -0.5 * attributes->width); + + //NOTE(martin): the distance d between the offset curve and the path must be between w/2-tolerance and w/2+tolerance + // thus, by constraining tolerance to be at most, 0.5*width, we can rewrite this condition like this: + // + // (w/2-tolerance)^2 < d^2 < (w/2+tolerance)^2 + // + // we compute the maximum overshoot outside these bounds and split the curve at the corresponding parameter + + f32 tolerance = minimum(attributes->tolerance, 0.5 * attributes->width); + f32 d2LowBound = Square(0.5 * attributes->width - attributes->tolerance); + f32 d2HighBound = Square(0.5 * attributes->width + attributes->tolerance); + + f32 maxOvershoot = 0; + f32 maxOvershootParameter = 0; + + for(int i=0; i maxOvershoot) + { + maxOvershoot = overshoot; + maxOvershootParameter = t; + } + } + + if(maxOvershoot > 0) + { + //TODO(martin): split at maxErrorParameter and recurse + vec2 splitLeft[4]; + vec2 splitRight[4]; + mg_cubic_split(p, maxOvershootParameter, splitLeft, splitRight); + mg_render_stroke_cubic(context, splitLeft, zIndex, attributes); + mg_render_stroke_cubic(context, splitRight, zIndex, attributes); + } + else + { + //NOTE(martin): push the actual fill commands for the offset contour + + u32 zIndex = mg_get_next_z_index(context); + + mg_render_fill_cubic(context, positiveOffsetHull, zIndex, attributes->color); + mg_render_fill_cubic(context, negativeOffsetHull, zIndex, attributes->color); + + //NOTE(martin): add base triangles + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, positiveOffsetHull[0]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, positiveOffsetHull[3]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, negativeOffsetHull[3]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, negativeOffsetHull[0]), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; + } + #undef CHECK_SAMPLE_COUNT +} + +void mg_stroke_cap(mg_canvas_data* context, vec2 p0, vec2 direction, mg_attributes* attributes) +{ + //NOTE(martin): compute the tangent and normal vectors (multiplied by half width) at the cap point + + f32 dn = sqrt(Square(direction.x) + Square(direction.y)); + f32 alpha = 0.5 * attributes->width/dn; + + vec2 n0 = {-alpha*direction.y, + alpha*direction.x}; + + vec2 m0 = {alpha*direction.x, + alpha*direction.y}; + + u32 zIndex = mg_get_next_z_index(context); + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x + n0.x, p0.y + n0.y}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x + n0.x + m0.x, p0.y + n0.y + m0.y}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x - n0.x + m0.x, p0.y - n0.y + m0.y}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x - n0.x, p0.y - n0.y}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + indices[0] = baseIndex; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; +} + +void mg_stroke_joint(mg_canvas_data* context, + vec2 p0, + vec2 t0, + vec2 t1, + mg_attributes* attributes) +{ + //NOTE(martin): compute the normals at the joint point + f32 norm_t0 = sqrt(Square(t0.x) + Square(t0.y)); + f32 norm_t1 = sqrt(Square(t1.x) + Square(t1.y)); + + vec2 n0 = {-t0.y, t0.x}; + n0.x /= norm_t0; + n0.y /= norm_t0; + + vec2 n1 = {-t1.y, t1.x}; + n1.x /= norm_t1; + n1.y /= norm_t1; + + //NOTE(martin): the sign of the cross product determines if the normals are facing outwards or inwards the angle. + // we flip them to face outwards if needed + f32 crossZ = n0.x*n1.y - n0.y*n1.x; + if(crossZ > 0) + { + n0.x *= -1; + n0.y *= -1; + n1.x *= -1; + n1.y *= -1; + } + + u32 zIndex = mg_get_next_z_index(context); + + //NOTE(martin): use the same code as hull offset to find mitter point... + /*NOTE(martin): let vector u = (n0+n1) and vector v = pIntersect - p1 + then v = u * (2*offset / norm(u)^2) + (this can be derived from writing the pythagoras theorems in the triangles of the joint) + */ + f32 halfW = 0.5 * attributes->width; + vec2 u = {n0.x + n1.x, n0.y + n1.y}; + f32 uNormSquare = u.x*u.x + u.y*u.y; + f32 alpha = attributes->width / uNormSquare; + vec2 v = {u.x * alpha, u.y * alpha}; + + f32 excursionSquare = uNormSquare * Square(alpha - attributes->width/4); + + if( attributes->joint == MG_JOINT_MITER + && excursionSquare <= Square(attributes->maxJointExcursion)) + { + vec2 mitterPoint = {p0.x + v.x, p0.y + v.y}; + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p0), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x + n0.x*halfW, p0.y + n0.y*halfW}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, mitterPoint), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x + n1.x*halfW, p0.y + n1.y*halfW}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + indices[0] = baseIndex; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; + } + else + { + //NOTE(martin): add a bevel joint + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 3); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, p0), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x + n0.x*halfW, p0.y + n0.y*halfW}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + mg_push_vertex(context, + mg_mat2x3_mul(context->transform, (vec2){p0.x + n1.x*halfW, p0.y + n1.y*halfW}), + (vec4){1, 1, 1, 1}, + attributes->color, + zIndex); + + DEBUG_ASSERT(!isnan(n0.x) && !isnan(n0.y) && !isnan(n1.x) && !isnan(n1.y)); + + indices[0] = baseIndex; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + } +} + +void mg_render_stroke_element(mg_canvas_data* context, + mg_path_elt* element, + mg_attributes* attributes, + vec2 currentPoint, + vec2* startTangent, + vec2* endTangent, + vec2* endPoint) +{ + vec2 controlPoints[4] = {currentPoint, element->p[0], element->p[1], element->p[2]}; + int endPointIndex = 0; + u32 zIndex = mg_get_next_z_index(context); + + switch(element->type) + { + case MG_PATH_LINE: + mg_render_stroke_line(context, controlPoints, zIndex, attributes); + endPointIndex = 1; + break; + + case MG_PATH_QUADRATIC: + mg_render_stroke_quadratic(context, controlPoints, zIndex, attributes); + endPointIndex = 2; + break; + + case MG_PATH_CUBIC: + mg_render_stroke_cubic(context, controlPoints, zIndex, attributes); + endPointIndex = 3; + break; + + case MG_PATH_MOVE: + ASSERT(0, "should be unreachable"); + break; + } + + *startTangent = (vec2){.x = controlPoints[1].x - controlPoints[0].x, + .y = controlPoints[1].y - controlPoints[0].y}; + + *endTangent = (vec2){controlPoints[endPointIndex].x - controlPoints[endPointIndex-1].x, + controlPoints[endPointIndex].y - controlPoints[endPointIndex-1].y}; + + *endPoint = controlPoints[endPointIndex]; + +} + +u32 mg_render_stroke_subpath(mg_canvas_data* context, + mg_path_elt* elements, + mg_path_descriptor* path, + mg_attributes* attributes, + u32 startIndex, + vec2 startPoint) +{ + u32 eltCount = path->count; + DEBUG_ASSERT(startIndex < eltCount); + + vec2 currentPoint = startPoint; + vec2 endPoint = {0, 0}; + vec2 previousEndTangent = {0, 0}; + vec2 firstTangent = {0, 0}; + vec2 startTangent = {0, 0}; + vec2 endTangent = {0, 0}; + + //NOTE(martin): render first element and compute first tangent + mg_render_stroke_element(context, elements + startIndex, attributes, currentPoint, &startTangent, &endTangent, &endPoint); + + firstTangent = startTangent; + previousEndTangent = endTangent; + currentPoint = endPoint; + + //NOTE(martin): render subsequent elements along with their joints + u32 eltIndex = startIndex + 1; + for(; + eltIndexjoint != MG_JOINT_NONE) + { + mg_stroke_joint(context, currentPoint, previousEndTangent, startTangent, attributes); + } + previousEndTangent = endTangent; + currentPoint = endPoint; + } + u32 subPathEltCount = eltIndex - (startIndex+1); + + //NOTE(martin): draw end cap / joint. We ensure there's at least two segments to draw a closing joint + if( subPathEltCount > 1 + && startPoint.x == endPoint.x + && startPoint.y == endPoint.y) + { + if(attributes->joint != MG_JOINT_NONE) + { + //NOTE(martin): add a closing joint if the path is closed + mg_stroke_joint(context, endPoint, endTangent, firstTangent, attributes); + } + } + else if(attributes->cap == MG_CAP_SQUARE) + { + //NOTE(martin): add start and end cap + mg_stroke_cap(context, startPoint, (vec2){-startTangent.x, -startTangent.y}, attributes); + mg_stroke_cap(context, endPoint, startTangent, attributes); + } + + return(eltIndex); +} + + +void mg_render_stroke(mg_canvas_data* context, + mg_path_elt* elements, + mg_path_descriptor* path, + mg_attributes* attributes) +{ + u32 eltCount = path->count; + DEBUG_ASSERT(eltCount); + + vec2 startPoint = path->startPoint; + u32 startIndex = 0; + + while(startIndex < eltCount) + { + //NOTE(martin): eliminate leading moves + while(startIndex < eltCount && elements[startIndex].type == MG_PATH_MOVE) + { + startPoint = elements[startIndex].p[0]; + startIndex++; + } + + if(startIndex < eltCount) + { + startIndex = mg_render_stroke_subpath(context, elements, path, attributes, startIndex, startPoint); + } + } +} + +//----------------------------------------------------------------------------------------------------------- +// Fast shapes primitives +//----------------------------------------------------------------------------------------------------------- + +void mg_render_rectangle_fill(mg_canvas_data* context, mp_rect rect, mg_attributes* attributes) +{ + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + u32 zIndex = mg_get_next_z_index(context); + + vec2 points[4] = {{rect.x, rect.y}, + {rect.x + rect.w, rect.y}, + {rect.x + rect.w, rect.y + rect.h}, + {rect.x, rect.y + rect.h}}; + + vec4 cubic = {1, 1, 1, 1}; + for(int i=0; i<4; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, points[i]); + mg_push_vertex(context, pos, cubic, attributes->color, zIndex); + } + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; +} + +void mg_render_rectangle_stroke(mg_canvas_data* context, mp_rect rect, mg_attributes* attributes) +{ + //NOTE(martin): stroke a rectangle by fill two scaled rectangles with the same zIndex. + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 12); + + u32 zIndex = mg_get_next_z_index(context); + + //NOTE(martin): limit stroke width to the minimum dimension of the rectangle + f32 width = minimum(attributes->width, minimum(rect.w, rect.h)); + f32 halfW = width/2; + + vec2 outerPoints[4] = {{rect.x - halfW, rect.y - halfW}, + {rect.x + rect.w + halfW, rect.y - halfW}, + {rect.x + rect.w + halfW, rect.y + rect.h + halfW}, + {rect.x - halfW, rect.y + rect.h + halfW}}; + + vec2 innerPoints[4] = {{rect.x + halfW, rect.y + halfW}, + {rect.x + rect.w - halfW, rect.y + halfW}, + {rect.x + rect.w - halfW, rect.y + rect.h - halfW}, + {rect.x + halfW, rect.y + rect.h - halfW}}; + + vec4 cubic = {1, 1, 1, 1}; + for(int i=0; i<4; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, outerPoints[i]); + mg_push_vertex(context, pos, cubic, attributes->color, zIndex); + } + + for(int i=0; i<4; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, innerPoints[i]); + mg_push_vertex(context, pos, cubic, attributes->color, zIndex); + } + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; + indices[6] = baseIndex + 4; + indices[7] = baseIndex + 5; + indices[8] = baseIndex + 6; + indices[9] = baseIndex + 4; + indices[10] = baseIndex + 6; + indices[11] = baseIndex + 7; +} + +void mg_render_fill_arc_corner(mg_canvas_data* context, f32 x, f32 y, f32 rx, f32 ry, u32 zIndex, mg_color color) +{ + //NOTE(martin): draw a precomputed arc corner, using a bezier approximation + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + static const vec4 cubics[4] = {{-3.76797, -9.76362, 5.47912, -1}, + {-4.19896, -9.45223, 7.534, -1}, + {-4.19896, -7.534, 9.45223, -1}, + {-3.76797, -5.47912, 9.76362, -1}}; + + f32 cx = rx*4*(sqrt(2)-1)/3; + f32 cy = ry*4*(sqrt(2)-1)/3; + + vec2 points[4] = {{x, y + ry}, + {x, y + ry - cy}, + {x + rx - cx, y}, + {x + rx, y}}; + + for(int i=0; i<4; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, points[i]); + mg_push_vertex(context, pos, cubics[i], color, zIndex); + } + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; +} + +void mg_render_rounded_rectangle_fill_with_z_index(mg_canvas_data* context, + mg_rounded_rect rect, + mg_attributes* attributes, + u32 zIndex) +{ + //NOTE(martin): draw a rounded rectangle by drawing a normal rectangle and 4 corners, + // approximating an arc by a precomputed bezier curve + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 18); + + //NOTE(martin): inner cutted corner rectangle + vec2 points[8] = {{rect.x + rect.r, rect.y}, + {rect.x + rect.w - rect.r, rect.y}, + {rect.x + rect.w, rect.y + rect.r}, + {rect.x + rect.w, rect.y + rect.h - rect.r}, + {rect.x + rect.w - rect.r, rect.y + rect.h}, + {rect.x + rect.r, rect.y + rect.h}, + {rect.x, rect.y + rect.h - rect.r}, + {rect.x, rect.y + rect.r}}; + + vec4 cubic = {1, 1, 1, 1}; + + for(int i=0; i<8; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, points[i]); + mg_push_vertex(context, pos, cubic, attributes->color, zIndex); + } + + static const i32 fanIndices[18] = { 0, 1, 2, 0, 2, 3, 0, 3, 4, 0, 4, 5, 0, 5, 6, 0, 6, 7 }; // inner fan + for(int i=0; i<18; i++) + { + indices[i] = fanIndices[i] + baseIndex; + } + + mg_render_fill_arc_corner(context, rect.x, rect.y, rect.r, rect.r, zIndex, attributes->color); + mg_render_fill_arc_corner(context, rect.x + rect.w, rect.y, -rect.r, rect.r, zIndex, attributes->color); + mg_render_fill_arc_corner(context, rect.x + rect.w, rect.y + rect.h, -rect.r, -rect.r, zIndex, attributes->color); + mg_render_fill_arc_corner(context, rect.x, rect.y + rect.h, rect.r, -rect.r, zIndex, attributes->color); +} + + +void mg_render_rounded_rectangle_fill(mg_canvas_data* context, + mg_rounded_rect rect, + mg_attributes* attributes) +{ + u32 zIndex = mg_get_next_z_index(context); + mg_render_rounded_rectangle_fill_with_z_index(context, rect, attributes, zIndex); +} + +void mg_render_rounded_rectangle_stroke(mg_canvas_data* context, + mg_rounded_rect rect, + mg_attributes* attributes) +{ + //NOTE(martin): stroke rounded rectangle by filling two scaled rounded rectangles with the same zIndex + f32 width = minimum(attributes->width, minimum(rect.w, rect.h)); + f32 halfW = width/2; + + mg_rounded_rect inner = {rect.x + halfW, rect.y + halfW, rect.w - width, rect.h - width, rect.r - halfW}; + mg_rounded_rect outer = {rect.x - halfW, rect.y - halfW, rect.w + width, rect.h + width, rect.r + halfW}; + + u32 zIndex = mg_get_next_z_index(context); + mg_render_rounded_rectangle_fill_with_z_index(context, outer, attributes, zIndex); + mg_render_rounded_rectangle_fill_with_z_index(context, inner, attributes, zIndex); +} + +void mg_render_ellipse_fill_with_z_index(mg_canvas_data* context, + mp_rect rect, + mg_attributes* attributes, + u32 zIndex) +{ + //NOTE(martin): draw a filled ellipse by drawing a diamond and 4 corners, + // approximating an arc by a precomputed bezier curve + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + f32 rx = rect.w/2; + f32 ry = rect.h/2; + + //NOTE(martin): inner diamond + vec2 points[4] = {{rect.x, rect.y + ry}, + {rect.x + rx, rect.y}, + {rect.x + rect.w, rect.y + ry}, + {rect.x + rx, rect.y + rect.h}}; + + vec4 cubic = {1, 1, 1, 1}; + + for(int i=0; i<4; i++) + { + vec2 pos = mg_mat2x3_mul(context->transform, points[i]); + mg_push_vertex(context, pos, cubic, attributes->color, zIndex); + } + + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; + + mg_render_fill_arc_corner(context, rect.x, rect.y, rx, ry, zIndex, attributes->color); + mg_render_fill_arc_corner(context, rect.x + rect.w, rect.y, -rx, ry, zIndex, attributes->color); + mg_render_fill_arc_corner(context, rect.x + rect.w, rect.y + rect.h, -rx, -ry, zIndex, attributes->color); + mg_render_fill_arc_corner(context, rect.x, rect.y + rect.h, rx, -ry, zIndex, attributes->color); +} + +void mg_render_ellipse_fill(mg_canvas_data* context, mp_rect rect, mg_attributes* attributes) +{ + u32 zIndex = mg_get_next_z_index(context); + mg_render_ellipse_fill_with_z_index(context, rect, attributes, zIndex); +} + +void mg_render_ellipse_stroke(mg_canvas_data* context, mp_rect rect, mg_attributes* attributes) +{ + //NOTE(martin): stroke by filling two scaled ellipsis with the same zIndex + f32 width = minimum(attributes->width, minimum(rect.w, rect.h)); + f32 halfW = width/2; + + mp_rect inner = {rect.x + halfW, rect.y + halfW, rect.w - width, rect.h - width}; + mp_rect outer = {rect.x - halfW, rect.y - halfW, rect.w + width, rect.h + width}; + + u32 zIndex = mg_get_next_z_index(context); + mg_render_ellipse_fill_with_z_index(context, outer, attributes, zIndex); + mg_render_ellipse_fill_with_z_index(context, inner, attributes, zIndex); +} + +void mg_render_image(mg_canvas_data* context, mg_image image, mp_rect rect) +{ + mg_image_data* imageData = mg_image_ptr_from_handle(context, image); + if(!imageData) + { + return; + } + + u32 baseIndex = mg_vertices_base_index(context); + i32* indices = mg_reserve_indices(context, 6); + + u32 zIndex = mg_get_next_z_index(context); + + vec2 points[4] = {{rect.x, rect.y}, + {rect.x + rect.w, rect.y}, + {rect.x + rect.w, rect.y + rect.h}, + {rect.x, rect.y + rect.h}}; + + vec2 uv[4] = {{imageData->rect.x + 0.5, imageData->rect.y + 0.5}, + {imageData->rect.x + imageData->rect.w - 0.5, imageData->rect.y + 0.5}, + {imageData->rect.x + imageData->rect.w - 0.5, imageData->rect.y + imageData->rect.h - 0.5}, + {imageData->rect.x + 0.5, imageData->rect.y + imageData->rect.h - 0.5}}; + + vec4 cubic = {1, 1, 1, 1}; + mg_color color = {1, 1, 1, 1}; + + for(int i=0; i<4; i++) + { + vec2 transformedUV = {uv[i].x / MG_ATLAS_SIZE, uv[i].y / MG_ATLAS_SIZE}; + + vec2 pos = mg_mat2x3_mul(context->transform, points[i]); + mg_push_textured_vertex(context, pos, cubic, transformedUV, color, zIndex); + } + indices[0] = baseIndex + 0; + indices[1] = baseIndex + 1; + indices[2] = baseIndex + 2; + indices[3] = baseIndex + 0; + indices[4] = baseIndex + 2; + indices[5] = baseIndex + 3; +} + +void mg_render_rounded_image(mg_canvas_data* context, mg_image image, mg_rounded_rect rect, mg_attributes* attributes) +{ + mg_image_data* imageData = mg_image_ptr_from_handle(context, image); + if(!imageData) + { + return; + } + + //////////////////////////////////////////////////////////////////////////////// + //TODO: this does not work for rotated rectangles + //////////////////////////////////////////////////////////////////////////////// + vec2 uvMin = {(imageData->rect.x + 0.5) / MG_ATLAS_SIZE, (imageData->rect.y + 0.5) / MG_ATLAS_SIZE}; + vec2 uvMax = {(imageData->rect.x + imageData->rect.w - 0.5) / MG_ATLAS_SIZE, (imageData->rect.y + imageData->rect.h - 0.5) / MG_ATLAS_SIZE}; + mp_rect uvRect = {uvMin.x, uvMin.y, uvMax.x - uvMin.x, uvMax.y - uvMin.y}; + + vec2 pMin = mg_mat2x3_mul(context->transform, (vec2){rect.x, rect.y}); + vec2 pMax = mg_mat2x3_mul(context->transform, (vec2){rect.x + rect.w, rect.y + rect.h}); + mp_rect pRect = {pMin.x, pMin.y, pMax.x - pMin.x, pMax.y - pMin.y}; + + u32 startIndex = mg_vertices_base_index(context); + + mgc_vertex_layout* layout = &context->painter->vertexLayout; + + attributes->color = (mg_color){1, 1, 1, 1}; + mg_render_rounded_rectangle_fill(context, rect, attributes); + + u32 indexCount = mg_vertices_base_index(context) - startIndex; + + for(int i=0; iposBuffer) + index*layout->posStride); + vec2* uv = (vec2*)(((char*)layout->uvBuffer) + index*layout->uvStride); + + vec2 coordInBoundingSpace = {(pos->x - pRect.x)/pRect.w, + (pos->y - pRect.y)/pRect.h}; + + vec2 mappedUV = {uvRect.x + coordInBoundingSpace.x * uvRect.w, + uvRect.y + coordInBoundingSpace.y * uvRect.h}; + + *uv = mappedUV; + } + +} + +//----------------------------------------------------------------------------------------------------------- +// Graphics canvas +//----------------------------------------------------------------------------------------------------------- + +#ifdef MG_IMPLEMENTS_BACKEND_METAL + typedef struct mg_metal_surface mg_metal_surface; + mg_canvas_painter* mg_metal_painter_create_for_surface(mg_metal_surface* surface, mp_rect viewPort); +#endif + +mg_canvas_painter* mg_canvas_painter_create_for_surface(mg_surface surfaceHandle, mp_rect viewPort) +{ + mg_surface_info* surface = mg_surface_ptr_from_handle(surfaceHandle); + if(!surface) + { + return(0); + } + + switch(surface->backend) + { + case MG_BACKEND_METAL: + return(mg_metal_painter_create_for_surface((mg_metal_surface*)surface, viewPort)); + + default: + return(0); + } +} + +void mg_canvas_reset_streams(mg_canvas_data* context) +{ + ListInit(&context->streamsFreeList); + for(int i=0; istreamsFreeList, &context->streams[i].freeListElt); + context->streams[i].generation = 1; + } + context->currentStream = mg_stream_alloc(context); +} + +mg_canvas mg_canvas_create(mg_surface surface, mp_rect viewPort) +{ + mg_canvas_data* context = mg_canvas_alloc(); + if(!context) + { + return(mg_canvas_nil()); + } + + context->path.startIndex = 0; + context->path.count = 0; + context->nextZIndex = 1; + context->attributes.color = (mg_color){0, 0, 0, 1}; + context->attributes.tolerance = 1; + context->attributes.width = 10; + context->attributes.clip = (mp_rect){-FLT_MAX/2, -FLT_MAX/2, FLT_MAX, FLT_MAX}; + + context->transform = (mg_mat2x3){{1, 0, 0, + 0, 1, 0}}; + + mg_canvas_reset_streams(context); + + context->painter = mg_canvas_painter_create_for_surface(surface, viewPort); + if(!context->painter) + { + LOG_ERROR("Couldn't create painter for surface\n"); + mg_canvas_recycle(context); + return(mg_canvas_nil()); + } + + mg_canvas handle = mg_canvas_handle_from_ptr(context); + + //NOTE: create a blank image + u8 bytes[4] = {255, 255, 255, 255}; + context->blankImage = mg_image_create_from_rgba8(handle, 1, 1, bytes); + + return(handle); +} + +void mg_canvas_destroy(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_canvas_recycle(context); +} + +void mg_canvas_viewport(mg_canvas handle, mp_rect viewport) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->painter->setViewPort(context->painter, viewport); +} + +void mg_canvas_begin_draw(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->vertexCount = 0; + context->indexCount = 0; + context->path.startIndex = 0; + context->path.count = 0; +} + +mg_mat2x3 mg_matrix_stack_top(mg_canvas_data* context) +{ + if(context->matrixStackSize == 0) + { + return((mg_mat2x3){1, 0, 0, + 0, 1, 0}); + } + else + { + return(context->matrixStack[context->matrixStackSize-1]); + } +} + +void mg_matrix_stack_push(mg_canvas_data* context, mg_mat2x3 transform) +{ + if(context->matrixStackSize >= MG_MATRIX_STACK_MAX_DEPTH) + { + LOG_ERROR("matrix stack overflow\n"); + } + else + { + context->matrixStack[context->matrixStackSize] = transform; + context->matrixStackSize++; + context->transform = transform; + } +} + +void mg_matrix_stack_pop(mg_canvas_data* context) +{ + if(context->matrixStackSize == 0) + { + LOG_ERROR("matrix stack underflow\n"); + } + else + { + context->matrixStackSize--; + context->transform = mg_matrix_stack_top(context); + } +} + + +mp_rect mg_clip_stack_top(mg_canvas_data* context) +{ + if(context->clipStackSize == 0) + { + return((mp_rect){-FLT_MAX/2, -FLT_MAX/2, FLT_MAX, FLT_MAX}); + } + else + { + return(context->clipStack[context->clipStackSize-1]); + } +} + +void mg_clip_stack_push(mg_canvas_data* context, mp_rect clip) +{ + if(context->clipStackSize >= MG_CLIP_STACK_MAX_DEPTH) + { + LOG_ERROR("clip stack overflow\n"); + } + else + { + context->clipStack[context->clipStackSize] = clip; + context->clipStackSize++; + context->clip = clip; + } +} + +void mg_clip_stack_pop(mg_canvas_data* context) +{ + if(context->clipStackSize == 0) + { + LOG_ERROR("clip stack underflow\n"); + } + else + { + context->clipStackSize--; + context->clip = mg_clip_stack_top(context); + } +} + +void mg_do_clip_push(mg_canvas_data* context, mp_rect clip) +{ + //NOTE(martin): transform clip + vec2 p0 = mg_mat2x3_mul(context->transform, (vec2){clip.x, clip.y}); + vec2 p1 = mg_mat2x3_mul(context->transform, (vec2){clip.x + clip.w, clip.y}); + vec2 p2 = mg_mat2x3_mul(context->transform, (vec2){clip.x + clip.w, clip.y + clip.h}); + vec2 p3 = mg_mat2x3_mul(context->transform, (vec2){clip.x, clip.y + clip.h}); + + f32 x0 = minimum(p0.x, minimum(p1.x, minimum(p2.x, p3.x))); + f32 y0 = minimum(p0.y, minimum(p1.y, minimum(p2.y, p3.y))); + f32 x1 = maximum(p0.x, maximum(p1.x, maximum(p2.x, p3.x))); + f32 y1 = maximum(p0.y, maximum(p1.y, maximum(p2.y, p3.y))); + + mp_rect current = mg_clip_stack_top(context); + + //NOTE(martin): intersect with current clip + x0 = maximum(current.x, x0); + y0 = maximum(current.y, y0); + x1 = minimum(current.x + current.w, x1); + y1 = minimum(current.y + current.h, y1); + + mp_rect r = {x0, y0, maximum(0, x1-x0), maximum(0, y1-y0)}; + mg_clip_stack_push(context, r); +} + +void mg_canvas_flush(mg_canvas contextHandle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(contextHandle); + if(!context) + { + return; + } + + mg_color clearColor = {0, 0, 0, 1}; + + u32 count = context->primitiveCount; + + mg_stream_data* stream = context->currentStream; + DEBUG_ASSERT(stream); + DEBUG_ASSERT(stream->count <= count); + + u32 nextIndex = stream->firstCommand; + + mg_reset_z_index(context); + context->transform = (mg_mat2x3){1, 0, 0, + 0, 1, 0}; + context->clip = (mp_rect){-FLT_MAX/2, -FLT_MAX/2, FLT_MAX, FLT_MAX}; + + for(int i=0; icount; i++) + { + if(nextIndex >= count) + { + LOG_ERROR("invalid location '%i' in graphics command buffer would cause an overrun\n", nextIndex); + break; + } + mg_primitive* primitive = &(context->primitives[nextIndex]); + nextIndex++; + + switch(primitive->cmd) + { + case MG_CMD_CLEAR: + { + //NOTE(martin): clear buffers + context->vertexCount = 0; + context->indexCount = 0; + + clearColor = primitive->attributes.color; + } break; + + case MG_CMD_FILL: + { + u32 zIndex = mg_get_next_z_index(context); + mg_render_fill(context, + context->pathElements + primitive->path.startIndex, + &primitive->path, + zIndex, + primitive->attributes.color); + } break; + + case MG_CMD_STROKE: + { + mg_render_stroke(context, + context->pathElements + primitive->path.startIndex, + &primitive->path, + &primitive->attributes); + } break; + + + case MG_CMD_RECT_FILL: + mg_render_rectangle_fill(context, primitive->rect, &primitive->attributes); + break; + + case MG_CMD_RECT_STROKE: + mg_render_rectangle_stroke(context, primitive->rect, &primitive->attributes); + break; + + case MG_CMD_ROUND_RECT_FILL: + mg_render_rounded_rectangle_fill(context, primitive->roundedRect, &primitive->attributes); + break; + + case MG_CMD_ROUND_RECT_STROKE: + mg_render_rounded_rectangle_stroke(context, primitive->roundedRect, &primitive->attributes); + break; + + case MG_CMD_ELLIPSE_FILL: + mg_render_ellipse_fill(context, primitive->rect, &primitive->attributes); + break; + + case MG_CMD_ELLIPSE_STROKE: + mg_render_ellipse_stroke(context, primitive->rect, &primitive->attributes); + break; + + case MG_CMD_JUMP: + { + if(primitive->jump == ~0) + { + //NOTE(martin): normal end of stream marker + goto exit_command_loop; + } + else if(primitive->jump >= count) + { + LOG_ERROR("invalid jump location '%i' in graphics command buffer\n", primitive->jump); + goto exit_command_loop; + } + else + { + nextIndex = primitive->jump; + } + } break; + + case MG_CMD_MATRIX_PUSH: + { + mg_mat2x3 transform = mg_matrix_stack_top(context); + mg_matrix_stack_push(context, mg_mat2x3_mul_m(transform, primitive->matrix)); + } break; + + case MG_CMD_MATRIX_POP: + { + mg_matrix_stack_pop(context); + } break; + + case MG_CMD_CLIP_PUSH: + { + //TODO(martin): use only aligned rect and avoid this + mp_rect r = {primitive->rect.x, primitive->rect.y, primitive->rect.w, primitive->rect.h}; + mg_do_clip_push(context, r); + } break; + + case MG_CMD_CLIP_POP: + { + mg_clip_stack_pop(context); + } break; + + case MG_CMD_IMAGE_DRAW: + { + mg_render_image(context, primitive->attributes.image, primitive->rect); + } break; + + case MG_CMD_ROUNDED_IMAGE_DRAW: + { + mg_render_rounded_image(context, primitive->attributes.image, primitive->roundedRect, &primitive->attributes); + } break; + + } + } + exit_command_loop: ; + + if(context->painter && context->painter->drawBuffers) + { + context->painter->drawBuffers(context->painter, context->vertexCount, context->indexCount, clearColor); + } + + //NOTE(martin): clear buffers + context->primitiveCount = 0; + context->vertexCount = 0; + context->indexCount = 0; + + context->path.startIndex = 0; + context->path.count = 0; + + //NOTE(martin): invalidate all streams + context->frameCounter++; + mg_canvas_reset_streams(context); +} + +void mg_push_command(mg_canvas_data* context, mg_primitive primitive) +{ + //NOTE(martin): push primitive and updates current stream, eventually patching a pending jump. + ASSERT(context->primitiveCount < MG_MAX_PRIMITIVE_COUNT); + context->primitives[context->primitiveCount] = primitive; + context->primitiveCount++; + + mg_stream_data* stream = context->currentStream; + DEBUG_ASSERT(stream); + + u32 lastCommand = context->primitiveCount-1; + if(stream->pendingJump) + { + DEBUG_ASSERT(context->primitives[stream->lastCommand].cmd == MG_CMD_JUMP); + context->primitives[stream->lastCommand].jump = lastCommand; + stream->pendingJump = false; + } + stream->lastCommand = lastCommand; + stream->count++; +} + +//----------------------------------------------------------------------------------------------------------- +// Graphics command streams +//----------------------------------------------------------------------------------------------------------- + +mg_stream mg_stream_create(mg_canvas contextHandle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(contextHandle); + if(!context) + { + return(mg_stream_null_handle()); + } + + mg_stream_data* stream = mg_stream_alloc(context); + if(!stream) + { + return(mg_stream_null_handle()); + } + + return(mg_stream_handle_from_ptr(context, stream)); +} + +mg_stream mg_stream_swap(mg_canvas contextHandle, mg_stream streamHandle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(contextHandle); + if(!context) + { + return(mg_stream_null_handle()); + } + mg_stream_data* stream = mg_stream_ptr_from_handle(context, streamHandle); + if(!stream) + { + return(mg_stream_null_handle()); + } + + if(stream == context->currentStream) + { + //NOTE(martin): prevent swapping for itself + return(mg_stream_null_handle()); + } + + //NOTE(martin): push an unitialized jump to the current stream + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_JUMP, .jump = ~0}); + context->currentStream->pendingJump = true; + + //NOTE(martin): set the new current stream + mg_stream_data* oldStream = context->currentStream; + context->currentStream = stream; + + if(!stream->count) + { + //NOTE(martin): if stream is new, set the first command to the current offset in command buffer + stream->firstCommand = context->primitiveCount; + stream->lastCommand = stream->firstCommand; + } + else + { + //NOTE(martin): else, we just make sure there is a pending jump to be patched when we push the next command + DEBUG_ASSERT(stream->pendingJump); + } + + return(mg_stream_handle_from_ptr(context, oldStream)); +} + +void mg_stream_append(mg_canvas contextHandle, mg_stream streamHandle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(contextHandle); + if(!context) + { + return; + } + mg_stream_data* stream = mg_stream_ptr_from_handle(context, streamHandle); + if(!stream) + { + return; + } + if(stream == context->currentStream) + { + //NOTE(martin): prevent appending to itself + return; + } + + if(stream->count) + { + //NOTE(martin): push a jump set to the appended stream's first command + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_JUMP, .jump = stream->firstCommand}); + + //NOTE(martin): update current stream last command and make sure it's a jump + context->currentStream->lastCommand = stream->lastCommand; + context->currentStream->count += stream->count; + context->currentStream->pendingJump = true; + + DEBUG_ASSERT(context->primitives[context->currentStream->lastCommand].cmd == MG_CMD_JUMP); + } + + //NOTE(martin): invalidate appended stream + mg_stream_recycle(context, stream); +} + +//----------------------------------------------------------------------------------------------------------- +// Fonts management +//----------------------------------------------------------------------------------------------------------- + +mg_font mg_font_nil() +{ + return((mg_font){0}); +} + +mg_font mg_font_create_from_memory(u32 size, byte* buffer, u32 rangeCount, unicode_range* ranges) +{ + mg_font_info* fontInfo = ListPopEntry(&__mgInfo.fontFreeList, mg_font_info, freeListElt); + if(!fontInfo) + { + if(__mgInfo.fontsNextIndex < MG_FONT_MAX_COUNT) + { + fontInfo = &(__mgInfo.fonts[__mgInfo.fontsNextIndex]); + __mgInfo.fontsNextIndex++; + memset(fontInfo, 0, sizeof(*fontInfo)); + fontInfo->generation = 1; + } + else + { + LOG_ERROR("can't allocate new font\n"); + return((mg_font){0}); + } + } + + stbtt_fontinfo stbttFontInfo; + stbtt_InitFont(&stbttFontInfo, buffer, 0); + + //NOTE(martin): load font metrics data + fontInfo->unitsPerEm = 1./stbtt_ScaleForMappingEmToPixels(&stbttFontInfo, 1); + + int ascent, descent, lineGap, x0, x1, y0, y1; + stbtt_GetFontVMetrics(&stbttFontInfo, &ascent, &descent, &lineGap); + stbtt_GetFontBoundingBox(&stbttFontInfo, &x0, &y0, &x1, &y1); + + fontInfo->extents.ascent = ascent; + fontInfo->extents.descent = -descent; + fontInfo->extents.leading = lineGap; + fontInfo->extents.width = x1 - x0; + + stbtt_GetCodepointBox(&stbttFontInfo, 'x', &x0, &y0, &x1, &y1); + fontInfo->extents.xHeight = y1 - y0; + + stbtt_GetCodepointBox(&stbttFontInfo, 'M', &x0, &y0, &x1, &y1); + fontInfo->extents.capHeight = y1 - y0; + + //NOTE(martin): load codepoint ranges + fontInfo->rangeCount = rangeCount; + fontInfo->glyphMap = malloc_array(mg_glyph_map_entry, rangeCount); + fontInfo->glyphCount = 0; + + for(int i=0; iglyphMap[i].range = ranges[i]; + fontInfo->glyphMap[i].firstGlyphIndex = fontInfo->glyphCount + 1; + fontInfo->glyphCount += ranges[i].count; + } + + fontInfo->glyphs = malloc_array(mg_glyph_info, fontInfo->glyphCount); + + //NOTE(martin): first do a count of outlines + int outlineCount = 0; + for(int rangeIndex=0; rangeIndexglyphMap[rangeIndex].range.firstCodePoint; + u32 firstGlyphIndex = fontInfo->glyphMap[rangeIndex].firstGlyphIndex; + u32 endGlyphIndex = firstGlyphIndex + fontInfo->glyphMap[rangeIndex].range.count; + + for(int glyphIndex = firstGlyphIndex; + glyphIndex < endGlyphIndex; glyphIndex++) + { + int stbttGlyphIndex = stbtt_FindGlyphIndex(&stbttFontInfo, codePoint); + if(stbttGlyphIndex == 0) + { + //NOTE(martin): the codepoint is not found in the font + codePoint++; + continue; + } + //NOTE(martin): load glyph outlines + stbtt_vertex* vertices = 0; + outlineCount += stbtt_GetGlyphShape(&stbttFontInfo, stbttGlyphIndex, &vertices); + stbtt_FreeShape(&stbttFontInfo, vertices); + codePoint++; + } + } + //NOTE(martin): allocate outlines + fontInfo->outlines = malloc_array(mg_path_elt, outlineCount); + fontInfo->outlineCount = 0; + + //NOTE(martin): load metrics and outlines + for(int rangeIndex=0; rangeIndexglyphMap[rangeIndex].range.firstCodePoint; + u32 firstGlyphIndex = fontInfo->glyphMap[rangeIndex].firstGlyphIndex; + u32 endGlyphIndex = firstGlyphIndex + fontInfo->glyphMap[rangeIndex].range.count; + + for(int glyphIndex = firstGlyphIndex; + glyphIndex < endGlyphIndex; glyphIndex++) + { + mg_glyph_info* glyph = &(fontInfo->glyphs[glyphIndex-1]); + + int stbttGlyphIndex = stbtt_FindGlyphIndex(&stbttFontInfo, codePoint); + if(stbttGlyphIndex == 0) + { + //NOTE(martin): the codepoint is not found in the font, we zero the glyph info + memset(glyph, 0, sizeof(*glyph)); + codePoint++; + continue; + } + + glyph->exists = true; + glyph->codePoint = codePoint; + + //NOTE(martin): load glyph metric + int xAdvance, xBearing, x0, y0, x1, y1; + stbtt_GetGlyphHMetrics(&stbttFontInfo, stbttGlyphIndex, &xAdvance, &xBearing); + stbtt_GetGlyphBox(&stbttFontInfo, stbttGlyphIndex, &x0, &y0, &x1, &y1); + + glyph->extents.xAdvance = (f32)xAdvance; + glyph->extents.yAdvance = 0; + glyph->extents.xBearing = (f32)xBearing; + glyph->extents.yBearing = y0; + + glyph->extents.width = x1 - x0; + glyph->extents.height = y1 - y0; + + //NOTE(martin): load glyph outlines + + stbtt_vertex* vertices = 0; + int vertexCount = stbtt_GetGlyphShape(&stbttFontInfo, stbttGlyphIndex, &vertices); + + glyph->pathDescriptor = (mg_path_descriptor){.startIndex = fontInfo->outlineCount, + .count = vertexCount, + .startPoint = {0, 0}}; + + mg_path_elt* elements = fontInfo->outlines + fontInfo->outlineCount; + fontInfo->outlineCount += vertexCount; + vec2 currentPos = {0, 0}; + + for(int vertIndex = 0; vertIndex < vertexCount; vertIndex++) + { + f32 x = vertices[vertIndex].x; + f32 y = vertices[vertIndex].y; + f32 cx = vertices[vertIndex].cx; + f32 cy = vertices[vertIndex].cy; + f32 cx1 = vertices[vertIndex].cx1; + f32 cy1 = vertices[vertIndex].cy1; + + switch(vertices[vertIndex].type) + { + case STBTT_vmove: + elements[vertIndex].type = MG_PATH_MOVE; + elements[vertIndex].p[0] = (vec2){x, y}; + break; + + case STBTT_vline: + elements[vertIndex].type = MG_PATH_LINE; + elements[vertIndex].p[0] = (vec2){x, y}; + break; + + case STBTT_vcurve: + { + elements[vertIndex].type = MG_PATH_QUADRATIC; + elements[vertIndex].p[0] = (vec2){cx, cy}; + elements[vertIndex].p[1] = (vec2){x, y}; + } break; + + case STBTT_vcubic: + elements[vertIndex].type = MG_PATH_CUBIC; + elements[vertIndex].p[0] = (vec2){cx, cy}; + elements[vertIndex].p[1] = (vec2){cx1, cy1}; + elements[vertIndex].p[2] = (vec2){x, y}; + break; + } + currentPos = (vec2){x, y}; + } + stbtt_FreeShape(&stbttFontInfo, vertices); + codePoint++; + } + } + u64 h = ((u64)(fontInfo - __mgInfo.fonts))<<32 | ((u64)fontInfo->generation); + return((mg_font){h}); +} + +mg_font_info* mg_get_font_info(mg_font fontHandle) +{ + u32 fontIndex = (u32)(fontHandle.h >> 32); + u32 fontGeneration = (u32)(fontHandle.h & 0xffffffff); + + if( fontIndex >= MG_FONT_MAX_COUNT + || fontIndex >= __mgInfo.fontsNextIndex + || fontGeneration != __mgInfo.fonts[fontIndex].generation) + { + LOG_WARNING("invalid font handle\n"); + return(0); + } + return(&(__mgInfo.fonts[fontIndex])); +} + +void mg_font_destroy(mg_font fontHandle) +{ + mg_font_info* fontInfo = mg_get_font_info(fontHandle); + if(!fontInfo) + { + return; + } + #ifdef DEBUG + if(fontInfo->generation == UINT32_MAX) + { + LOG_ERROR("font info generation wrap around\n"); + } + #endif + fontInfo->generation++; + + free(fontInfo->glyphMap); + free(fontInfo->glyphs); + free(fontInfo->outlines); + + ListPush(&__mgInfo.fontFreeList, &fontInfo->freeListElt); +} + +str32 mg_font_get_glyph_indices_from_font_info(mg_font_info* fontInfo, str32 codePoints, str32 backing) +{ + u64 count = minimum(codePoints.len, backing.len); + + for(int i = 0; irangeCount; rangeIndex++) + { + if(codePoints.ptr[i] >= fontInfo->glyphMap[rangeIndex].range.firstCodePoint + && codePoints.ptr[i] < (fontInfo->glyphMap[rangeIndex].range.firstCodePoint + fontInfo->glyphMap[rangeIndex].range.count)) + { + u32 rangeOffset = codePoints.ptr[i] - fontInfo->glyphMap[rangeIndex].range.firstCodePoint; + glyphIndex = fontInfo->glyphMap[rangeIndex].firstGlyphIndex + rangeOffset; + break; + } + } + if(glyphIndex && !fontInfo->glyphs[glyphIndex].exists) + { + backing.ptr[i] = 0; + } + backing.ptr[i] = glyphIndex; + } + str32 res = {.len = count, .ptr = backing.ptr}; + return(res); +} + +u32 mg_font_get_glyph_index_from_font_info(mg_font_info* fontInfo, utf32 codePoint) +{ + u32 glyphIndex = 0; + str32 codePoints = {1, &codePoint}; + str32 backing = {1, &glyphIndex}; + mg_font_get_glyph_indices_from_font_info(fontInfo, codePoints, backing); + return(glyphIndex); +} + +str32 mg_font_get_glyph_indices(mg_font font, str32 codePoints, str32 backing) +{ + mg_font_info* fontInfo = mg_get_font_info(font); + if(!fontInfo) + { + return((str32){}); + } + return(mg_font_get_glyph_indices_from_font_info(fontInfo, codePoints, backing)); +} + +str32 mg_font_push_glyph_indices(mg_font font, mem_arena* arena, str32 codePoints) +{ + u32* buffer = mem_arena_alloc_array(arena, u32, codePoints.len); + str32 backing = {codePoints.len, buffer}; + return(mg_font_get_glyph_indices(font, codePoints, backing)); +} + +u32 mg_font_get_glyph_index(mg_font font, utf32 codePoint) +{ + u32 glyphIndex = 0; + str32 codePoints = {1, &codePoint}; + str32 backing = {1, &glyphIndex}; + mg_font_get_glyph_indices(font, codePoints, backing); + return(glyphIndex); +} + +mg_glyph_info* mg_font_get_glyph_info(mg_font_info* fontInfo, u32 glyphIndex) +{ + DEBUG_ASSERT(glyphIndex); + DEBUG_ASSERT(glyphIndex < fontInfo->glyphCount); + return(&(fontInfo->glyphs[glyphIndex-1])); +} + +int mg_font_get_extents(mg_font font, mg_font_extents* outExtents) +{ + mg_font_info* fontInfo = mg_get_font_info(font); + if(!fontInfo) + { + return(-1); + } + *outExtents = fontInfo->extents; + return(0); +} + +f32 mg_font_get_scale_for_em_pixels(mg_font font, f32 emSize) +{ + mg_font_info* fontInfo = mg_get_font_info(font); + if(!fontInfo) + { + return(0); + } + return(emSize/fontInfo->unitsPerEm); +} + +void mg_font_get_glyph_extents_from_font_info(mg_font_info* fontInfo, + str32 glyphIndices, + mg_text_extents* outExtents) +{ + for(int i=0; i= fontInfo->glyphCount) + { + continue; + } + mg_glyph_info* glyph = mg_font_get_glyph_info(fontInfo, glyphIndices.ptr[i]); + outExtents[i] = glyph->extents; + } +} + +int mg_font_get_glyph_extents(mg_font font, str32 glyphIndices, mg_text_extents* outExtents) +{ + mg_font_info* fontInfo = mg_get_font_info(font); + if(!fontInfo) + { + return(-1); + } + mg_font_get_glyph_extents_from_font_info(fontInfo, glyphIndices, outExtents); + return(0); +} + +int mg_font_get_codepoint_extents(mg_font font, utf32 codePoint, mg_text_extents* outExtents) +{ + mg_font_info* fontInfo = mg_get_font_info(font); + if(!fontInfo) + { + return(-1); + } + u32 glyphIndex = 0; + str32 codePoints = {1, &codePoint}; + str32 backing = {1, &glyphIndex}; + str32 glyphs = mg_font_get_glyph_indices_from_font_info(fontInfo, codePoints, backing); + mg_font_get_glyph_extents_from_font_info(fontInfo, glyphs, outExtents); + return(0); +} + +mp_rect mg_text_bounding_box(mg_font font, f32 fontSize, str8 text) +{ + if(!text.len || !text.ptr) + { + return((mp_rect){}); + } + + mg_font_info* fontInfo = mg_get_font_info(font); + if(!fontInfo) + { + return((mp_rect){}); + } + + mem_arena* scratch = mem_scratch(); + str32 codePoints = utf8_push_to_codepoints(scratch, text); + str32 glyphIndices = mg_font_push_glyph_indices(font, scratch, codePoints); + + //NOTE(martin): find width of missing character + //TODO(martin): should cache that at font creation... + mg_text_extents missingGlyphExtents; + u32 missingGlyphIndex = mg_font_get_glyph_index_from_font_info(fontInfo, 0xfffd); + + if(missingGlyphIndex) + { + mg_font_get_glyph_extents_from_font_info(fontInfo, (str32){1, &missingGlyphIndex}, &missingGlyphExtents); + } + else + { + //NOTE(martin): could not find replacement glyph, try to get an 'x' to get a somewhat correct width + // to render an empty rectangle. Otherwise just render with the max font width + f32 boxWidth = fontInfo->extents.width * 0.8; + f32 xBearing = fontInfo->extents.width * 0.1; + f32 xAdvance = fontInfo->extents.width; + + missingGlyphIndex = mg_font_get_glyph_index_from_font_info(fontInfo, 'x'); + if(missingGlyphIndex) + { + mg_font_get_glyph_extents_from_font_info(fontInfo, (str32){1, &missingGlyphIndex}, &missingGlyphExtents); + } + else + { + missingGlyphExtents.xBearing = fontInfo->extents.width * 0.1; + missingGlyphExtents.yBearing = 0; + missingGlyphExtents.width = fontInfo->extents.width * 0.8; + missingGlyphExtents.xAdvance = fontInfo->extents.width; + missingGlyphExtents.yAdvance = 0; + } + } + + //NOTE(martin): accumulate text extents + f32 width = 0; + f32 x = 0; + f32 y = 0; + f32 lineHeight = fontInfo->extents.descent + fontInfo->extents.ascent; + + for(int i=0; i= fontInfo->glyphCount) + { + extents = missingGlyphExtents; + } + else + { + glyph = mg_font_get_glyph_info(fontInfo, glyphIndices.ptr[i]); + extents = glyph->extents; + } + x += extents.xAdvance; + y += extents.yAdvance; + + if(glyph && glyph->codePoint == '\n') + { + width = maximum(width, x); + x = 0; + y += lineHeight + fontInfo->extents.leading; + } + } + width = maximum(width, x); + + f32 fontScale = mg_font_get_scale_for_em_pixels(font, fontSize); + mp_rect rect = {0, -fontInfo->extents.ascent * fontScale, width * fontScale, (y + lineHeight) * fontScale }; + return(rect); +} + +//----------------------------------------------------------------------------------------------------------- +// Transform matrix settings +//----------------------------------------------------------------------------------------------------------- + +void mg_matrix_push(mg_canvas handle, mg_mat2x3 matrix) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_MATRIX_PUSH, .matrix = matrix}); +} + +void mg_matrix_pop(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_MATRIX_POP}); +} +//----------------------------------------------------------------------------------------------------------- +// Graphics attributes settings +//----------------------------------------------------------------------------------------------------------- + +void mg_set_color(mg_canvas handle, mg_color color) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.color = color; +} + +void mg_set_color_rgba(mg_canvas handle, f32 r, f32 g, f32 b, f32 a) +{ + mg_set_color(handle, (mg_color){r, g, b, a}); +} + +void mg_set_width(mg_canvas handle, f32 width) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.width = width; +} + +void mg_set_tolerance(mg_canvas handle, f32 tolerance) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.tolerance = tolerance; +} + +void mg_set_joint(mg_canvas handle, mg_joint_type joint) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.joint = joint; +} + +void mg_set_max_joint_excursion(mg_canvas handle, f32 maxJointExcursion) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.maxJointExcursion = maxJointExcursion; +} + +void mg_set_cap(mg_canvas handle, mg_cap_type cap) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.cap = cap; +} + +void mg_set_font(mg_canvas handle, mg_font font) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.font = font; +} + +void mg_set_font_size(mg_canvas handle, f32 fontSize) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->attributes.fontSize = fontSize; +} + +void mg_set_text_flip(mg_canvas handle, bool flip) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + context->textFlip = flip; +} + + + +mg_color mg_get_color(mg_canvas handle) +{ + mg_color color = {0}; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + color = context->attributes.color; + } + return(color); +} + +f32 mg_get_width(mg_canvas handle) +{ + f32 width = 0; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + width = context->attributes.width; + } + return(width); +} + +f32 mg_get_tolerance(mg_canvas handle) +{ + f32 tolerance = 0; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + tolerance = context->attributes.tolerance; + } + return(tolerance); +} + +mg_joint_type mg_get_joint(mg_canvas handle) +{ + mg_joint_type joint = 0; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + joint = context->attributes.joint; + } + return(joint); +} + +f32 mg_get_max_joint_excursion(mg_canvas handle) +{ + f32 maxJointExcursion = 0; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + maxJointExcursion = context->attributes.maxJointExcursion; + } + return(maxJointExcursion); +} + +mg_cap_type mg_get_cap(mg_canvas handle) +{ + mg_cap_type cap = 0; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + cap = context->attributes.cap; + } + return(cap); +} + +mg_font mg_get_font(mg_canvas handle) +{ + mg_font font = mg_font_nil(); + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + font = context->attributes.font; + } + return(font); +} + +f32 mg_get_font_size(mg_canvas handle) +{ + f32 fontSize = 0; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + fontSize = context->attributes.fontSize; + } + return(fontSize); +} + +bool mg_get_text_flip(mg_canvas handle) +{ + bool flip = false; + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + flip = context->textFlip; + } + return(flip); +} + + + +//----------------------------------------------------------------------------------------------------------- +// Clip +//----------------------------------------------------------------------------------------------------------- + +void mg_clip_push(mg_canvas handle, f32 x, f32 y, f32 w, f32 h) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_CLIP_PUSH, + .rect = (mp_rect){x, y, w, h}}); +} + +void mg_clip_pop(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_CLIP_POP}); +} + +//----------------------------------------------------------------------------------------------------------- +// Path construction +//----------------------------------------------------------------------------------------------------------- + +void mg_new_path(mg_canvas_data* context) +{ + context->path.startIndex += context->path.count; + context->path.count = 0; + context->subPathStartPoint = context->subPathLastPoint; + context->path.startPoint = context->subPathStartPoint; +} + +void mg_clear(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, (mg_primitive){.cmd = MG_CMD_CLEAR, .attributes = context->attributes}); +} + +void mg_get_current_position(mg_canvas handle, f32* x, f32* y) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + *x = context->subPathLastPoint.x; + *y = context->subPathLastPoint.y; +} + +void mg_path_push_elements(mg_canvas_data* context, u32 count, mg_path_elt* elements) +{ + ASSERT(context->path.count + context->path.startIndex + count <= MG_MAX_PATH_ELEMENT_COUNT); + memcpy(context->pathElements + context->path.startIndex + context->path.count, elements, count*sizeof(mg_path_elt)); + context->path.count += count; +} + +void mg_path_push_element(mg_canvas_data* context, mg_path_elt elt) +{ + mg_path_push_elements(context, 1, &elt); +} + + + +void mg_move_to(mg_canvas handle, f32 x, f32 y) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_path_push_element(context, ((mg_path_elt){.type = MG_PATH_MOVE, .p[0] = {x, y}})); + context->subPathStartPoint = (vec2){x, y}; + context->subPathLastPoint = (vec2){x, y}; +} + +void mg_line_to(mg_canvas handle, f32 x, f32 y) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_path_push_element(context, ((mg_path_elt){.type = MG_PATH_LINE, .p[0] = {x, y}})); + context->subPathLastPoint = (vec2){x, y}; +} + +void mg_quadratic_to(mg_canvas handle, f32 x1, f32 y1, f32 x2, f32 y2) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_path_push_element(context, ((mg_path_elt){.type = MG_PATH_QUADRATIC, .p = {{x1, y1}, {x2, y2}}})); + context->subPathLastPoint = (vec2){x2, y2}; +} + +void mg_cubic_to(mg_canvas handle, f32 x1, f32 y1, f32 x2, f32 y2, f32 x3, f32 y3) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_path_push_element(context, ((mg_path_elt){.type = MG_PATH_CUBIC, .p = {{x1, y1}, {x2, y2}, {x3, y3}}})); + context->subPathLastPoint = (vec2){x3, y3}; +} + +void mg_close_path(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + if( context->subPathStartPoint.x != context->subPathLastPoint.x + || context->subPathStartPoint.y != context->subPathLastPoint.y) + { + mg_line_to(handle, context->subPathStartPoint.x, context->subPathStartPoint.y); + } + context->subPathStartPoint = context->subPathLastPoint; +} + +mp_rect mg_glyph_outlines_from_font_info(mg_canvas_data* context, mg_font_info* fontInfo, str32 glyphIndices) +{ + mg_canvas contextHandle = mg_canvas_handle_from_ptr(context); + + f32 startX = context->subPathLastPoint.x; + f32 startY = context->subPathLastPoint.y; + f32 maxWidth = 0; + + f32 scale = context->attributes.fontSize/fontInfo->unitsPerEm; + + for(int i=0; isubPathLastPoint.x; + f32 yOffset = context->subPathLastPoint.y; + f32 flip = context->textFlip ? -1 : 1; + + if(!glyphIndex || glyphIndex >= fontInfo->glyphCount) + { + LOG_WARNING("code point is not present in font ranges\n"); + //NOTE(martin): try to find the replacement character + glyphIndex = mg_font_get_glyph_index_from_font_info(fontInfo, 0xfffd); + if(!glyphIndex) + { + //NOTE(martin): could not find replacement glyph, try to get an 'x' to get a somewhat correct width + // to render an empty rectangle. Otherwise just render with the max font width + f32 boxWidth = fontInfo->extents.width * 0.8; + f32 xBearing = fontInfo->extents.width * 0.1; + f32 xAdvance = fontInfo->extents.width; + + glyphIndex = mg_font_get_glyph_index_from_font_info(fontInfo, 'x'); + if(glyphIndex) + { + mg_glyph_info* glyph = &(fontInfo->glyphs[glyphIndex]); + boxWidth = glyph->extents.width; + xBearing = glyph->extents.xBearing; + xAdvance = glyph->extents.xAdvance; + } + f32 oldStrokeWidth = context->attributes.width; + + mg_set_width(contextHandle, boxWidth*0.005); + mg_rectangle_stroke(contextHandle, + xOffset + xBearing * scale, + yOffset, + boxWidth * scale * flip, + fontInfo->extents.capHeight*scale); + + mg_set_width(contextHandle, oldStrokeWidth); + mg_move_to(contextHandle, xOffset + xAdvance * scale, yOffset); + maxWidth = maximum(maxWidth, xOffset + xAdvance*scale - startX); + continue; + } + } + + mg_glyph_info* glyph = mg_font_get_glyph_info(fontInfo, glyphIndex); + + mg_path_push_elements(context, glyph->pathDescriptor.count, fontInfo->outlines + glyph->pathDescriptor.startIndex); + + mg_path_elt* elements = context->pathElements + context->path.count + context->path.startIndex - glyph->pathDescriptor.count; + for(int eltIndex=0; eltIndexpathDescriptor.count; eltIndex++) + { + for(int pIndex = 0; pIndex < 3; pIndex++) + { + elements[eltIndex].p[pIndex].x = elements[eltIndex].p[pIndex].x * scale + xOffset; + elements[eltIndex].p[pIndex].y = elements[eltIndex].p[pIndex].y * scale * flip + yOffset; + } + } + mg_move_to(contextHandle, xOffset + scale*glyph->extents.xAdvance, yOffset); + + maxWidth = maximum(maxWidth, xOffset + scale*glyph->extents.xAdvance - startX); + } + f32 lineHeight = (fontInfo->extents.ascent + fontInfo->extents.descent)*scale; + mp_rect box = {startX, startY, maxWidth, context->subPathLastPoint.y - startY + lineHeight }; + return(box); +} + +mp_rect mg_glyph_outlines(mg_canvas handle, str32 glyphIndices) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return((mp_rect){}); + } + mg_font_info* fontInfo = mg_get_font_info(context->attributes.font); + if(!fontInfo) + { + return((mp_rect){}); + } + return(mg_glyph_outlines_from_font_info(context, fontInfo, glyphIndices)); +} + +void mg_text_outlines(mg_canvas handle, str8 text) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_font_info* fontInfo = mg_get_font_info(context->attributes.font); + if(!fontInfo) + { + return; + } + + mem_arena* scratch = mem_scratch(); + str32 codePoints = utf8_push_to_codepoints(scratch, text); + str32 glyphIndices = mg_font_push_glyph_indices(context->attributes.font, scratch, codePoints); + + mg_glyph_outlines_from_font_info(context, fontInfo, glyphIndices); +} + +//----------------------------------------------------------------------------------------------------------- +// Path primitives commands +//----------------------------------------------------------------------------------------------------------- + +void mg_fill(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + if(context->path.count) + { + mg_push_command(context, ((mg_primitive){.cmd = MG_CMD_FILL, .path = context->path, .attributes = context->attributes})); + mg_new_path(context); + } +} + +void mg_stroke(mg_canvas handle) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + if(context->path.count) + { + mg_push_command(context, ((mg_primitive){.cmd = MG_CMD_STROKE, .path = context->path, .attributes = context->attributes})); + mg_new_path(context); + } +} + +//----------------------------------------------------------------------------------------------------------- +// Fast shapes primitives commands +//----------------------------------------------------------------------------------------------------------- + +void mg_rectangle_fill(mg_canvas handle, f32 x, f32 y, f32 w, f32 h) +{ +// DEBUG_ASSERT(w>=0 && h>=0); + + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_RECT_FILL, .rect = (mp_rect){x, y, w, h}, .attributes = context->attributes})); +} +void mg_rectangle_stroke(mg_canvas handle, f32 x, f32 y, f32 w, f32 h) +{ +// DEBUG_ASSERT(w>=0 && h>=0); + + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_RECT_STROKE, .rect = (mp_rect){x, y, w, h}, .attributes = context->attributes})); +} + +void mg_rounded_rectangle_fill(mg_canvas handle, f32 x, f32 y, f32 w, f32 h, f32 r) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_ROUND_RECT_FILL, + .roundedRect = (mg_rounded_rect){x, y, w, h, r}, + .attributes = context->attributes})); +} + +void mg_rounded_rectangle_stroke(mg_canvas handle, f32 x, f32 y, f32 w, f32 h, f32 r) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_ROUND_RECT_STROKE, + .roundedRect = (mg_rounded_rect){x, y, w, h, r}, + .attributes = context->attributes})); +} + +void mg_circle_fill(mg_canvas handle, f32 x, f32 y, f32 r) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_ELLIPSE_FILL, + .rect = (mp_rect){x-r, y-r, 2*r, 2*r}, + .attributes = context->attributes})); +} + +void mg_circle_stroke(mg_canvas handle, f32 x, f32 y, f32 r) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_ELLIPSE_STROKE, + .rect = (mp_rect){x-r, y-r, 2*r, 2*r}, + .attributes = context->attributes})); +} + +void mg_ellipse_fill(mg_canvas handle, f32 x, f32 y, f32 rx, f32 ry) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_ELLIPSE_FILL, + .rect = (mp_rect){x-rx, y-ry, 2*rx, 2*ry}, + .attributes = context->attributes})); +} + +void mg_ellipse_stroke(mg_canvas handle, f32 x, f32 y, f32 rx, f32 ry) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_push_command(context, + ((mg_primitive){.cmd = MG_CMD_ELLIPSE_STROKE, + .rect = (mp_rect){x-rx, y-ry, 2*rx, 2*ry}, + .attributes = context->attributes})); +} + +void mg_arc(mg_canvas handle, f32 x, f32 y, f32 r, f32 arcAngle, f32 startAngle) +{ + f32 endAngle = startAngle + arcAngle; + + while(startAngle < endAngle) + { + f32 smallAngle = minimum(endAngle - startAngle, M_PI/4.); + if(smallAngle < 0.001) + { + break; + } + + vec2 v0 = {cos(smallAngle/2), sin(smallAngle/2)}; + vec2 v1 = {(4-v0.x)/3, (1-v0.x)*(3-v0.x)/(3*v0.y)}; + vec2 v2 = {v1.x, -v1.y}; + vec2 v3 = {v0.x, -v0.y}; + + f32 rotAngle = smallAngle/2 + startAngle; + f32 rotCos = cos(rotAngle); + f32 rotSin = sin(rotAngle); + + mg_mat2x3 t = {r*rotCos, -r*rotSin, x, + r*rotSin, r*rotCos, y}; + + v0 = mg_mat2x3_mul(t, v0); + v1 = mg_mat2x3_mul(t, v1); + v2 = mg_mat2x3_mul(t, v2); + v3 = mg_mat2x3_mul(t, v3); + + mg_move_to(handle, v0.x, v0.y); + mg_cubic_to(handle, v1.x, v1.y, v2.x, v2.y, v3.x, v3.y); + + startAngle += smallAngle; + } +} + +//------------------------------------------------------------------------------------------ +//NOTE(martin): images +//------------------------------------------------------------------------------------------ + +#define STB_IMAGE_IMPLEMENTATION +#include"stb_image.h" + +mg_image mg_image_nil() { return((mg_image){0}); } + +mg_image mg_image_handle_from_ptr(mg_canvas_data* canvas, mg_image_data* imageData) +{ + DEBUG_ASSERT( (imageData - canvas->images) >= 0 + && (imageData - canvas->images) < MG_MAX_CONTEXTS); + + u64 h = ((u64)(imageData - canvas->images))<<32 + |((u64)(imageData->generation)); + return((mg_image){.h = h}); +} + +void mg_image_data_recycle(mg_canvas_data* canvas, mg_image_data* image) +{ + image->generation++; + ListPush(&canvas->imageFreeList, &image->listElt); +} + +mg_image mg_image_create_from_rgba8(mg_canvas handle, u32 width, u32 height, u8* bytes) +{ + mg_image image = mg_image_nil(); + + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(context) + { + mg_image_data* imageData = ListPopEntry(&context->imageFreeList, mg_image_data, listElt); + if(!imageData) + { + if(context->imageNextIndex < MG_IMAGE_MAX_COUNT) + { + imageData = &context->images[context->imageNextIndex]; + imageData->generation = 1; + context->imageNextIndex++; + } + else + { + LOG_ERROR("image pool full\n"); + } + } + + if(imageData) + { + if(context->atlasPos.x + width >= MG_ATLAS_SIZE) + { + context->atlasPos.x = 0; + context->atlasPos.y += context->atlasLineHeight; + } + if(context->atlasPos.x + width < MG_ATLAS_SIZE + && context->atlasPos.y + height < MG_ATLAS_SIZE) + { + imageData->rect = (mp_rect){context->atlasPos.x, + context->atlasPos.y, + width, + height}; + + context->atlasPos.x += width; + context->atlasLineHeight = maximum(context->atlasLineHeight, height); + + context->painter->atlasUpload(context->painter, imageData->rect, bytes); + image = mg_image_handle_from_ptr(context, imageData); + } + else + { + mg_image_data_recycle(context, imageData); + } + } + } + return(image); +} + +mg_image mg_image_create_from_data(mg_canvas canvas, str8 data, bool flip) +{ + mg_image image = mg_image_nil(); + int width, height, channels; + + stbi_set_flip_vertically_on_load(flip ? 1 : 0); + u8* pixels = stbi_load_from_memory((u8*)data.ptr, data.len, &width, &height, &channels, 4); + if(pixels) + { + image = mg_image_create_from_rgba8(canvas, width, height, pixels); + free(pixels); + } + return(image); +} + +mg_image mg_image_create_from_file(mg_canvas canvas, str8 path, bool flip) +{ + mg_image image = mg_image_nil(); + int width, height, channels; + + const char* cpath = str8_to_cstring(mem_scratch(), path); + + stbi_set_flip_vertically_on_load(flip ? 1 : 0); + u8* pixels = stbi_load(cpath, &width, &height, &channels, 4); + if(pixels) + { + image = mg_image_create_from_rgba8(canvas, width, height, pixels); + free(pixels); + } + return(image); +} + +void mg_image_destroy(mg_canvas handle, mg_image image) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + //TODO invalidate image handle, maybe free atlas area +} + +void mg_image_draw(mg_canvas handle, mg_image image, mp_rect rect) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_primitive primitive = {.cmd = MG_CMD_IMAGE_DRAW, + .rect = (mp_rect){rect.x, rect.y, rect.h, rect.w}, + .attributes = context->attributes}; + primitive.attributes.image = image; + + mg_push_command(context, primitive); +} + +void mg_rounded_image_draw(mg_canvas handle, mg_image image, mp_rect rect, f32 roundness) +{ + mg_canvas_data* context = mg_canvas_ptr_from_handle(handle); + if(!context) + { + return; + } + mg_primitive primitive = {.cmd = MG_CMD_ROUNDED_IMAGE_DRAW, + .roundedRect = {rect.x, rect.y, rect.h, rect.w, roundness}, + .attributes = context->attributes}; + primitive.attributes.image = image; + + mg_push_command(context, primitive); +} + + +#undef LOG_SUBSYSTEM diff --git a/src/graphics.h b/src/graphics.h new file mode 100644 index 0000000..faebd49 --- /dev/null +++ b/src/graphics.h @@ -0,0 +1,228 @@ + /************************************************************//** +* +* @file: graphics.h +* @author: Martin Fouilleul +* @date: 01/08/2022 +* @revision: +* +*****************************************************************/ +#ifndef __GRAPHICS_H_ +#define __GRAPHICS_H_ + +#include"mp_app.h" + +#ifdef __cplusplus +extern "C" { +#endif + +//------------------------------------------------------------------------------------------ +//NOTE(martin): graphics surface +//------------------------------------------------------------------------------------------ + +typedef struct mg_surface { u64 h; } mg_surface; + +typedef enum { MG_BACKEND_DUMMY, + MG_BACKEND_METAL, + //... + } mg_backend_id; + +void mg_init(); + +mg_surface mg_surface_nil(); +mg_surface mg_surface_create_for_window(mp_window window, mg_backend_id backend); +mg_surface mg_surface_create_for_view(mp_view view, mg_backend_id backend); +void mg_surface_destroy(mg_surface surface); + +void mg_surface_prepare(mg_surface surface); +void mg_surface_present(mg_surface surface); +void mg_surface_resize(mg_surface surface, int width, int height); + +vec2 mg_surface_size(mg_surface surface); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): canvas drawing structs +//------------------------------------------------------------------------------------------ + +typedef struct mg_canvas { u64 h; } mg_canvas; +typedef struct mg_stream { u64 h; } mg_stream; +typedef struct mg_font { u64 h; } mg_font; + +typedef struct mg_mat2x3 +{ + f32 m[6]; +} mg_mat2x3; + +typedef enum { MG_COORDS_2D_DISPLAY_CENTER, + MG_COORDS_2D_DISPLAY_TOP_LEFT, + MG_COORDS_2D_DISPLAY_BOTTOM_LEFT } mg_coordinate_system; + +typedef struct mg_color +{ + union + { + struct + { + f32 r; + f32 g; + f32 b; + f32 a; + }; + f32 c[4]; + }; +} mg_color; + +typedef enum {MG_JOINT_MITER = 0, + MG_JOINT_BEVEL, + MG_JOINT_NONE } mg_joint_type; + +typedef enum {MG_CAP_NONE = 0, + MG_CAP_SQUARE } mg_cap_type; + +typedef struct mg_font_extents +{ + f32 ascent; // the extent above the baseline (by convention a positive value extends above the baseline) + f32 descent; // the extent below the baseline (by convention, positive value extends below the baseline) + f32 leading; // spacing between one row's descent and the next row's ascent + f32 xHeight; // height of the lower case letter 'x' + f32 capHeight; // height of the upper case letter 'M' + f32 width; // maximum width of the font + +} mg_font_extents; + +typedef struct mg_text_extents +{ + f32 xBearing; + f32 yBearing; + f32 width; + f32 height; + f32 xAdvance; + f32 yAdvance; + +} mg_text_extents; + +//------------------------------------------------------------------------------------------ +//NOTE(martin): canvas lifetime and command streams +//------------------------------------------------------------------------------------------ +mg_canvas mg_canvas_create(mg_surface surface, mp_rect viewPort); +void mg_canvas_destroy(mg_canvas context); +void mg_canvas_viewport(mg_canvas, mp_rect viewPort); +void mg_canvas_flush(mg_canvas context); + +mg_stream mg_stream_create(mg_canvas context); +mg_stream mg_stream_swap(mg_canvas context, mg_stream stream); +void mg_stream_append(mg_canvas context, mg_stream stream); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): fonts management +//------------------------------------------------------------------------------------------ +mg_font mg_font_create_from_memory(u32 size, byte* buffer, u32 rangeCount, unicode_range* ranges); +void mg_font_destroy(mg_font font); + +//NOTE(martin): the following int valued functions return -1 if font is invalid or codepoint is not present in font// +//TODO(martin): add enum error codes + +int mg_font_get_extents(mg_font font, mg_font_extents* outExtents); +f32 mg_font_get_scale_for_em_pixels(mg_font font, f32 emSize); + +//NOTE(martin): if you need to process more than one codepoint, first convert your codepoints to glyph indices, then use the +// glyph index versions of the functions, which can take an array of glyph indices. + +str32 mg_font_get_glyph_indices(mg_font font, str32 codePoints, str32 backing); +str32 mg_font_push_glyph_indices(mg_font font, mem_arena* arena, str32 codePoints); +u32 mg_font_get_glyph_index(mg_font font, utf32 codePoint); + +int mg_font_get_codepoint_extents(mg_font font, utf32 codePoint, mg_text_extents* outExtents); + +int mg_font_get_glyph_extents(mg_font font, str32 glyphIndices, mg_text_extents* outExtents); + +mp_rect mg_text_bounding_box(mg_font font, f32 fontSize, str8 text); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): matrix settings +//------------------------------------------------------------------------------------------ +void mg_matrix_push(mg_canvas context, mg_mat2x3 matrix); +void mg_matrix_pop(mg_canvas context); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): clipping +//------------------------------------------------------------------------------------------ +void mg_clip_push(mg_canvas context, f32 x, f32 y, f32 w, f32 h); +void mg_clip_pop(mg_canvas context); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): graphics attributes setting/getting +//------------------------------------------------------------------------------------------ +void mg_set_clear_color(mg_canvas context, mg_color color); +void mg_set_clear_color_rgba(mg_canvas context, f32 r, f32 g, f32 b, f32 a); +void mg_set_color(mg_canvas context, mg_color color); +void mg_set_color_rgba(mg_canvas context, f32 r, f32 g, f32 b, f32 a); +void mg_set_width(mg_canvas context, f32 width); +void mg_set_tolerance(mg_canvas context, f32 tolerance); +void mg_set_joint(mg_canvas context, mg_joint_type joint); +void mg_set_max_joint_excursion(mg_canvas context, f32 maxJointExcursion); +void mg_set_cap(mg_canvas context, mg_cap_type cap); +void mg_set_font(mg_canvas context, mg_font font); +void mg_set_font_size(mg_canvas context, f32 size); +void mg_set_text_flip(mg_canvas context, bool flip); + +mg_color mg_get_clear_color(mg_canvas context); +mg_color mg_get_color(mg_canvas context); +f32 mg_get_width(mg_canvas context); +f32 mg_get_tolerance(mg_canvas context); +mg_joint_type mg_get_joint(mg_canvas context); +f32 mg_get_max_joint_excursion(mg_canvas context); +mg_cap_type mg_get_cap(mg_canvas context); +mg_font mg_get_font(mg_canvas context); +f32 mg_get_font_size(mg_canvas context); +bool mg_get_text_flip(mg_canvas context); +//------------------------------------------------------------------------------------------ +//NOTE(martin): path construction +//------------------------------------------------------------------------------------------ +void mg_get_current_position(mg_canvas context, f32* x, f32* y); +void mg_move_to(mg_canvas context, f32 x, f32 y); +void mg_line_to(mg_canvas context, f32 x, f32 y); +void mg_quadratic_to(mg_canvas context, f32 x1, f32 y1, f32 x2, f32 y2); +void mg_cubic_to(mg_canvas context, f32 x1, f32 y1, f32 x2, f32 y2, f32 x3, f32 y3); +void mg_close_path(mg_canvas context); + +mp_rect mg_glyph_outlines(mg_canvas context, str32 glyphIndices); +void mg_text_outlines(mg_canvas context, str8 string); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): clear/fill/stroke +//------------------------------------------------------------------------------------------ +void mg_clear(mg_canvas context); +void mg_fill(mg_canvas context); +void mg_stroke(mg_canvas context); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): 'fast' shapes primitives +//------------------------------------------------------------------------------------------ +void mg_rectangle_fill(mg_canvas context, f32 x, f32 y, f32 w, f32 h); +void mg_rectangle_stroke(mg_canvas context, f32 x, f32 y, f32 w, f32 h); +void mg_rounded_rectangle_fill(mg_canvas context, f32 x, f32 y, f32 w, f32 h, f32 r); +void mg_rounded_rectangle_stroke(mg_canvas context, f32 x, f32 y, f32 w, f32 h, f32 r); +void mg_ellipse_fill(mg_canvas context, f32 x, f32 y, f32 rx, f32 ry); +void mg_ellipse_stroke(mg_canvas context, f32 x, f32 y, f32 rx, f32 ry); +void mg_circle_fill(mg_canvas context, f32 x, f32 y, f32 r); +void mg_circle_stroke(mg_canvas context, f32 x, f32 y, f32 r); +void mg_arc(mg_canvas handle, f32 x, f32 y, f32 r, f32 arcAngle, f32 startAngle); + +//------------------------------------------------------------------------------------------ +//NOTE(martin): images +//------------------------------------------------------------------------------------------ +typedef struct mg_image { u64 h; } mg_image; + +mg_image mg_image_create_from_rgba8(mg_canvas canvas, u32 width, u32 height, u8* pixels); +mg_image mg_image_create_from_data(mg_canvas canvas, str8 data, bool flip); +mg_image mg_image_create_from_file(mg_canvas canvas, str8 path, bool flip); + +void mg_image_drestroy(mg_canvas canvas, mg_image image); + +void mg_image_draw(mg_canvas canvas, mg_image image, mp_rect rect); +void mg_rounded_image_draw(mg_canvas handle, mg_image image, mp_rect rect, f32 roundness); + +#ifdef __cplusplus +} // extern "C" +#endif +#endif //__GRAPHICS_H_ diff --git a/src/graphics_internal.h b/src/graphics_internal.h new file mode 100644 index 0000000..2f5a215 --- /dev/null +++ b/src/graphics_internal.h @@ -0,0 +1,90 @@ +/************************************************************//** +* +* @file: graphics_internal.h +* @author: Martin Fouilleul +* @date: 01/08/2022 +* @revision: +* +*****************************************************************/ +#ifndef __GRAPHICS_INTERNAL_H_ +#define __GRAPHICS_INTERNAL_H_ + +#include"graphics.h" + +#ifdef __cplusplus +extern "C" { +#endif + +typedef struct mg_surface_info mg_surface_info; + +typedef void (*mg_surface_prepare_proc)(mg_surface_info* surface); +typedef void (*mg_surface_present_proc)(mg_surface_info* surface); +typedef void (*mg_surface_resize_proc)(mg_surface_info* surface, u32 width, u32 height); +typedef vec2 (*mg_surface_get_size_proc)(mg_surface_info* surface); +typedef void (*mg_surface_destroy_proc)(mg_surface_info* surface); + +typedef struct mg_surface_info +{ + mg_backend_id backend; + mg_surface_destroy_proc destroy; + mg_surface_prepare_proc prepare; + mg_surface_present_proc present; + mg_surface_resize_proc resize; + mg_surface_get_size_proc getSize; + +} mg_surface_info; + +mg_surface mg_surface_alloc_handle(mg_surface_info* surface); + +typedef struct mgc_vertex_layout +{ + u32 maxVertexCount; + u32 maxIndexCount; + + void* posBuffer; + u32 posStride; + + void* cubicBuffer; + u32 cubicStride; + + void* uvBuffer; + u32 uvStride; + + void* colorBuffer; + u32 colorStride; + + void* zIndexBuffer; + u32 zIndexStride; + + void* clipsBuffer; + u32 clipsStride; + + void* indexBuffer; + u32 indexStride; + +} mgc_vertex_layout; + +typedef struct mg_canvas_painter mg_canvas_painter; +typedef void (*mgc_painter_destroy_proc)(mg_canvas_painter* painter); +typedef void (*mgc_painter_draw_buffers_proc)(mg_canvas_painter* painter, u32 vertexCount, u32 indexCount, mg_color clearColor); +typedef void (*mgc_painter_set_viewport_proc)(mg_canvas_painter* painter, mp_rect viewPort); +typedef void (*mgc_painter_atlas_upload_proc)(mg_canvas_painter* painter, mp_rect rect, u8* bytes); + +typedef struct mg_canvas_painter +{ + mg_surface_info* surface; + mgc_vertex_layout vertexLayout; + mgc_painter_destroy_proc destroy; + mgc_painter_draw_buffers_proc drawBuffers; + mgc_painter_set_viewport_proc setViewPort; + mgc_painter_atlas_upload_proc atlasUpload; + +} mg_canvas_painter; + +#define MG_ATLAS_SIZE 8192 + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__GRAPHICS_INTERNAL_H_ diff --git a/src/metal_painter.mm b/src/metal_painter.mm new file mode 100644 index 0000000..34af6a5 --- /dev/null +++ b/src/metal_painter.mm @@ -0,0 +1,419 @@ +/************************************************************//** +* +* @file: graphics.mm +* @author: Martin Fouilleul +* @date: 12/07/2020 +* @revision: +* +*****************************************************************/ +#import +#import +#include + +#include"graphics_internal.h" +#include"macro_helpers.h" +#include"osx_app.h" + +#include"metal_shader.h" + +#define LOG_SUBSYSTEM "Graphics" + +static const int MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH = 4<<20; + +typedef struct mg_metal_painter +{ + mg_canvas_painter interface; + + mg_metal_surface* surface; + + // permanent metal resources + id tilingPipeline; + id sortingPipeline; + id boxingPipeline; + id computePipeline; + id renderPipeline; + + mp_rect viewPort; + + // textures and buffers + id outTexture; + id atlasTexture; + id vertexBuffer; + id indexBuffer; + id tileCounters; + id tilesArray; + id triangleArray; + id boxArray; + +} mg_metal_painter; + +void mg_metal_painter_draw_buffers(mg_canvas_painter* interface, u32 vertexCount, u32 indexCount, mg_color clearColor) +{ + mg_metal_painter* painter = (mg_metal_painter*)interface; + mg_metal_surface* surface = painter->surface; + + @autoreleasepool + { + if(surface->commandBuffer == nil || surface->commandBuffer == nil) + { + mg_metal_surface_acquire_drawable_and_command_buffer(surface); + } + + ASSERT(indexCount * sizeof(i32) < [painter->indexBuffer length]); + + f32 scale = surface->metalLayer.contentsScale; + vector_uint2 viewportSize = {painter->viewPort.w * scale, painter->viewPort.h * scale}; + + //----------------------------------------------------------- + //NOTE(martin): encode the clear counter + //----------------------------------------------------------- + id blitEncoder = [surface->commandBuffer blitCommandEncoder]; + [blitEncoder fillBuffer: painter->tileCounters range: NSMakeRange(0, RENDERER_MAX_TILES*sizeof(uint)) value: 0]; + [blitEncoder endEncoding]; + + //----------------------------------------------------------- + //NOTE(martin): encode the boxing pass + //----------------------------------------------------------- + id boxEncoder = [surface->commandBuffer computeCommandEncoder]; + [boxEncoder setComputePipelineState: painter->boxingPipeline]; + [boxEncoder setBuffer: painter->vertexBuffer offset:0 atIndex: 0]; + [boxEncoder setBuffer: painter->indexBuffer offset:0 atIndex: 1]; + [boxEncoder setBuffer: painter->triangleArray offset:0 atIndex: 2]; + [boxEncoder setBuffer: painter->boxArray offset:0 atIndex: 3]; + [boxEncoder setBytes: &scale length: sizeof(float) atIndex: 4]; + + MTLSize boxGroupSize = MTLSizeMake(painter->boxingPipeline.maxTotalThreadsPerThreadgroup, 1, 1); + MTLSize boxGridSize = MTLSizeMake(indexCount/3, 1, 1); + + [boxEncoder dispatchThreads: boxGridSize threadsPerThreadgroup: boxGroupSize]; + [boxEncoder endEncoding]; + + //----------------------------------------------------------- + //NOTE(martin): encode the tiling pass + //----------------------------------------------------------- + + id tileEncoder = [surface->commandBuffer computeCommandEncoder]; + [tileEncoder setComputePipelineState: painter->tilingPipeline]; + [tileEncoder setBuffer: painter->boxArray offset:0 atIndex: 0]; + [tileEncoder setBuffer: painter->tileCounters offset:0 atIndex: 1]; + [tileEncoder setBuffer: painter->tilesArray offset:0 atIndex: 2]; + [tileEncoder setBytes: &viewportSize length: sizeof(vector_uint2) atIndex: 3]; + + [tileEncoder dispatchThreads: boxGridSize threadsPerThreadgroup: boxGroupSize]; + [tileEncoder endEncoding]; + + //----------------------------------------------------------- + //NOTE(martin): encode the sorting pass + //----------------------------------------------------------- + + id sortEncoder = [surface->commandBuffer computeCommandEncoder]; + [sortEncoder setComputePipelineState: painter->sortingPipeline]; + [sortEncoder setBuffer: painter->tileCounters offset:0 atIndex: 0]; + [sortEncoder setBuffer: painter->triangleArray offset:0 atIndex: 1]; + [sortEncoder setBuffer: painter->tilesArray offset:0 atIndex: 2]; + [sortEncoder setBytes: &viewportSize length: sizeof(vector_uint2) atIndex: 3]; + + u32 nTilesX = (viewportSize.x + RENDERER_TILE_SIZE - 1)/RENDERER_TILE_SIZE; + u32 nTilesY = (viewportSize.y + RENDERER_TILE_SIZE - 1)/RENDERER_TILE_SIZE; + + MTLSize sortGroupSize = MTLSizeMake(painter->boxingPipeline.maxTotalThreadsPerThreadgroup, 1, 1); + MTLSize sortGridSize = MTLSizeMake(nTilesX*nTilesY, 1, 1); + + [sortEncoder dispatchThreads: sortGridSize threadsPerThreadgroup: sortGroupSize]; + [sortEncoder endEncoding]; + + //----------------------------------------------------------- + //NOTE(martin): create compute encoder and encode commands + //----------------------------------------------------------- + vector_float4 clearColorVec4 = {clearColor.r, clearColor.g, clearColor.b, clearColor.a}; + + id encoder = [surface->commandBuffer computeCommandEncoder]; + [encoder setComputePipelineState:painter->computePipeline]; + [encoder setTexture: painter->outTexture atIndex: 0]; + [encoder setTexture: painter->atlasTexture atIndex: 1]; + [encoder setBuffer: painter->vertexBuffer offset:0 atIndex: 0]; + [encoder setBuffer: painter->tileCounters offset:0 atIndex: 1]; + [encoder setBuffer: painter->tilesArray offset:0 atIndex: 2]; + [encoder setBuffer: painter->triangleArray offset:0 atIndex: 3]; + [encoder setBuffer: painter->boxArray offset:0 atIndex: 4]; + [encoder setBytes: &clearColorVec4 length: sizeof(vector_float4) atIndex: 5]; + + //TODO: check that we don't exceed maxTotalThreadsPerThreadgroup + DEBUG_ASSERT(RENDERER_TILE_SIZE*RENDERER_TILE_SIZE <= painter->computePipeline.maxTotalThreadsPerThreadgroup); + MTLSize threadGridSize = MTLSizeMake(viewportSize.x, viewportSize.y, 1); + MTLSize threadGroupSize = MTLSizeMake(RENDERER_TILE_SIZE, RENDERER_TILE_SIZE, 1); + + [encoder dispatchThreads: threadGridSize threadsPerThreadgroup:threadGroupSize]; + [encoder endEncoding]; + + //----------------------------------------------------------- + //NOTE(martin): acquire drawable, create render encoder to blit texture to framebuffer + //----------------------------------------------------------- + + MTLViewport viewport = {painter->viewPort.x * scale, + painter->viewPort.y * scale, + painter->viewPort.w * scale, + painter->viewPort.h * scale, + 0, + 1}; + + MTLRenderPassDescriptor* renderPassDescriptor = [MTLRenderPassDescriptor renderPassDescriptor]; + renderPassDescriptor.colorAttachments[0].texture = surface->drawable.texture; + renderPassDescriptor.colorAttachments[0].storeAction = MTLStoreActionStore; + + id renderEncoder = [surface->commandBuffer renderCommandEncoderWithDescriptor:renderPassDescriptor]; + [renderEncoder setViewport: viewport]; + [renderEncoder setRenderPipelineState: painter->renderPipeline]; + [renderEncoder setFragmentTexture: painter->outTexture atIndex: 0]; + [renderEncoder drawPrimitives: MTLPrimitiveTypeTriangle + vertexStart: 0 + vertexCount: 3 ]; + [renderEncoder endEncoding]; + } +} + +void mg_metal_painter_viewport(mg_canvas_painter* interface, mp_rect viewPort) +{ + mg_metal_painter* painter = (mg_metal_painter*)interface; + mg_metal_surface* surface = painter->surface; + + painter->viewPort = viewPort; + + @autoreleasepool + { + f32 scale = surface->metalLayer.contentsScale; + CGSize drawableSize = (CGSize){.width = viewPort.w * scale, .height = viewPort.h * scale}; + + [painter->outTexture release]; + + MTLTextureDescriptor* texDesc = [[MTLTextureDescriptor alloc] init]; + texDesc.textureType = MTLTextureType2D; + texDesc.storageMode = MTLStorageModePrivate; + texDesc.usage = MTLTextureUsageShaderRead | MTLTextureUsageShaderWrite; + texDesc.pixelFormat = MTLPixelFormatBGRA8Unorm;// MTLPixelFormatBGRA8Unorm_sRGB; + texDesc.width = drawableSize.width; + texDesc.height = drawableSize.height; + + painter->outTexture = [surface->device newTextureWithDescriptor:texDesc]; + } +} + +void mg_metal_painter_update_vertex_layout(mg_metal_painter* painter) +{ + mg_metal_surface* surface = painter->surface; + + char* vertexBase = (char*)[painter->vertexBuffer contents]; + + painter->interface.vertexLayout = (mgc_vertex_layout){ + .maxVertexCount = MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH, + .maxIndexCount = MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH, + .posBuffer = vertexBase + offsetof(mg_vertex, pos), + .posStride = sizeof(mg_vertex), + .cubicBuffer = vertexBase + offsetof(mg_vertex, cubic), + .cubicStride = sizeof(mg_vertex), + .uvBuffer = vertexBase + offsetof(mg_vertex, uv), + .uvStride = sizeof(mg_vertex), + .colorBuffer = vertexBase + offsetof(mg_vertex, color), + .colorStride = sizeof(mg_vertex), + .zIndexBuffer = vertexBase + offsetof(mg_vertex, zIndex), + .zIndexStride = sizeof(mg_vertex), + .clipsBuffer = vertexBase + offsetof(mg_vertex, clip), + .clipsStride = sizeof(mg_vertex), + .indexBuffer = [painter->indexBuffer contents], + .indexStride = sizeof(int)}; +} + + +void mg_metal_painter_destroy(mg_canvas_painter* interface) +{ + mg_metal_painter* painter = (mg_metal_painter*)interface; + + @autoreleasepool + { + [painter->outTexture release]; + [painter->atlasTexture release]; + [painter->vertexBuffer release]; + [painter->indexBuffer release]; + [painter->tilesArray release]; + [painter->triangleArray release]; + [painter->boxArray release]; + [painter->computePipeline release]; + } +} + +void mg_metal_painter_atlas_upload(mg_canvas_painter* interface, mp_rect rect, u8* bytes) +{@autoreleasepool{ + mg_metal_painter* painter = (mg_metal_painter*)interface; + + MTLRegion region = MTLRegionMake2D(rect.x, rect.y, rect.w, rect.h); + [painter->atlasTexture replaceRegion:region + mipmapLevel:0 + withBytes:(void*)bytes + bytesPerRow: 4 * rect.w]; +}} + +extern "C" mg_canvas_painter* mg_metal_painter_create_for_surface(mg_metal_surface* surface, mp_rect viewPort) +{ + mg_metal_painter* painter = malloc_type(mg_metal_painter); + painter->surface = surface; + + //NOTE(martin): setup interface functions + painter->interface.drawBuffers = mg_metal_painter_draw_buffers; + painter->interface.setViewPort = mg_metal_painter_viewport; + painter->interface.destroy = mg_metal_painter_destroy; + painter->interface.atlasUpload = mg_metal_painter_atlas_upload; + + painter->viewPort = viewPort; + + @autoreleasepool + { + f32 scale = surface->metalLayer.contentsScale; + CGSize drawableSize = (CGSize){.width = viewPort.w * scale, .height = viewPort.h * scale}; + + //----------------------------------------------------------- + //NOTE(martin): create our output texture + //----------------------------------------------------------- + MTLTextureDescriptor* texDesc = [[MTLTextureDescriptor alloc] init]; + texDesc.textureType = MTLTextureType2D; + texDesc.storageMode = MTLStorageModePrivate; + texDesc.usage = MTLTextureUsageShaderRead | MTLTextureUsageShaderWrite; + texDesc.pixelFormat = MTLPixelFormatBGRA8Unorm;// MTLPixelFormatBGRA8Unorm_sRGB; + texDesc.width = drawableSize.width; + texDesc.height = drawableSize.height; + + painter->outTexture = [surface->device newTextureWithDescriptor:texDesc]; + //TODO(martin): retain ? + + //----------------------------------------------------------- + //NOTE(martin): create our atlas texture + //----------------------------------------------------------- + texDesc.textureType = MTLTextureType2D; + texDesc.storageMode = MTLStorageModeManaged; + texDesc.usage = MTLTextureUsageShaderRead; + texDesc.pixelFormat = MTLPixelFormatRGBA8Unorm; //MTLPixelFormatBGRA8Unorm; + texDesc.width = MG_ATLAS_SIZE; + texDesc.height = MG_ATLAS_SIZE; + + painter->atlasTexture = [surface->device newTextureWithDescriptor:texDesc]; + + //----------------------------------------------------------- + //NOTE(martin): create buffers for vertex and index + //----------------------------------------------------------- + + MTLResourceOptions bufferOptions = MTLResourceCPUCacheModeWriteCombined + | MTLResourceStorageModeShared; + + painter->indexBuffer = [surface->device newBufferWithLength: MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH*sizeof(int) + options: bufferOptions]; + + painter->vertexBuffer = [surface->device newBufferWithLength: MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH*sizeof(mg_vertex) + options: bufferOptions]; + + painter->tilesArray = [surface->device newBufferWithLength: RENDERER_TILE_BUFFER_SIZE*sizeof(int)*RENDERER_MAX_TILES + options: MTLResourceStorageModePrivate]; + + painter->triangleArray = [surface->device newBufferWithLength: MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH*sizeof(mg_triangle_data) + options: MTLResourceStorageModePrivate]; + + painter->boxArray = [surface->device newBufferWithLength: MG_METAL_PAINTER_DEFAULT_BUFFER_LENGTH*sizeof(vector_float4) + options: MTLResourceStorageModePrivate]; + + //TODO(martin): retain ? + //----------------------------------------------------------- + //NOTE(martin): create and initialize tile counters + //----------------------------------------------------------- + painter->tileCounters = [surface->device newBufferWithLength: RENDERER_MAX_TILES*sizeof(uint) + options: MTLResourceStorageModePrivate]; + id commandBuffer = [surface->commandQueue commandBuffer]; + id blitEncoder = [commandBuffer blitCommandEncoder]; + [blitEncoder fillBuffer: painter->tileCounters range: NSMakeRange(0, RENDERER_MAX_TILES*sizeof(uint)) value: 0]; + [blitEncoder endEncoding]; + [commandBuffer commit]; + + //----------------------------------------------------------- + //NOTE(martin): load the library + //----------------------------------------------------------- + + //TODO(martin): filepath magic to find metallib path when not in the working directory + char* shaderPath = 0; + mp_app_get_resource_path("../resources/metal_shader.metallib", &shaderPath); + NSString* metalFileName = [[NSString alloc] initWithCString: shaderPath encoding: NSUTF8StringEncoding]; + free(shaderPath); + NSError* err = 0; + id library = [surface->device newLibraryWithFile: metalFileName error:&err]; + if(err != nil) + { + const char* errStr = [[err localizedDescription] UTF8String]; + LOG_ERROR("error : %s\n", errStr); + return(0); + } + id tilingFunction = [library newFunctionWithName:@"TileKernel"]; + id sortingFunction = [library newFunctionWithName:@"SortKernel"]; + id boxingFunction = [library newFunctionWithName:@"BoundingBoxKernel"]; + id computeFunction = [library newFunctionWithName:@"RenderKernel"]; + id vertexFunction = [library newFunctionWithName:@"VertexShader"]; + id fragmentFunction = [library newFunctionWithName:@"FragmentShader"]; + + //----------------------------------------------------------- + //NOTE(martin): setup our data layout and pipeline state + //----------------------------------------------------------- + NSError* error = NULL; + painter->computePipeline = [surface->device newComputePipelineStateWithFunction: computeFunction + error:&error]; + ASSERT(painter->computePipeline); + + MTLComputePipelineDescriptor* tilingPipelineDesc = [[MTLComputePipelineDescriptor alloc] init]; + tilingPipelineDesc.computeFunction = tilingFunction; +// tilingPipelineDesc.threadGroupSizeIsMultipleOfThreadExecutionWidth = true; + + painter->tilingPipeline = [surface->device newComputePipelineStateWithDescriptor: tilingPipelineDesc + options: MTLPipelineOptionNone + reflection: nil + error: &error]; + + MTLComputePipelineDescriptor* sortingPipelineDesc = [[MTLComputePipelineDescriptor alloc] init]; + sortingPipelineDesc.computeFunction = sortingFunction; +// sortingPipelineDesc.threadGroupSizeIsMultipleOfThreadExecutionWidth = true; + + painter->sortingPipeline = [surface->device newComputePipelineStateWithDescriptor: sortingPipelineDesc + options: MTLPipelineOptionNone + reflection: nil + error: &error]; + + MTLComputePipelineDescriptor* boxingPipelineDesc = [[MTLComputePipelineDescriptor alloc] init]; + boxingPipelineDesc.computeFunction = boxingFunction; +// boxingPipelineDesc.threadGroupSizeIsMultipleOfThreadExecutionWidth = true; + + painter->boxingPipeline = [surface->device newComputePipelineStateWithDescriptor: boxingPipelineDesc + options: MTLPipelineOptionNone + reflection: nil + error: &error]; + //----------------------------------------------------------- + //NOTE(martin): setup our render pipeline state + //----------------------------------------------------------- + // create and initialize the pipeline state descriptor + MTLRenderPipelineDescriptor *pipelineStateDescriptor = [[MTLRenderPipelineDescriptor alloc] init]; + pipelineStateDescriptor.label = @"My simple pipeline"; + pipelineStateDescriptor.vertexFunction = vertexFunction; + pipelineStateDescriptor.fragmentFunction = fragmentFunction; + pipelineStateDescriptor.colorAttachments[0].pixelFormat = surface->metalLayer.pixelFormat; + + // create render pipeline + painter->renderPipeline = [surface->device newRenderPipelineStateWithDescriptor: pipelineStateDescriptor error:&err]; + if(err != nil) + { + const char* errStr = [[err localizedDescription] UTF8String]; + const char* descStr = [[err localizedFailureReason] UTF8String]; + const char* recovStr = [[err localizedRecoverySuggestion] UTF8String]; + LOG_ERROR("(%li) %s. %s. %s\n", [err code], errStr, descStr, recovStr); + return(0); + } + } + + mg_metal_painter_update_vertex_layout(painter); + + return((mg_canvas_painter*)painter); +} + + +#undef LOG_SUBSYSTEM diff --git a/src/metal_shader.h b/src/metal_shader.h new file mode 100644 index 0000000..67c4a27 --- /dev/null +++ b/src/metal_shader.h @@ -0,0 +1,49 @@ +/************************************************************//** +* +* @file: metal_shader.h +* @author: Martin Fouilleul +* @date: 01/08/2022 +* @revision: +* +*****************************************************************/ +#ifndef __METAL_RENDERER_H_ +#define __METAL_RENDERER_H_ + +#include + +#define RENDERER_TILE_BUFFER_SIZE 4096 +#define RENDERER_TILE_SIZE 16 +#define RENDERER_MAX_TILES 65536 + +#define RENDERER_DEBUG_TILE_VISITED 0xf00d +#define RENDERER_DEBUG_TILE_BUFFER_OVERFLOW 0xdead + +typedef struct mg_vertex +{ + vector_float2 pos; // position + vector_float4 cubic; // canonical implicit curve space coordinates + vector_float2 uv; // texture coordinates + vector_float4 color; + vector_float4 clip; + int zIndex; + +} mg_vertex; + +typedef struct mg_triangle_data +{ + uint i0; + uint i1; + uint i2; + uint zIndex; + + vector_float2 p0; + vector_float2 p1; + vector_float2 p2; + + int bias0; + int bias1; + int bias2; + +} mg_triangle_data; + +#endif //__METAL_RENDERER_H_ diff --git a/src/metal_shader.metal b/src/metal_shader.metal new file mode 100644 index 0000000..86a9b27 --- /dev/null +++ b/src/metal_shader.metal @@ -0,0 +1,320 @@ + +#include +#include + +#include"metal_shader.h" + +using namespace metal; + +struct vs_out +{ + float4 pos [[position]]; + float2 uv; +}; + +vertex vs_out VertexShader(ushort vid [[vertex_id]]) +{ + vs_out out; + out.uv = float2((vid << 1) & 2, vid & 2); + out.pos = float4(out.uv * float2(2, 2) + float2(-1, -1), 0, 1); + return(out); +} + +fragment float4 FragmentShader(vs_out i [[stage_in]], texture2d tex [[texture(0)]]) +{ + constexpr sampler smp(mip_filter::nearest, mag_filter::linear, min_filter::linear); + return(float4(tex.sample(smp, i.uv).rgb, 1)); +} + + +bool is_top_left(float2 a, float2 b) +{ + return( (a.y == b.y && b.x < a.x) + ||(b.y < a.y)); +} + +kernel void BoundingBoxKernel(constant mg_vertex* vertexBuffer [[buffer(0)]], + constant uint* indexBuffer [[buffer(1)]], + device mg_triangle_data* triangleArray [[buffer(2)]], + device float4* boxArray [[buffer(3)]], + constant float* contentsScaling [[buffer(4)]], + uint gid [[thread_position_in_grid]]) +{ + uint triangleIndex = gid; + uint vertexIndex = triangleIndex*3; + + uint i0 = indexBuffer[vertexIndex]; + uint i1 = indexBuffer[vertexIndex+1]; + uint i2 = indexBuffer[vertexIndex+2]; + + float2 p0 = vertexBuffer[i0].pos * contentsScaling[0]; + float2 p1 = vertexBuffer[i1].pos * contentsScaling[0]; + float2 p2 = vertexBuffer[i2].pos * contentsScaling[0]; + + //NOTE(martin): compute triangle bounding box + float2 boxMin = min(min(p0, p1), p2); + float2 boxMax = max(max(p0, p1), p2); + + //NOTE(martin): clip bounding box against clip rect + vector_float4 clip = contentsScaling[0]*vertexBuffer[indexBuffer[vertexIndex]].clip; + float2 clipMin(clip.x, clip.y); + float2 clipMax(clip.x + clip.z-1, clip.y + clip.w-1); + + //NOTE(martin): intersect with current clip + boxMin = max(boxMin, clipMin); + boxMax = min(boxMax, clipMax); + + //NOTE(martin): reorder triangle counter-clockwise and compute bias for each edge + float cw = (p1 - p0).x*(p2 - p0).y - (p1 - p0).y*(p2 - p0).x; + if(cw < 0) + { + uint tmpIndex = i1; + i1 = i2; + i2 = tmpIndex; + + float2 tmpPoint = p1; + p1 = p2; + p2 = tmpPoint; + } + int bias0 = is_top_left(p1, p2) ? 0 : -1; + int bias1 = is_top_left(p2, p0) ? 0 : -1; + int bias2 = is_top_left(p0, p1) ? 0 : -1; + + //NOTE(martin): fill triangle data + boxArray[triangleIndex] = float4(boxMin.x, boxMin.y, boxMax.x, boxMax.y); + + triangleArray[triangleIndex].zIndex = vertexBuffer[i0].zIndex; + triangleArray[triangleIndex].i0 = i0; + triangleArray[triangleIndex].i1 = i1; + triangleArray[triangleIndex].i2 = i2; + triangleArray[triangleIndex].p0 = p0; + triangleArray[triangleIndex].p1 = p1; + triangleArray[triangleIndex].p2 = p2; + triangleArray[triangleIndex].bias0 = bias0; + triangleArray[triangleIndex].bias1 = bias1; + triangleArray[triangleIndex].bias2 = bias2; +} + +kernel void TileKernel(const device float4* boxArray [[buffer(0)]], + device volatile atomic_uint* tileCounters [[buffer(1)]], + device uint* tilesArray [[buffer(2)]], + constant vector_uint2* viewport [[buffer(3)]], + uint gid [[thread_position_in_grid]]) +{ + uint2 tilesMatrixDim = (*viewport + RENDERER_TILE_SIZE - 1) / RENDERER_TILE_SIZE; + uint nTilesX = tilesMatrixDim.x; + uint nTilesY = tilesMatrixDim.y; + + uint triangleIndex = gid; + uint4 box = uint4(floor(boxArray[triangleIndex]/RENDERER_TILE_SIZE)); + uint xMin = max((uint)0, box.x); + uint yMin = max((uint)0, box.y); + uint xMax = min(box.z, nTilesX-1); + uint yMax = min(box.w, nTilesY-1); + + for(uint y = yMin; y <= yMax; y++) + { + for(uint x = xMin ; x <= xMax; x++) + { + uint tileIndex = y*nTilesX + x; + device uint* tileBuffer = tilesArray + tileIndex*RENDERER_TILE_BUFFER_SIZE; + uint counter = atomic_fetch_add_explicit(&(tileCounters[tileIndex]), 1, memory_order_relaxed); + tileBuffer[counter] = triangleIndex; + } + } +} + +kernel void SortKernel(const device uint* tileCounters [[buffer(0)]], + const device mg_triangle_data* triangleArray [[buffer(1)]], + device uint* tilesArray [[buffer(2)]], + constant vector_uint2* viewport [[buffer(3)]], + uint gid [[thread_position_in_grid]]) +{ + uint tileIndex = gid; + device uint* tileBuffer = tilesArray + tileIndex*RENDERER_TILE_BUFFER_SIZE; + uint tileBufferSize = tileCounters[tileIndex]; + + for(int eltIndex=0; eltIndex < (int)tileBufferSize; eltIndex++) + { + uint elt = tileBuffer[eltIndex]; + uint eltZIndex = triangleArray[elt].zIndex; + + int backIndex = eltIndex-1; + for(; backIndex >= 0; backIndex--) + { + uint backElt = tileBuffer[backIndex]; + uint backEltZIndex = triangleArray[backElt].zIndex; + if(eltZIndex >= backEltZIndex) + { + break; + } + else + { + tileBuffer[backIndex+1] = backElt; + } + } + tileBuffer[backIndex+1] = elt; + } +} + +float orient2d(float2 a, float2 b, float2 c) +{ + ////////////////////////////////////////////////////////////////////////////////////////// + //TODO(martin): FIX this. This is a **horrible** quick hack to fix the precision issues + // arising when a, b, and c are close. But it degrades when a, c, and c + // are big. The proper solution is to change the expression to avoid + // precision loss but I'm too busy/lazy to do it now. + ////////////////////////////////////////////////////////////////////////////////////////// + a *= 10; + b *= 10; + c *= 10; + return((b.x-a.x)*(c.y-a.y) - (b.y-a.y)*(c.x-a.x)); +} + +kernel void RenderKernel(texture2d outTexture [[texture(0)]], + texture2d texAtlas [[texture(1)]], + const device mg_vertex* vertexBuffer [[buffer(0)]], + device uint* tileCounters [[buffer(1)]], + const device uint* tilesArray [[buffer(2)]], + const device mg_triangle_data* triangleArray [[buffer(3)]], + const device float4* boxArray [[buffer(4)]], + constant vector_float4* clearColor [[buffer(5)]], + uint2 gid [[thread_position_in_grid]], + uint2 tgid [[threadgroup_position_in_grid]], + uint2 threadsPerThreadgroup [[threads_per_threadgroup]], + uint2 gridSize [[threads_per_grid]]) +{ + //TODO: guard against thread group size not equal to tile size? + const uint2 tilesMatrixDim = (gridSize + RENDERER_TILE_SIZE - 1) / RENDERER_TILE_SIZE; + const uint2 tilePos = tgid * threadsPerThreadgroup / RENDERER_TILE_SIZE; + const uint tileIndex = tilePos.y * tilesMatrixDim.x + tilePos.x; + const device uint* tileBuffer = tilesArray + tileIndex * RENDERER_TILE_BUFFER_SIZE; + + const uint tileBufferSize = tileCounters[tileIndex]; + +#ifdef RENDERER_DEBUG_TILES + //NOTE(martin): color code debug values and show the tile grid + if((gid.x % RENDERER_TILE_SIZE == 0) || (gid.y % RENDERER_TILE_SIZE == 0)) + { + outTexture.write(float4(0, 0, 0, 1), gid); + return; + } + if(tileBufferSize <= 0) + { + outTexture.write(float4(0, 1, 0, 1), gid); + return; + } +#endif + const float2 sampleOffsets[6] = { float2(5./12, 5./12), + float2(-3./12, 3./12), + float2(1./12, 1./12), + float2(3./12, -1./12), + float2(-5./12, -3./12), + float2(-1./12, -5./12)}; + + int zIndices[6]; + uint flipCounts[6]; + float4 pixelColors[6]; + float4 nextColors[6]; + for(int i=0; i<6; i++) + { + zIndices[i] = -1; + flipCounts[i] = 0; + pixelColors[i] = *clearColor; + nextColors[i] = *clearColor; + } + + for(uint tileBufferIndex=0; tileBufferIndex < tileBufferSize; tileBufferIndex++) + { + float4 box = boxArray[tileBuffer[tileBufferIndex]]; + const device mg_triangle_data* triangle = &triangleArray[tileBuffer[tileBufferIndex]]; + + float2 p0 = triangle->p0; + float2 p1 = triangle->p1; + float2 p2 = triangle->p2; + + int bias0 = triangle->bias0; + int bias1 = triangle->bias1; + int bias2 = triangle->bias2; + + const device mg_vertex* v0 = &(vertexBuffer[triangle->i0]); + const device mg_vertex* v1 = &(vertexBuffer[triangle->i1]); + const device mg_vertex* v2 = &(vertexBuffer[triangle->i2]); + + float4 cubic0 = v0->cubic; + float4 cubic1 = v1->cubic; + float4 cubic2 = v2->cubic; + + float2 uv0 = v0->uv; + float2 uv1 = v1->uv; + float2 uv2 = v2->uv; + + int zIndex = v0->zIndex; + float4 color = v0->color; + + for(int i=0; i<6; i++) + { + float2 samplePoint = (float2)gid + sampleOffsets[i]; + + //NOTE(martin): cull if pixel is outside box + if(samplePoint.x < box.x || samplePoint.x > box.z || samplePoint.y < box.y || samplePoint.y > box.w) + { + continue; + } + + float w0 = orient2d(p1, p2, samplePoint); + float w1 = orient2d(p2, p0, samplePoint); + float w2 = orient2d(p0, p1, samplePoint); + + if(((int)w0+bias0) >= 0 && ((int)w1+bias1) >= 0 && ((int)w2+bias2) >= 0) + { + float4 cubic = (cubic0*w0 + cubic1*w1 + cubic2*w2)/(w0+w1+w2); + float2 uv = (uv0*w0 + uv1*w1 + uv2*w2)/(w0+w1+w2); + + constexpr sampler smp(mip_filter::nearest, mag_filter::linear, min_filter::linear); + float4 texColor = texAtlas.sample(smp, uv); + + //TODO(martin): this is a quick and dirty fix for solid polygons where we use + // cubic = (1, 1, 1, 1) on all vertices, which can cause small errors to + // flip the sign. + // We should really use another value that always lead to <= 0, but we must + // make sure we never share these vertices with bezier shapes. + // Alternatively, an ugly (but maybe less than this one) solution would be + // to check if uvs are equal on all vertices of the triangle and always render + // those triangles. + float eps = 0.0001; + if(cubic.w*(cubic.x*cubic.x*cubic.x - cubic.y*cubic.z) <= eps) + { + if(zIndex == zIndices[i]) + { + flipCounts[i]++; + } + else + { + if(flipCounts[i] & 0x01) + { + pixelColors[i] = nextColors[i]; + } + + float4 nextCol = color*texColor; + nextColors[i] = pixelColors[i]*(1-nextCol.a) +nextCol.a*nextCol; + + zIndices[i] = zIndex; + flipCounts[i] = 1; + } + } + } + } + } + float4 out = float4(0, 0, 0, 0); + for(int i=0; i<6; i++) + { + if(flipCounts[i] & 0x01) + { + pixelColors[i] = nextColors[i]; + } + out += pixelColors[i]; + } + out = float4(out.xyz/6, 1); + outTexture.write(out, gid); +} diff --git a/src/metal_surface.mm b/src/metal_surface.mm new file mode 100644 index 0000000..d5ab513 --- /dev/null +++ b/src/metal_surface.mm @@ -0,0 +1,231 @@ +/************************************************************//** +* +* @file: graphics.mm +* @author: Martin Fouilleul +* @date: 12/07/2020 +* @revision: +* +*****************************************************************/ +#import +#import +#include + +#include"graphics_internal.h" +#include"macro_helpers.h" +#include"osx_app.h" + +#define LOG_SUBSYSTEM "Graphics" + +static const u32 MP_METAL_MAX_DRAWABLES_IN_FLIGHT = 3; + +typedef struct mg_metal_surface +{ + mg_surface_info interface; + + // permanent metal resources + NSView* view; + id device; + CAMetalLayer* metalLayer; + id commandQueue; + + // transient metal resources + id drawable; + id commandBuffer; + + dispatch_semaphore_t drawableSemaphore; + +} mg_metal_surface; + +void mg_metal_surface_acquire_drawable_and_command_buffer(mg_metal_surface* surface) +{@autoreleasepool{ + /*WARN(martin): this is super important + When the app is put in the background, it seems that if there are buffers in flight, the drawables to + can be leaked. This causes the gpu to allocate more and more drawables, until the app crashes. + (note: the drawable objects themselves are released once the app comes back to the forefront, but the + memory allocated in the GPU is never freed...) + + In background the gpu seems to create drawable if none is available instead of actually + blocking on nextDrawable. These drawable never get freed. + This is not a problem if our shader is fast enough, since a previous drawable becomes + available before we finish the frame. But we want to protect against it anyway + + The normal blocking mechanism of nextDrawable seems useless, so we implement our own scheme by + counting the number of drawables available with a semaphore that gets decremented here and + incremented in the presentedHandler of the drawable. + Thus we ensure that we don't consume more drawables than we are able to draw. + + //TODO: we _also_ should stop trying to render if we detect that the app is in the background + or occluded, but we can't count only on this because there is a potential race between the + notification of background mode and the rendering. + + //TODO: We should set a reasonable timeout and skip the frame and log an error in case we are stalled + for too long. + */ + dispatch_semaphore_wait(surface->drawableSemaphore, DISPATCH_TIME_FOREVER); + + surface->drawable = [surface->metalLayer nextDrawable]; + ASSERT(surface->drawable != nil); + + //TODO: make this a weak reference if we use ARC + dispatch_semaphore_t semaphore = surface->drawableSemaphore; + [surface->drawable addPresentedHandler:^(id drawable){ + dispatch_semaphore_signal(semaphore); + }]; + + //NOTE(martin): create a command buffer + surface->commandBuffer = [surface->commandQueue commandBuffer]; + + [surface->commandBuffer retain]; + [surface->drawable retain]; +}} + +void mg_metal_surface_prepare(mg_surface_info* interface) +{ + mg_metal_surface* surface = (mg_metal_surface*)interface; + mg_metal_surface_acquire_drawable_and_command_buffer(surface); +} + +void mg_metal_surface_present(mg_surface_info* interface) +{ + mg_metal_surface* surface = (mg_metal_surface*)interface; + @autoreleasepool + { + //NOTE(martin): present drawable and commit command buffer + [surface->commandBuffer presentDrawable: surface->drawable]; + [surface->commandBuffer commit]; + [surface->commandBuffer waitUntilCompleted]; + + //NOTE(martin): acquire next frame resources + [surface->commandBuffer release]; + surface->commandBuffer = nil; + [surface->drawable release]; + surface->drawable = nil; + } +} + +void mg_metal_surface_destroy(mg_surface_info* interface) +{ + mg_metal_surface* surface = (mg_metal_surface*)interface; + + @autoreleasepool + { + [surface->commandQueue release]; + [surface->metalLayer release]; + [surface->device release]; + } +} + +void mg_metal_surface_resize(mg_surface_info* interface, u32 width, u32 height) +{ + mg_metal_surface* surface = (mg_metal_surface*)interface; + + @autoreleasepool + { + //TODO(martin): actually detect scaling + CGRect frame = {{0, 0}, {width, height}}; + [surface->view setFrame: frame]; + f32 scale = surface->metalLayer.contentsScale; + CGSize drawableSize = (CGSize){.width = width * scale, .height = height * scale}; + surface->metalLayer.drawableSize = drawableSize; + } +} + +vec2 mg_metal_surface_get_size(mg_surface_info* interface) +{ + mg_metal_surface* surface = (mg_metal_surface*)interface; + + @autoreleasepool + { + //TODO(martin): actually detect scaling + CGRect frame = surface->view.frame; + return((vec2){frame.size.width, frame.size.height}); + } +} + +static const f32 MG_METAL_SURFACE_CONTENTS_SCALING = 2; + +extern "C" mg_surface mg_metal_surface_create_for_view(mp_view view) +{ + mp_view_data* viewData = mp_view_ptr_from_handle(view); + if(!viewData) + { + return(mg_surface_nil()); + } + else + { + mg_metal_surface* surface = (mg_metal_surface*)malloc(sizeof(mg_metal_surface)); + + //NOTE(martin): setup interface functions + surface->interface.backend = MG_BACKEND_METAL; + surface->interface.destroy = mg_metal_surface_destroy; + surface->interface.prepare = mg_metal_surface_prepare; + surface->interface.present = mg_metal_surface_present; + surface->interface.resize = mg_metal_surface_resize; + surface->interface.getSize = mg_metal_surface_get_size; + + @autoreleasepool + { + surface->drawableSemaphore = dispatch_semaphore_create(MP_METAL_MAX_DRAWABLES_IN_FLIGHT); + + //----------------------------------------------------------- + //NOTE(martin): create a metal device and a metal layer and + //----------------------------------------------------------- + surface->device = MTLCreateSystemDefaultDevice(); + [surface->device retain]; + surface->metalLayer = [CAMetalLayer layer]; + [surface->metalLayer retain]; + surface->metalLayer.device = surface->device; + + surface->view = viewData->nsView; + + //----------------------------------------------------------- + //NOTE(martin): set the size and scaling + //----------------------------------------------------------- + CGSize size = surface->view.bounds.size; + size.width *= MG_METAL_SURFACE_CONTENTS_SCALING; + size.height *= MG_METAL_SURFACE_CONTENTS_SCALING; + surface->metalLayer.drawableSize = size; + surface->metalLayer.contentsScale = MG_METAL_SURFACE_CONTENTS_SCALING; + + surface->metalLayer.pixelFormat = MTLPixelFormatBGRA8Unorm; + + surface->view.wantsLayer = YES; + surface->view.layer = surface->metalLayer; + + //NOTE(martin): handling resizing + surface->view.layerContentsRedrawPolicy = NSViewLayerContentsRedrawDuringViewResize; + surface->metalLayer.autoresizingMask = kCALayerHeightSizable | kCALayerWidthSizable; + surface->metalLayer.needsDisplayOnBoundsChange = YES; + + //----------------------------------------------------------- + //NOTE(martin): create a command queue + //----------------------------------------------------------- + surface->commandQueue = [surface->device newCommandQueue]; + [surface->commandQueue retain]; + + //NOTE(martin): command buffer and drawable are set on demand and at the end of each present() call + surface->drawable = nil; + surface->commandBuffer = nil; + } + + mg_surface handle = mg_surface_alloc_handle((mg_surface_info*)surface); + viewData->surface = handle; + + return(handle); + } +} + +extern "C" mg_surface mg_metal_surface_create_for_window(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(!windowData) + { + return(mg_surface_nil()); + } + else + { + return(mg_metal_surface_create_for_view(windowData->mainView)); + } +} + +#undef LOG_SUBSYSTEM diff --git a/src/milepost.c b/src/milepost.c new file mode 100644 index 0000000..7c4da4d --- /dev/null +++ b/src/milepost.c @@ -0,0 +1,29 @@ +/************************************************************//** +* +* @file: milepost.c +* @author: Martin Fouilleul +* @date: 13/02/2021 +* @revision: +* +*****************************************************************/ + +#include"util/debug_log.c" +#include"memory.c" +#include"strings.c" +#include"ringbuffer.c" +#include"util/utf8.c" +#include"util/hash.c" + +#include"platform/unix_base_allocator.c" + +#include"graphics.c" +#include"ui.c" + +//TODO: guard these under platform-specific #ifdefs +#include"platform/osx_clock.c" +/* + +#include"platform/unix_rng.c" +#include"platform/posix_thread.c" +#include"platform/posix_socket.c" +*/ diff --git a/src/milepost.h b/src/milepost.h new file mode 100644 index 0000000..e4603be --- /dev/null +++ b/src/milepost.h @@ -0,0 +1,31 @@ +/************************************************************//** +* +* @file: milepost.h +* @author: Martin Fouilleul +* @date: 13/02/2021 +* @revision: +* +*****************************************************************/ +#ifndef __MILEPOST_H_ +#define __MILEPOST_H_ + + +#include"typedefs.h" +#include"macro_helpers.h" +#include"debug_log.h" +#include"lists.h" +#include"memory.h" +#include"strings.h" +#include"utf8.h" + +#include"platform_clock.h" +/* +#include"platform_rng.h" +#include"platform_socket.h" +#include"platform_thread.h" +*/ +#include"mp_app.h" +#include"graphics.h" +#include"ui.h" + +#endif //__MILEPOST_H_ diff --git a/src/milepost.mm b/src/milepost.mm new file mode 100644 index 0000000..98124c6 --- /dev/null +++ b/src/milepost.mm @@ -0,0 +1,12 @@ +/************************************************************//** +* +* @file: milepost.mm +* @author: Martin Fouilleul +* @date: 13/02/2021 +* @revision: +* +*****************************************************************/ + +#include"osx_app.mm" +#include"metal_surface.mm" +#include"metal_painter.mm" diff --git a/src/mp_app.h b/src/mp_app.h new file mode 100644 index 0000000..ab54328 --- /dev/null +++ b/src/mp_app.h @@ -0,0 +1,413 @@ +/************************************************************//** +* +* @file: platform_app.h +* @author: Martin Fouilleul +* @date: 16/05/2020 +* @revision: +* +*****************************************************************/ +#ifndef __PLATFORM_APP_H_ +#define __PLATFORM_APP_H_ + +#include"typedefs.h" +#include"utf8.h" +#include"lists.h" + +#ifdef __cplusplus +extern "C" { +#endif + +//-------------------------------------------------------------------- +// Typedefs, enums and constants +//-------------------------------------------------------------------- + +typedef struct mp_window { u64 h; } mp_window; +typedef struct mp_view { u64 h; } mp_view; + +typedef enum { MP_MOUSE_CURSOR_ARROW, + MP_MOUSE_CURSOR_RESIZE_0, + MP_MOUSE_CURSOR_RESIZE_90, + MP_MOUSE_CURSOR_RESIZE_45, + MP_MOUSE_CURSOR_RESIZE_135, + MP_MOUSE_CURSOR_TEXT } mp_mouse_cursor; + +typedef i32 mp_window_style; +static const mp_window_style MP_WINDOW_STYLE_NO_TITLE = 0x01<<0, + MP_WINDOW_STYLE_FIXED_SIZE = 0x01<<1, + MP_WINDOW_STYLE_NO_CLOSE = 0x01<<2, + MP_WINDOW_STYLE_NO_MINIFY = 0x01<<3, + MP_WINDOW_STYLE_NO_FOCUS = 0x01<<4, + MP_WINDOW_STYLE_FLOAT = 0x01<<5, + MP_WINDOW_STYLE_POPUPMENU = 0x01<<6, + MP_WINDOW_STYLE_NO_BUTTONS = 0x01<<7; + +typedef enum { MP_EVENT_NONE, + MP_EVENT_KEYBOARD_MODS, + MP_EVENT_KEYBOARD_KEY, + MP_EVENT_KEYBOARD_CHAR, + MP_EVENT_MOUSE_BUTTON, + MP_EVENT_MOUSE_MOVE, + MP_EVENT_MOUSE_WHEEL, + MP_EVENT_MOUSE_ENTER, + MP_EVENT_MOUSE_LEAVE, + MP_EVENT_WINDOW_RESIZE, + MP_EVENT_WINDOW_MOVE, + MP_EVENT_WINDOW_FOCUS, + MP_EVENT_WINDOW_UNFOCUS, + MP_EVENT_WINDOW_HIDE, + MP_EVENT_WINDOW_SHOW, + MP_EVENT_WINDOW_CLOSE, + MP_EVENT_CLIPBOARD, + MP_EVENT_PATHDROP, + MP_EVENT_FRAME, + MP_EVENT_QUIT } mp_event_type; + +typedef enum { MP_KEY_NO_ACTION, + MP_KEY_PRESS, + MP_KEY_RELEASE, + MP_KEY_REPEAT } mp_key_action; + +typedef i32 mp_key_code; + +static const mp_key_code MP_KEY_UNKNOWN = -1, + MP_KEY_SPACE = 32, + MP_KEY_APOSTROPHE = 39, // ' + MP_KEY_COMMA = 44, // , + MP_KEY_MINUS = 45, // - + MP_KEY_PERIOD = 46, // . + MP_KEY_SLASH = 47, // / + MP_KEY_0 = 48, + MP_KEY_1 = 49, + MP_KEY_2 = 50, + MP_KEY_3 = 51, + MP_KEY_4 = 52, + MP_KEY_5 = 53, + MP_KEY_6 = 54, + MP_KEY_7 = 55, + MP_KEY_8 = 56, + MP_KEY_9 = 57, + MP_KEY_SEMICOLON = 59, // ; + MP_KEY_EQUAL = 61, // = + MP_KEY_A = 65, + MP_KEY_B = 66, + MP_KEY_C = 67, + MP_KEY_D = 68, + MP_KEY_E = 69, + MP_KEY_F = 70, + MP_KEY_G = 71, + MP_KEY_H = 72, + MP_KEY_I = 73, + MP_KEY_J = 74, + MP_KEY_K = 75, + MP_KEY_L = 76, + MP_KEY_M = 77, + MP_KEY_N = 78, + MP_KEY_O = 79, + MP_KEY_P = 80, + MP_KEY_Q = 81, + MP_KEY_R = 82, + MP_KEY_S = 83, + MP_KEY_T = 84, + MP_KEY_U = 85, + MP_KEY_V = 86, + MP_KEY_W = 87, + MP_KEY_X = 88, + MP_KEY_Y = 89, + MP_KEY_Z = 90, + MP_KEY_LEFT_BRACKET = 91, // [ + MP_KEY_BACKSLASH = 92, // \ */ + MP_KEY_RIGHT_BRACKET = 93, // ] + MP_KEY_GRAVE_ACCENT = 96, // ` + MP_KEY_WORLD_1 = 161, // non-US #1 + MP_KEY_WORLD_2 = 162, // non-US #2 + MP_KEY_ESCAPE = 256, + MP_KEY_ENTER = 257, + MP_KEY_TAB = 258, + MP_KEY_BACKSPACE = 259, + MP_KEY_INSERT = 260, + MP_KEY_DELETE = 261, + MP_KEY_RIGHT = 262, + MP_KEY_LEFT = 263, + MP_KEY_DOWN = 264, + MP_KEY_UP = 265, + MP_KEY_PAGE_UP = 266, + MP_KEY_PAGE_DOWN = 267, + MP_KEY_HOME = 268, + MP_KEY_END = 269, + MP_KEY_CAPS_LOCK = 280, + MP_KEY_SCROLL_LOCK = 281, + MP_KEY_NUM_LOCK = 282, + MP_KEY_PRINT_SCREEN = 283, + MP_KEY_PAUSE = 284, + MP_KEY_F1 = 290, + MP_KEY_F2 = 291, + MP_KEY_F3 = 292, + MP_KEY_F4 = 293, + MP_KEY_F5 = 294, + MP_KEY_F6 = 295, + MP_KEY_F7 = 296, + MP_KEY_F8 = 297, + MP_KEY_F9 = 298, + MP_KEY_F10 = 299, + MP_KEY_F11 = 300, + MP_KEY_F12 = 301, + MP_KEY_F13 = 302, + MP_KEY_F14 = 303, + MP_KEY_F15 = 304, + MP_KEY_F16 = 305, + MP_KEY_F17 = 306, + MP_KEY_F18 = 307, + MP_KEY_F19 = 308, + MP_KEY_F20 = 309, + MP_KEY_F21 = 310, + MP_KEY_F22 = 311, + MP_KEY_F23 = 312, + MP_KEY_F24 = 313, + MP_KEY_F25 = 314, + MP_KEY_KP_0 = 320, + MP_KEY_KP_1 = 321, + MP_KEY_KP_2 = 322, + MP_KEY_KP_3 = 323, + MP_KEY_KP_4 = 324, + MP_KEY_KP_5 = 325, + MP_KEY_KP_6 = 326, + MP_KEY_KP_7 = 327, + MP_KEY_KP_8 = 328, + MP_KEY_KP_9 = 329, + MP_KEY_KP_DECIMAL = 330, + MP_KEY_KP_DIVIDE = 331, + MP_KEY_KP_MULTIPLY = 332, + MP_KEY_KP_SUBTRACT = 333, + MP_KEY_KP_ADD = 334, + MP_KEY_KP_ENTER = 335, + MP_KEY_KP_EQUAL = 336, + MP_KEY_LEFT_SHIFT = 340, + MP_KEY_LEFT_CONTROL = 341, + MP_KEY_LEFT_ALT = 342, + MP_KEY_LEFT_SUPER = 343, + MP_KEY_RIGHT_SHIFT = 344, + MP_KEY_RIGHT_CONTROL = 345, + MP_KEY_RIGHT_ALT = 346, + MP_KEY_RIGHT_SUPER = 347, + MP_KEY_MENU = 348; + +static const mp_key_code MP_KEY_MAX = MP_KEY_MENU; + +typedef u8 mp_key_mods; +static const mp_key_mods MP_KEYMOD_NONE = 0x00, + MP_KEYMOD_ALT = 0x01, + MP_KEYMOD_SHIFT = 0x02, + MP_KEYMOD_CTRL = 0x04, + MP_KEYMOD_CMD = 0x08; + +typedef i32 mp_mouse_button; +static const mp_mouse_button MP_MOUSE_LEFT = 0x00, + MP_MOUSE_RIGHT = 0x01, + MP_MOUSE_MIDDLE = 0x02, + MP_MOUSE_EXT1 = 0x03, + MP_MOUSE_EXT2 = 0x04; + +static const u32 MP_KEY_COUNT = MP_KEY_MAX+1, + MP_MOUSE_BUTTON_COUNT = 5; + +typedef struct mp_key_event // keyboard and mouse buttons input +{ + mp_key_action action; + mp_key_code code; + mp_key_mods mods; + char label[8]; + u8 labelLen; + int clickCount; +} mp_key_event; + +typedef struct mp_char_event // character input +{ + utf32 codepoint; + char sequence[8]; + u8 seqLen; +} mp_char_event; + +typedef struct mp_move_event // mouse move/scroll +{ + f32 x; + f32 y; + f32 deltaX; + f32 deltaY; + mp_key_mods mods; +} mp_move_event; + +typedef struct mp_frame_event // window resize / move +{ + mp_rect rect; +} mp_frame_event; + +typedef struct mp_event +{ + //TODO clipboard and path drop + + mp_window window; + mp_event_type type; + + union + { + mp_key_event key; + mp_char_event character; + mp_move_event move; + mp_frame_event frame; + }; + + //TODO(martin): chain externally ? + list_elt list; +} mp_event; + +//-------------------------------------------------------------------- +// app management +//-------------------------------------------------------------------- + +void mp_init(); +void mp_terminate(); + +bool mp_should_quit(); +void mp_do_quit(); +void mp_request_quit(); +void mp_cancel_quit(); + +void mp_set_cursor(mp_mouse_cursor cursor); + +//-------------------------------------------------------------------- +// window management +//-------------------------------------------------------------------- + +//#include"graphics.h" + +bool mp_window_handle_is_null(mp_window window); +mp_window mp_window_null_handle(); + +mp_window mp_window_create(mp_rect contentRect, const char* title, mp_window_style style); +void mp_window_destroy(mp_window window); + +bool mp_window_should_close(mp_window window); +void mp_window_request_close(mp_window window); +void mp_window_cancel_close(mp_window window); + +void* mp_window_native_pointer(mp_window window); + +void mp_window_center(mp_window window); + +bool mp_window_is_hidden(mp_window window); +bool mp_window_is_focused(mp_window window); + +void mp_window_hide(mp_window window); +void mp_window_focus(mp_window window); +void mp_window_send_to_back(mp_window window); +void mp_window_bring_to_front(mp_window window); + +void mp_window_bring_to_front_and_focus(mp_window window); + +mp_rect mp_window_content_rect_for_frame_rect(mp_rect frameRect, mp_window_style style); +mp_rect mp_window_frame_rect_for_content_rect(mp_rect contentRect, mp_window_style style); + +mp_rect mp_window_get_content_rect(mp_window window); +mp_rect mp_window_get_absolute_content_rect(mp_window window); +mp_rect mp_window_get_frame_rect(mp_window window); + +void mp_window_set_content_rect(mp_window window, mp_rect contentRect); +void mp_window_set_frame_rect(mp_window window, mp_rect frameRect); +void mp_window_set_frame_size(mp_window window, int width, int height); +void mp_window_set_content_size(mp_window window, int width, int height); + +//-------------------------------------------------------------------- +// View management +//-------------------------------------------------------------------- + +mp_view mp_view_nil(); +bool mp_view_is_nil(mp_view view); +mp_view mp_view_create(mp_window window, mp_rect frame); +void mp_view_destroy(mp_view view); +void mp_view_set_frame(mp_view view, mp_rect frame); + +/*TODO +mp_view mp_view_bring_to_front(mp_view view); +mp_view mp_view_send_to_back(mp_view view); +*/ +//-------------------------------------------------------------------- +// Main loop throttle +//-------------------------------------------------------------------- + +void mp_set_target_fps(u32 fps); // or use wait vblank? + +//-------------------------------------------------------------------- +// Events handling +//-------------------------------------------------------------------- +void mp_pump_events(f64 timeout); +bool mp_next_event(mp_event* event); + +//-------------------------------------------------------------------- +// Input state polling +//-------------------------------------------------------------------- +bool mp_input_key_down(mp_key_code key); +bool mp_input_key_pressed(mp_key_code key); +bool mp_input_key_released(mp_key_code key); +mp_key_mods mp_input_key_mods(); + +str8 mp_key_to_label(mp_key_code key); +mp_key_code mp_label_to_key(str8 label); + +bool mp_input_mouse_down(mp_mouse_button button); +bool mp_input_mouse_pressed(mp_mouse_button button); +bool mp_input_mouse_released(mp_mouse_button button); +bool mp_input_mouse_clicked(mp_mouse_button button); +bool mp_input_mouse_double_clicked(mp_mouse_button button); + +vec2 mp_input_mouse_position(); +vec2 mp_input_mouse_delta(); +vec2 mp_input_mouse_wheel(); + +str32 mp_input_text_utf32(mem_arena* arena); +str8 mp_input_text_utf8(mem_arena* arena); +//-------------------------------------------------------------------- +// app resources +//-------------------------------------------------------------------- +int mp_app_get_resource_path(const char* name, char** result); +str8 mp_app_get_executable_path(mem_arena* arena); + +//-------------------------------------------------------------------- +// Clipboard +//-------------------------------------------------------------------- +void mp_clipboard_clear(); + +void mp_clipboard_set_string(str8 string); +str8 mp_clipboard_get_string(mem_arena* arena); +str8 mp_clipboard_copy_string(str8 backing); + +bool mp_clipboard_has_tag(const char* tag); +void mp_clipboard_set_data_for_tag(const char* tag, str8 data); +str8 mp_clipboard_get_data_for_tag(mem_arena* arena, const char* tag); + +//-------------------------------------------------------------------- +// native open/save/alert windows +//-------------------------------------------------------------------- + +str8 mp_open_dialog(mem_arena* arena, + const char* title, + const char* defaultPath, + int filterCount, + const char** filters, + bool directory); + +str8 mp_save_dialog(mem_arena* arena, + const char* title, + const char* defaultPath, + int filterCount, + const char** filters); + +int mp_alert_popup(const char* title, + const char* message, + u32 count, + const char** options); + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__PLATFORM_APP_H_ diff --git a/src/osx_app.h b/src/osx_app.h new file mode 100644 index 0000000..cd98b0b --- /dev/null +++ b/src/osx_app.h @@ -0,0 +1,58 @@ +/************************************************************//** +* +* @file: osx_app.h +* @author: Martin Fouilleul +* @date: 12/02/2021 +* @revision: +* +*****************************************************************/ +#ifndef __OSX_APP_H_ +#define __OSX_APP_H_ + +#import +#import +#include"mp_app.h" +#include"graphics.h" + +struct mp_window_data +{ + list_elt freeListElt; + u32 generation; + + NSWindow* nsWindow; + NSView* nsView; + NSObject* nsWindowDelegate; + + mp_rect contentRect; + mp_rect frameRect; + mp_window_style style; + + bool shouldClose; //TODO could be in status flags + bool hidden; + + mp_view mainView; +}; + +struct mp_view_data +{ + list_elt freeListElt; + u32 generation; + + mp_window window; + NSView* nsView; + mg_surface surface; +}; + +@interface MPNativeWindow : NSWindow +{ + mp_window_data* mpWindow; +} +- (id)initWithMPWindow:(mp_window_data*) window contentRect:(NSRect) rect styleMask:(uint32) style; +@end + +mp_window_data* mp_window_ptr_from_handle(mp_window handle); +mp_view_data* mp_view_ptr_from_handle(mp_view handle); + + + +#endif //__OSX_APP_H_ diff --git a/src/osx_app.mm b/src/osx_app.mm new file mode 100644 index 0000000..dabb0ec --- /dev/null +++ b/src/osx_app.mm @@ -0,0 +1,2401 @@ +//***************************************************************** +// +// $file: osx_app.cpp $ +// $author: Martin Fouilleul $ +// $date: 16/05/2020 $ +// $revision: $ +// $note: (C) 2020 by Martin Fouilleul - all rights reserved $ +// +//***************************************************************** + +#include // malloc/free + +#include"osx_app.h" +#include"lists.h" +#include"ringbuffer.h" +#include"memory.h" +#include"macro_helpers.h" +#include"platform_clock.h" + +#define LOG_SUBSYSTEM "Application" + +//-------------------------------------------------------------------- +// mp window struct and utility functions +//-------------------------------------------------------------------- + +static mp_rect mp_osx_to_user_screen_rect(mp_rect rect) +{ + @autoreleasepool + { + NSRect screenRect = [[NSScreen mainScreen] frame]; + rect.y = screenRect.size.height - rect.y - rect.h; + } + return(rect); +} + +static mp_rect mp_user_to_osx_screen_rect(mp_rect rect) +{ + @autoreleasepool + { + NSRect screenRect = [[NSScreen mainScreen] frame]; + rect.y = screenRect.size.height - rect.y - rect.h; + } + return(rect); +} + +static void mp_window_update_rect_cache(mp_window_data* window) +{ + @autoreleasepool + { + NSRect frame = [window->nsWindow frame]; + window->frameRect = mp_osx_to_user_screen_rect((mp_rect){frame.origin.x, frame.origin.y, frame.size.width, frame.size.height}); + + const NSRect contentRect = [[window->nsWindow contentView] frame]; + + window->contentRect = (mp_rect){ contentRect.origin.x, + contentRect.origin.y, + contentRect.size.width, + contentRect.size.height }; + + window->contentRect.y = window->frameRect.h - window->contentRect.y - window->contentRect.h; + } +} + +static u32 mp_osx_get_window_style_mask(mp_window_style style) +{ + u32 mask = 0; + if(style & MP_WINDOW_STYLE_NO_TITLE) + { + mask = NSWindowStyleMaskBorderless; + } + else + { + mask = NSWindowStyleMaskTitled; + } + + if(!(style & MP_WINDOW_STYLE_FIXED_SIZE)) + { + mask |= NSWindowStyleMaskResizable; + } + if(!(style & MP_WINDOW_STYLE_NO_CLOSE)) + { + mask |= NSWindowStyleMaskClosable; + } + if(!(style & MP_WINDOW_STYLE_NO_MINIFY)) + { + mask |= NSWindowStyleMaskMiniaturizable; + } + return(mask); +} + +//--------------------------------------------------------------- +// App struct and utility functions +//--------------------------------------------------------------- + +const u32 MP_APP_MAX_WINDOWS = 128; +const u32 MP_APP_MAX_VIEWS = 128; + +typedef struct mp_frame_stats +{ + f64 start; + f64 workTime; + f64 remainingTime; + f64 targetFramePeriod; +} mp_frame_stats; + +typedef struct mp_key_utf8 +{ + u8 labelLen; + char label[8]; +} mp_key_utf8; + +typedef struct mp_key_state +{ + u64 lastUpdate; + u32 transitionCounter; + bool down; + bool clicked; + bool doubleClicked; + +} mp_key_state; + +typedef struct mp_keyboard_state +{ + mp_key_state keys[MP_KEY_COUNT]; + mp_key_mods mods; +} mp_keyboard_state; + +typedef struct mp_mouse_state +{ + u64 lastUpdate; + vec2 pos; + vec2 delta; + vec2 wheel; + + union + { + mp_key_state buttons[MP_MOUSE_BUTTON_COUNT]; + struct + { + mp_key_state left; + mp_key_state right; + mp_key_state middle; + mp_key_state ext1; + mp_key_state ext2; + }; + }; +} mp_mouse_state; + +const u32 MP_INPUT_TEXT_BACKING_SIZE = 64; + +typedef struct mp_text_state +{ + u64 lastUpdate; + utf32 backing[MP_INPUT_TEXT_BACKING_SIZE]; + str32 codePoints; +} mp_text_state; + +typedef struct mp_input_state +{ + u64 frameCounter; + mp_keyboard_state keyboard; + mp_mouse_state mouse; + mp_text_state text; +} mp_input_state; + +struct mp_app_data +{ + bool init; + bool shouldQuit; + + mp_input_state inputState; + ringbuffer eventQueue; + + mp_frame_stats frameStats; + + NSTimer* frameTimer; + NSCursor* cursor; + mp_window_data windowPool[MP_APP_MAX_WINDOWS]; + list_info windowFreeList; + + mp_view_data viewPool[MP_APP_MAX_VIEWS]; + list_info viewFreeList; + + + TISInputSourceRef kbLayoutInputSource; + void* kbLayoutUnicodeData; + id kbLayoutListener; + mp_key_utf8 mpKeyToLabel[256]; + + int mpKeysToNative[MP_KEY_MAX]; + int nativeToMPKeys[256]; + +}; + +static mp_app_data __mpAppData = {}; + +//////////////////////////////////////////////////////////////////////////////////////////////////////// + +void mp_init_window_handles() +{ + ListInit(&__mpAppData.windowFreeList); + for(int i=0; i>32; + u32 generation = handle.h & 0xffffffff; + if(index >= MP_APP_MAX_WINDOWS) + { + return(0); + } + mp_window_data* window = &__mpAppData.windowPool[index]; + if(window->generation != generation) + { + return(0); + } + else + { + return(window); + } +} + +mp_window mp_window_handle_from_ptr(mp_window_data* window) +{ + DEBUG_ASSERT( (window - __mpAppData.windowPool) >= 0 + && (window - __mpAppData.windowPool) < MP_APP_MAX_WINDOWS); + + u64 h = ((u64)(window - __mpAppData.windowPool))<<32 + | ((u64)window->generation); + + return((mp_window){h}); +} + +void mp_window_recycle_ptr(mp_window_data* window) +{ + window->generation++; + ListPush(&__mpAppData.windowFreeList, &window->freeListElt); +} + + +void mp_init_view_handles() +{ + ListInit(&__mpAppData.viewFreeList); + for(int i=0; i>32; + u32 generation = handle.h & 0xffffffff; + if(index >= MP_APP_MAX_VIEWS) + { + return(0); + } + mp_view_data* view = &__mpAppData.viewPool[index]; + if(view->generation != generation) + { + return(0); + } + else + { + return(view); + } +} + +mp_view mp_view_handle_from_ptr(mp_view_data* view) +{ + DEBUG_ASSERT( (view - __mpAppData.viewPool) >= 0 + && (view - __mpAppData.viewPool) < MP_APP_MAX_VIEWS); + + u64 h = ((u64)(view - __mpAppData.viewPool))<<32 + | ((u64)view->generation); + + return((mp_view){h}); +} + +void mp_view_recycle_ptr(mp_view_data* view) +{ + view->generation++; + ListPush(&__mpAppData.viewFreeList, &view->freeListElt); +} + +/* +void mp_app_set_process_event_callback(mp_app_process_event_callback callback, void* userData) +{ + DEBUG_ASSERT(callback); + __mpAppData.userData = userData; + __mpAppData.processEvent = callback; +} +*/ +///////////////////////////////////////////////////////////////////////////////////////////////////////// + + +static void mp_init_osx_keys() +{ + memset(__mpAppData.nativeToMPKeys, MP_KEY_UNKNOWN, 256*sizeof(int)); + + __mpAppData.nativeToMPKeys[0x1D] = MP_KEY_0; + __mpAppData.nativeToMPKeys[0x12] = MP_KEY_1; + __mpAppData.nativeToMPKeys[0x13] = MP_KEY_2; + __mpAppData.nativeToMPKeys[0x14] = MP_KEY_3; + __mpAppData.nativeToMPKeys[0x15] = MP_KEY_4; + __mpAppData.nativeToMPKeys[0x17] = MP_KEY_5; + __mpAppData.nativeToMPKeys[0x16] = MP_KEY_6; + __mpAppData.nativeToMPKeys[0x1A] = MP_KEY_7; + __mpAppData.nativeToMPKeys[0x1C] = MP_KEY_8; + __mpAppData.nativeToMPKeys[0x19] = MP_KEY_9; + __mpAppData.nativeToMPKeys[0x00] = MP_KEY_A; + __mpAppData.nativeToMPKeys[0x0B] = MP_KEY_B; + __mpAppData.nativeToMPKeys[0x08] = MP_KEY_C; + __mpAppData.nativeToMPKeys[0x02] = MP_KEY_D; + __mpAppData.nativeToMPKeys[0x0E] = MP_KEY_E; + __mpAppData.nativeToMPKeys[0x03] = MP_KEY_F; + __mpAppData.nativeToMPKeys[0x05] = MP_KEY_G; + __mpAppData.nativeToMPKeys[0x04] = MP_KEY_H; + __mpAppData.nativeToMPKeys[0x22] = MP_KEY_I; + __mpAppData.nativeToMPKeys[0x26] = MP_KEY_J; + __mpAppData.nativeToMPKeys[0x28] = MP_KEY_K; + __mpAppData.nativeToMPKeys[0x25] = MP_KEY_L; + __mpAppData.nativeToMPKeys[0x2E] = MP_KEY_M; + __mpAppData.nativeToMPKeys[0x2D] = MP_KEY_N; + __mpAppData.nativeToMPKeys[0x1F] = MP_KEY_O; + __mpAppData.nativeToMPKeys[0x23] = MP_KEY_P; + __mpAppData.nativeToMPKeys[0x0C] = MP_KEY_Q; + __mpAppData.nativeToMPKeys[0x0F] = MP_KEY_R; + __mpAppData.nativeToMPKeys[0x01] = MP_KEY_S; + __mpAppData.nativeToMPKeys[0x11] = MP_KEY_T; + __mpAppData.nativeToMPKeys[0x20] = MP_KEY_U; + __mpAppData.nativeToMPKeys[0x09] = MP_KEY_V; + __mpAppData.nativeToMPKeys[0x0D] = MP_KEY_W; + __mpAppData.nativeToMPKeys[0x07] = MP_KEY_X; + __mpAppData.nativeToMPKeys[0x10] = MP_KEY_Y; + __mpAppData.nativeToMPKeys[0x06] = MP_KEY_Z; + + __mpAppData.nativeToMPKeys[0x27] = MP_KEY_APOSTROPHE; + __mpAppData.nativeToMPKeys[0x2A] = MP_KEY_BACKSLASH; + __mpAppData.nativeToMPKeys[0x2B] = MP_KEY_COMMA; + __mpAppData.nativeToMPKeys[0x18] = MP_KEY_EQUAL; + __mpAppData.nativeToMPKeys[0x32] = MP_KEY_GRAVE_ACCENT; + __mpAppData.nativeToMPKeys[0x21] = MP_KEY_LEFT_BRACKET; + __mpAppData.nativeToMPKeys[0x1B] = MP_KEY_MINUS; + __mpAppData.nativeToMPKeys[0x2F] = MP_KEY_PERIOD; + __mpAppData.nativeToMPKeys[0x1E] = MP_KEY_RIGHT_BRACKET; + __mpAppData.nativeToMPKeys[0x29] = MP_KEY_SEMICOLON; + __mpAppData.nativeToMPKeys[0x2C] = MP_KEY_SLASH; + __mpAppData.nativeToMPKeys[0x0A] = MP_KEY_WORLD_1; + + __mpAppData.nativeToMPKeys[0x33] = MP_KEY_BACKSPACE; + __mpAppData.nativeToMPKeys[0x39] = MP_KEY_CAPS_LOCK; + __mpAppData.nativeToMPKeys[0x75] = MP_KEY_DELETE; + __mpAppData.nativeToMPKeys[0x7D] = MP_KEY_DOWN; + __mpAppData.nativeToMPKeys[0x77] = MP_KEY_END; + __mpAppData.nativeToMPKeys[0x24] = MP_KEY_ENTER; + __mpAppData.nativeToMPKeys[0x35] = MP_KEY_ESCAPE; + __mpAppData.nativeToMPKeys[0x7A] = MP_KEY_F1; + __mpAppData.nativeToMPKeys[0x78] = MP_KEY_F2; + __mpAppData.nativeToMPKeys[0x63] = MP_KEY_F3; + __mpAppData.nativeToMPKeys[0x76] = MP_KEY_F4; + __mpAppData.nativeToMPKeys[0x60] = MP_KEY_F5; + __mpAppData.nativeToMPKeys[0x61] = MP_KEY_F6; + __mpAppData.nativeToMPKeys[0x62] = MP_KEY_F7; + __mpAppData.nativeToMPKeys[0x64] = MP_KEY_F8; + __mpAppData.nativeToMPKeys[0x65] = MP_KEY_F9; + __mpAppData.nativeToMPKeys[0x6D] = MP_KEY_F10; + __mpAppData.nativeToMPKeys[0x67] = MP_KEY_F11; + __mpAppData.nativeToMPKeys[0x6F] = MP_KEY_F12; + __mpAppData.nativeToMPKeys[0x69] = MP_KEY_F13; + __mpAppData.nativeToMPKeys[0x6B] = MP_KEY_F14; + __mpAppData.nativeToMPKeys[0x71] = MP_KEY_F15; + __mpAppData.nativeToMPKeys[0x6A] = MP_KEY_F16; + __mpAppData.nativeToMPKeys[0x40] = MP_KEY_F17; + __mpAppData.nativeToMPKeys[0x4F] = MP_KEY_F18; + __mpAppData.nativeToMPKeys[0x50] = MP_KEY_F19; + __mpAppData.nativeToMPKeys[0x5A] = MP_KEY_F20; + __mpAppData.nativeToMPKeys[0x73] = MP_KEY_HOME; + __mpAppData.nativeToMPKeys[0x72] = MP_KEY_INSERT; + __mpAppData.nativeToMPKeys[0x7B] = MP_KEY_LEFT; + __mpAppData.nativeToMPKeys[0x3A] = MP_KEY_LEFT_ALT; + __mpAppData.nativeToMPKeys[0x3B] = MP_KEY_LEFT_CONTROL; + __mpAppData.nativeToMPKeys[0x38] = MP_KEY_LEFT_SHIFT; + __mpAppData.nativeToMPKeys[0x37] = MP_KEY_LEFT_SUPER; + __mpAppData.nativeToMPKeys[0x6E] = MP_KEY_MENU; + __mpAppData.nativeToMPKeys[0x47] = MP_KEY_NUM_LOCK; + __mpAppData.nativeToMPKeys[0x79] = MP_KEY_PAGE_DOWN; + __mpAppData.nativeToMPKeys[0x74] = MP_KEY_PAGE_UP; + __mpAppData.nativeToMPKeys[0x7C] = MP_KEY_RIGHT; + __mpAppData.nativeToMPKeys[0x3D] = MP_KEY_RIGHT_ALT; + __mpAppData.nativeToMPKeys[0x3E] = MP_KEY_RIGHT_CONTROL; + __mpAppData.nativeToMPKeys[0x3C] = MP_KEY_RIGHT_SHIFT; + __mpAppData.nativeToMPKeys[0x36] = MP_KEY_RIGHT_SUPER; + __mpAppData.nativeToMPKeys[0x31] = MP_KEY_SPACE; + __mpAppData.nativeToMPKeys[0x30] = MP_KEY_TAB; + __mpAppData.nativeToMPKeys[0x7E] = MP_KEY_UP; + + __mpAppData.nativeToMPKeys[0x52] = MP_KEY_KP_0; + __mpAppData.nativeToMPKeys[0x53] = MP_KEY_KP_1; + __mpAppData.nativeToMPKeys[0x54] = MP_KEY_KP_2; + __mpAppData.nativeToMPKeys[0x55] = MP_KEY_KP_3; + __mpAppData.nativeToMPKeys[0x56] = MP_KEY_KP_4; + __mpAppData.nativeToMPKeys[0x57] = MP_KEY_KP_5; + __mpAppData.nativeToMPKeys[0x58] = MP_KEY_KP_6; + __mpAppData.nativeToMPKeys[0x59] = MP_KEY_KP_7; + __mpAppData.nativeToMPKeys[0x5B] = MP_KEY_KP_8; + __mpAppData.nativeToMPKeys[0x5C] = MP_KEY_KP_9; + __mpAppData.nativeToMPKeys[0x45] = MP_KEY_KP_ADD; + __mpAppData.nativeToMPKeys[0x41] = MP_KEY_KP_DECIMAL; + __mpAppData.nativeToMPKeys[0x4B] = MP_KEY_KP_DIVIDE; + __mpAppData.nativeToMPKeys[0x4C] = MP_KEY_KP_ENTER; + __mpAppData.nativeToMPKeys[0x51] = MP_KEY_KP_EQUAL; + __mpAppData.nativeToMPKeys[0x43] = MP_KEY_KP_MULTIPLY; + __mpAppData.nativeToMPKeys[0x4E] = MP_KEY_KP_SUBTRACT; + + memset(__mpAppData.mpKeysToNative, 0, sizeof(int)*MP_KEY_COUNT); + for(int nativeKey=0; nativeKey<256; nativeKey++) + { + mp_key_code mpKey = __mpAppData.nativeToMPKeys[nativeKey]; + if(mpKey) + { + __mpAppData.mpKeysToNative[mpKey] = nativeKey; + } + } +} + +static int mp_convert_osx_key(unsigned short nsCode) +{ + if(nsCode >= 265) + { + return(MP_KEY_UNKNOWN); + } + else + { + return(__mpAppData.nativeToMPKeys[nsCode]); + } +} + +static mp_key_mods mp_convert_osx_mods(NSUInteger nsFlags) +{ + mp_key_mods mods = MP_KEYMOD_NONE; + if(nsFlags & NSEventModifierFlagShift) + { + mods |= MP_KEYMOD_SHIFT; + } + if(nsFlags & NSEventModifierFlagControl) + { + mods |= MP_KEYMOD_CTRL; + } + if(nsFlags & NSEventModifierFlagOption) + { + mods |= MP_KEYMOD_ALT; + } + if(nsFlags & NSEventModifierFlagCommand) + { + mods |= MP_KEYMOD_CMD; + } + return(mods); +} + +static void mp_update_keyboard_layout() +{ + if(__mpAppData.kbLayoutInputSource) + { + CFRelease(__mpAppData.kbLayoutInputSource); + __mpAppData.kbLayoutInputSource = 0; + __mpAppData.kbLayoutUnicodeData = nil; + } + + __mpAppData.kbLayoutInputSource = TISCopyCurrentKeyboardLayoutInputSource(); + if(!__mpAppData.kbLayoutInputSource) + { + LOG_ERROR("Failed to load keyboard layout input source"); + } + + __mpAppData.kbLayoutUnicodeData = TISGetInputSourceProperty(__mpAppData.kbLayoutInputSource, + kTISPropertyUnicodeKeyLayoutData); + if(!__mpAppData.kbLayoutUnicodeData) + { + LOG_ERROR("Failed to load keyboard layout unicode data"); + } + + memset(__mpAppData.mpKeyToLabel, 0, sizeof(mp_key_utf8)*MP_KEY_COUNT); + + for(int key=0; keylabelLen, keyInfo->label); + return(label); +} + +mp_key_code mp_label_to_key(str8 label) +{ + mp_key_code res = MP_KEY_UNKNOWN; + for(int key=0; keylastUpdate != frameCounter) + { + key->transitionCounter = 0; + key->clicked = false; + key->doubleClicked = false; + key->lastUpdate = frameCounter; + } + + if(key->down != down) + { + key->transitionCounter++; + } + key->down = down; +} + +static void mp_update_mouse_move(f32 x, f32 y, f32 deltaX, f32 deltaY) +{ + u64 frameCounter = __mpAppData.inputState.frameCounter; + mp_mouse_state* mouse = &__mpAppData.inputState.mouse; + if(mouse->lastUpdate != frameCounter) + { + mouse->delta = (vec2){0, 0}; + mouse->wheel = (vec2){0, 0}; + mouse->lastUpdate = frameCounter; + } + mouse->pos = (vec2){x, y}; + mouse->delta.x += deltaX; + mouse->delta.y += deltaY; +} + +static void mp_update_mouse_wheel(f32 deltaX, f32 deltaY) +{ + u64 frameCounter = __mpAppData.inputState.frameCounter; + mp_mouse_state* mouse = &__mpAppData.inputState.mouse; + if(mouse->lastUpdate != frameCounter) + { + mouse->delta = (vec2){0, 0}; + mouse->wheel = (vec2){0, 0}; + mouse->lastUpdate = frameCounter; + } + mouse->wheel.x += deltaX; + mouse->wheel.y += deltaY; +} + +static void mp_update_text(utf32 codepoint) +{ + u64 frameCounter = __mpAppData.inputState.frameCounter; + mp_text_state* text = &__mpAppData.inputState.text; + + if(text->lastUpdate != frameCounter) + { + text->codePoints.len = 0; + } + text->codePoints.ptr = text->backing; + if(text->codePoints.len < MP_INPUT_TEXT_BACKING_SIZE) + { + text->codePoints.ptr[text->codePoints.len] = codepoint; + text->codePoints.len++; + } + else + { + LOG_WARNING("too many input codepoints per frame, dropping input"); + } +} + +static void mp_queue_event(mp_event* event) +{ + if(ringbuffer_write_available(&__mpAppData.eventQueue) < sizeof(mp_event)) + { + LOG_ERROR("event queue full\n"); + } + else + { + u32 written = ringbuffer_write(&__mpAppData.eventQueue, sizeof(mp_event), (u8*)event); + DEBUG_ASSERT(written == sizeof(mp_event)); + } +} + +//--------------------------------------------------------------- +// Application and app delegate +//--------------------------------------------------------------- + +@interface MPApplication : NSApplication +@end + +@implementation MPApplication +-(void)noOpThread:(id)object +{} +@end + +@interface MPAppDelegate : NSObject +@end + +@implementation MPAppDelegate + +- (NSApplicationTerminateReply)applicationShouldTerminate:(NSApplication *)sender +{ + //NOTE: We set shouldQuit to true and send a Quit event + // We then return a value to cancel the direct termination because we still + // want to execte the code after mp_event_loop(). If the user didn't set shouldQuit to + // false, mp_event_loop() will exit, and the user can execute any cleanup needed and + // exit the program. + + __mpAppData.shouldQuit = true; + mp_event event = {}; + event.type = MP_EVENT_QUIT; + mp_queue_event(&event); + + return(NSTerminateCancel); +} + +- (void)applicationWillFinishLaunching:(NSNotification *)notification +{@autoreleasepool{ + + //NOTE(martin): add a menu for quit, and a corresponding key equivalent. + // this allows to quit the application when there is no window + // left to catch our Cmd-Q key equivalent + NSMenu* bar = [[NSMenu alloc] init]; + [NSApp setMainMenu:bar]; + + NSMenuItem* appMenuItem = + [bar addItemWithTitle:@"" action:NULL keyEquivalent:@""]; + NSMenu* appMenu = [[NSMenu alloc] init]; + [appMenuItem setSubmenu:appMenu]; + + [appMenu addItemWithTitle: @"Quit" + action: @selector(terminate:) + keyEquivalent: @"q"]; +}} + +- (void)timerElapsed:(NSTimer*)timer +{ + mp_event event = {}; + event.type = MP_EVENT_FRAME; + mp_queue_event(&event); +} + +- (void)applicationDidFinishLaunching:(NSNotification *)notification +{@autoreleasepool{ + //WARN(martin): the order of these calls seem to matter a lot for properly showing the menu bar + // with other orderings, the menu doesn't display before the application is put out of + // focus and on focus again... This is flaky undocumented behaviour, so although it is + // fixed by the current ordering, we expect the problem to show up again in future + // versions of macOS. + + //NOTE(martin): send a dummy event to wake-up the run loop and exit from the run loop. + NSEvent* event = [NSEvent otherEventWithType:NSEventTypeApplicationDefined + location:NSMakePoint(0, 0) + modifierFlags:0 + timestamp:0 + windowNumber:0 + context:nil + subtype:0 + data1:0 + data2:0]; + [NSApp postEvent:event atStart:YES]; + [NSApp stop:nil]; +}} + +@end // @implementation MPAppDelegate + +//--------------------------------------------------------------- +// Custom NSWindow +//--------------------------------------------------------------- + +@implementation MPNativeWindow +- (id)initWithMPWindow:(mp_window_data*) window contentRect:(NSRect) rect styleMask:(uint32) style +{ + mpWindow = window; + return([self initWithContentRect:rect styleMask:style backing:NSBackingStoreBuffered defer:NO]); +} +- (BOOL)canBecomeKeyWindow +{ + return(!(mpWindow->style & MP_WINDOW_STYLE_NO_FOCUS)); +} +@end //@implementation MPNativeWindow + +//--------------------------------------------------------------- +// Custom NSWindow delegate +//--------------------------------------------------------------- + +@interface MPNativeWindowDelegate : NSObject +{ + mp_window_data* mpWindow; +} +- (id)initWithMPWindow:(mp_window_data*) window; +@end + +@implementation MPNativeWindowDelegate + +- (id)initWithMPWindow:(mp_window_data*) window +{ + self = [super init]; + if(self != nil) + { + mpWindow = window; + } + return(self); +} + +- (void)windowDidBecomeKey:(NSNotification*)notification +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(mpWindow); + event.type = MP_EVENT_WINDOW_FOCUS; + + mpWindow->hidden = false; + + mp_queue_event(&event); +} + +- (void)windowDidResignKey:(NSNotification*)notification +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(mpWindow); + event.type = MP_EVENT_WINDOW_UNFOCUS; + + mp_queue_event(&event); +} + +- (void)windowDidMove:(NSNotification *)notification +{ + const NSRect contentRect = [[mpWindow->nsWindow contentView] frame]; + + mp_window_update_rect_cache(mpWindow); + + mp_event event = {}; + event.window = mp_window_handle_from_ptr(mpWindow); + event.type = MP_EVENT_WINDOW_MOVE; + event.frame.rect.x = contentRect.origin.x; + event.frame.rect.y = contentRect.origin.y; + event.frame.rect.w = contentRect.size.width; + event.frame.rect.h = contentRect.size.height; + + mp_queue_event(&event); +} + +- (void)windowDidResize:(NSNotification *)notification +{ + const NSRect contentRect = [[mpWindow->nsWindow contentView] frame]; + + mp_window_update_rect_cache(mpWindow); + + mp_event event = {}; + event.window = mp_window_handle_from_ptr(mpWindow); + event.type = MP_EVENT_WINDOW_RESIZE; + event.frame.rect.x = contentRect.origin.x; + event.frame.rect.y = contentRect.origin.y; + event.frame.rect.w = contentRect.size.width; + event.frame.rect.h = contentRect.size.height; + + mp_rect viewFrame = {0, 0, contentRect.size.width, contentRect.size.height}; + mp_view_set_frame(mpWindow->mainView, viewFrame); + + mp_queue_event(&event); +} + +- (void)windowWillClose:(NSNotification *)notification +{ + mpWindow->nsWindow = nil; + [mpWindow->nsView release]; + mpWindow->nsView = nil; + [mpWindow->nsWindowDelegate release]; + mpWindow->nsWindowDelegate = nil; + + mp_window_recycle_ptr(mpWindow); +} + +- (BOOL)windowShouldClose:(id)sender +{ + mpWindow->shouldClose = true; + + mp_event event = {}; + event.window = mp_window_handle_from_ptr(mpWindow); + event.type = MP_EVENT_WINDOW_CLOSE; + + mp_queue_event(&event); + + return(mpWindow->shouldClose); +} + +@end //@implementation MPNativeWindowDelegate + +//--------------------------------------------------------------- +// Custom NSView +//--------------------------------------------------------------- + +@interface MPNativeView : NSView +{ + mp_window_data* window; + NSTrackingArea* trackingArea; + NSMutableAttributedString* markedText; +} +- (id)initWithMPWindow:(mp_window_data*) mpWindow; +@end + +@implementation MPNativeView + +- (id)initWithMPWindow:(mp_window_data*) mpWindow +{ + self = [super init]; + if(self != nil) + { + window = mpWindow; + mpWindow->nsView = self; + + NSTrackingAreaOptions trackingOptions = NSTrackingMouseEnteredAndExited + | NSTrackingMouseMoved + | NSTrackingCursorUpdate + | NSTrackingActiveInActiveApp //TODO maybe change that to allow multi-window mouse events... + | NSTrackingEnabledDuringMouseDrag + | NSTrackingInVisibleRect + | NSTrackingAssumeInside ; + + trackingArea = [[NSTrackingArea alloc] initWithRect:[self bounds] options:trackingOptions owner:self userInfo:nil]; + [self addTrackingArea:trackingArea]; + markedText = [[NSMutableAttributedString alloc] init]; + } + return(self); +} + +- (void)dealloc +{ + [trackingArea release]; + [markedText release]; + [super dealloc]; +} + +-(void)drawRect:(NSRect)dirtyRect +{ + if(window->style & MP_WINDOW_STYLE_NO_TITLE) + { + [NSGraphicsContext saveGraphicsState]; + NSBezierPath *path = [NSBezierPath bezierPathWithRoundedRect:[self frame] xRadius:5 yRadius:5]; + [path addClip]; + [[NSColor whiteColor] set]; + NSRectFill([self frame]); + } + + if(window->style & MP_WINDOW_STYLE_NO_TITLE) + { + [NSGraphicsContext restoreGraphicsState]; + [window->nsWindow invalidateShadow]; + } +} + +- (BOOL)acceptsFirstReponder +{ + return(YES); +} + +- (void)cursorUpdate:(NSEvent*)event +{ + if(__mpAppData.cursor) + { + [__mpAppData.cursor set]; + } + else + { + [[NSCursor arrowCursor] set]; + } +} + +- (void)mouseDown:(NSEvent *)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_BUTTON; + event.key.action = MP_KEY_PRESS; + event.key.code = MP_MOUSE_LEFT; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + event.key.clickCount = [nsEvent clickCount]; + + mp_update_key_state(&__mpAppData.inputState.mouse.buttons[event.key.code], true); + if(event.key.clickCount >= 1) + { + __mpAppData.inputState.mouse.buttons[event.key.code].clicked = true; + } + if(event.key.clickCount >= 2) + { + __mpAppData.inputState.mouse.buttons[event.key.code].doubleClicked = true; + } + + mp_queue_event(&event); + + [window->nsWindow makeFirstResponder:self]; +} + +- (void)mouseUp:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_BUTTON; + event.key.action = MP_KEY_RELEASE; + event.key.code = MP_MOUSE_LEFT; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + event.key.clickCount = [nsEvent clickCount]; + + mp_update_key_state(&__mpAppData.inputState.mouse.buttons[event.key.code], false); + + mp_queue_event(&event); +} + +- (void)rightMouseDown:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_BUTTON; + event.key.action = MP_KEY_PRESS; + event.key.code = MP_MOUSE_RIGHT; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_key_state(&__mpAppData.inputState.mouse.buttons[event.key.code], true); + + mp_queue_event(&event); +} + +- (void)rightMouseUp:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_BUTTON; + event.key.action = MP_KEY_RELEASE; + event.key.code = MP_MOUSE_RIGHT; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_key_state(&__mpAppData.inputState.mouse.buttons[event.key.code], false); + + mp_queue_event(&event); +} + +- (void)otherMouseDown:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_BUTTON; + event.key.action = MP_KEY_PRESS; + event.key.code = [nsEvent buttonNumber]; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_key_state(&__mpAppData.inputState.mouse.buttons[event.key.code], true); + + mp_queue_event(&event); +} + +- (void)otherMouseUp:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_BUTTON; + event.key.action = MP_KEY_RELEASE; + event.key.code = [nsEvent buttonNumber]; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_key_state(&__mpAppData.inputState.mouse.buttons[event.key.code], false); + + mp_queue_event(&event); +} + +- (void)mouseDragged:(NSEvent*)nsEvent +{ + [self mouseMoved:nsEvent]; +} + +- (void)mouseMoved:(NSEvent*)nsEvent +{ + NSPoint p = [self convertPoint:[nsEvent locationInWindow] fromView:nil]; + + NSRect frame = [[window->nsWindow contentView] frame]; + mp_event event = {}; + event.type = MP_EVENT_MOUSE_MOVE; + event.window = mp_window_handle_from_ptr(window); + event.move.x = p.x; + event.move.y = p.y; + event.move.deltaX = [nsEvent deltaX]; + event.move.deltaY = -[nsEvent deltaY]; + event.move.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_mouse_move(p.x, p.y, event.move.deltaX, event.move.deltaY); + + mp_queue_event(&event); +} + +- (void)scrollWheel:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_WHEEL; + + double factor = [nsEvent hasPreciseScrollingDeltas] ? 0.1 : 1.0; + event.move.x = 0; + event.move.y = 0; + event.move.deltaX = [nsEvent scrollingDeltaX]*factor; + event.move.deltaY = -[nsEvent scrollingDeltaY]*factor; + event.move.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_mouse_wheel(event.move.deltaX, event.move.deltaY); + + mp_queue_event(&event); +} + +- (void)mouseExited:(NSEvent *)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_LEAVE; + mp_queue_event(&event); +} + +- (void)mouseEntered:(NSEvent *)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_MOUSE_ENTER; + mp_queue_event(&event); +} + + + +- (void)keyDown:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_KEYBOARD_KEY; + event.key.action = MP_KEY_PRESS; + event.key.code = mp_convert_osx_key([nsEvent keyCode]); + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + str8 label = mp_key_to_label(event.key.code); + event.key.labelLen = label.len; + memcpy(event.key.label, label.ptr, label.len); + + mp_update_key_state(&__mpAppData.inputState.keyboard.keys[event.key.code], true); + + mp_queue_event(&event); + + [self interpretKeyEvents:@[nsEvent]]; +} + +- (void)keyUp:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_KEYBOARD_KEY; + event.key.action = MP_KEY_RELEASE; + event.key.code = mp_convert_osx_key([nsEvent keyCode]); + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_key_state(&__mpAppData.inputState.keyboard.keys[event.key.code], false); + + mp_queue_event(&event); +} + +- (void) flagsChanged:(NSEvent*)nsEvent +{ + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_KEYBOARD_MODS; + event.key.mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + mp_update_key_mods(event.key.mods); + + mp_queue_event(&event); +} + +- (BOOL)performKeyEquivalent:(NSEvent*)nsEvent +{ + if([nsEvent modifierFlags] & NSEventModifierFlagCommand) + { + if([nsEvent charactersIgnoringModifiers] == [NSString stringWithUTF8String:"w"]) + { + [window->nsWindow performClose:self]; + return(YES); + } + else if([nsEvent charactersIgnoringModifiers] == [NSString stringWithUTF8String:"q"]) + { + __mpAppData.shouldQuit = true; + + mp_event event = {}; + event.type = MP_EVENT_QUIT; + + mp_queue_event(&event); + + //[NSApp terminate:self]; + return(YES); + } + } + + return([super performKeyEquivalent:nsEvent]); +} + +- (BOOL)hasMarkedText +{ + return([markedText length] > 0); +} + +static const NSRange kEmptyRange = { NSNotFound, 0 }; + +- (NSRange)markedRange +{ + if([markedText length] > 0) + { + return(NSMakeRange(0, [markedText length] - 1)); + } + else + { + return(kEmptyRange); + } +} + +- (NSRange)selectedRange +{ + return(kEmptyRange); +} + +- (void)setMarkedText:(id)string + selectedRange:(NSRange)selectedRange + replacementRange:(NSRange)replacementRange +{ + [markedText release]; + + if([string isKindOfClass:[NSAttributedString class]]) + { + markedText = [[NSMutableAttributedString alloc] initWithAttributedString:string]; + } + else + { + markedText = [[NSMutableAttributedString alloc] initWithString:string]; + } +} + +- (void)unmarkText +{ + [[markedText mutableString] setString:@""]; +} + +- (NSArray*)validAttributesForMarkedText +{ + return([NSArray array]); +} + +- (NSAttributedString*)attributedSubstringForProposedRange:(NSRange)range + actualRange:(NSRangePointer)actualRange +{ + return(nil); +} + +- (NSUInteger)characterIndexForPoint:(NSPoint)point +{ + return(0); +} + +- (NSRect)firstRectForCharacterRange:(NSRange)range + actualRange:(NSRangePointer)actualRange +{ + NSRect frame = [window->nsView frame]; + return(NSMakeRect(frame.origin.x, frame.origin.y, 0.0, 0.0)); +} + +- (void)insertText:(id)string replacementRange:(NSRange)replacementRange +{ + NSString* characters; + NSEvent* nsEvent = [NSApp currentEvent]; + mp_key_mods mods = mp_convert_osx_mods([nsEvent modifierFlags]); + + if([string isKindOfClass:[NSAttributedString class]]) + { + characters = [string string]; + } + else + { + characters = (NSString*) string; + } + + NSRange range = NSMakeRange(0, [characters length]); + while (range.length) + { + utf32 codepoint = 0; + + if ([characters getBytes:&codepoint + maxLength:sizeof(codepoint) + usedLength:NULL + encoding:NSUTF32StringEncoding + options:0 + range:range + remainingRange:&range]) + { + if(codepoint >= 0xf700 && codepoint <= 0xf7ff) + { + continue; + } + + mp_event event = {}; + event.window = mp_window_handle_from_ptr(window); + event.type = MP_EVENT_KEYBOARD_CHAR; + event.character.codepoint = codepoint; + + str8 seq = utf8_encode(event.character.sequence, event.character.codepoint); + event.character.seqLen = seq.len; + + mp_update_text(codepoint); + + mp_queue_event(&event); + } + } + [self unmarkText]; +} + +- (void)doCommandBySelector:(SEL)selector +{ +} + +@end //@implementation MPNativeView + +/* +void mp_sleep_nanoseconds(u64 nanoseconds) +{ + timespec rqtp; + rqtp.tv_sec = nanoseconds / 1000000000; + rqtp.tv_nsec = nanoseconds - rqtp.tv_sec * 1000000000; + nanosleep(&rqtp, 0); +} + +static mach_timebase_info_data_t __machTimeBase__ = {1,1}; + +u64 mp_get_elapsed_nanoseconds() +{ + //NOTE(martin): according to the documentation, mach_absolute_time() does not + // increment when the system is asleep + u64 now = mach_absolute_time(); + now *= __machTimeBase__.numer; + now /= __machTimeBase__.denom; + return(now); +} + +f64 mp_get_elapsed_seconds() +{ + return(1.e-9*(f64)mp_get_elapsed_nanoseconds()); +} +*/ + +//*************************************************************** +// public API +//*************************************************************** + +//--------------------------------------------------------------- +// App public API +//--------------------------------------------------------------- + +void mp_init() +{@autoreleasepool { + if(!__mpAppData.init) + { + memset(&__mpAppData, 0, sizeof(__mpAppData)); + + mp_clock_init(); + + LOG_MESSAGE("init keys\n"); + mp_init_osx_keys(); + mp_update_keyboard_layout(); + mp_install_keyboard_layout_listener(); + + LOG_MESSAGE("init handles\n"); + mp_init_window_handles(); + mp_init_view_handles(); + + LOG_MESSAGE("init event queue\n"); + ringbuffer_init(&__mpAppData.eventQueue, 16); + + [MPApplication sharedApplication]; + MPAppDelegate* delegate = [[MPAppDelegate alloc] init]; + [NSApp setDelegate: delegate]; + + [NSThread detachNewThreadSelector:@selector(noOpThread:) + toTarget:NSApp + withObject:nil]; + + __mpAppData.init = true; + + LOG_MESSAGE("run application\n"); + [NSApp run]; + + [NSApp setActivationPolicy:NSApplicationActivationPolicyRegular]; + [NSApp activateIgnoringOtherApps:YES]; + + } +}} + +void mp_terminate() +{ + //TODO: proper app data cleanup (eg delegate, etc) + if(__mpAppData.init) + { + __mpAppData = (mp_app_data){0}; + } +} + +bool mp_should_quit() +{ + return(__mpAppData.shouldQuit); +} + +void mp_do_quit() +{ + __mpAppData.shouldQuit = true; +} + +void mp_cancel_quit() +{ + __mpAppData.shouldQuit = false; +} + +void mp_request_quit() +{ + __mpAppData.shouldQuit = true; + mp_event event = {}; + event.type = MP_EVENT_QUIT; + mp_queue_event(&event); +} + +void mp_set_cursor(mp_mouse_cursor cursor) +{ + switch(cursor) + { + case MP_MOUSE_CURSOR_ARROW: + { + __mpAppData.cursor = [NSCursor arrowCursor]; + } break; + case MP_MOUSE_CURSOR_RESIZE_0: + { + __mpAppData.cursor = [[NSCursor class] performSelector:@selector(_windowResizeEastWestCursor)]; + } break; + case MP_MOUSE_CURSOR_RESIZE_90: + { + __mpAppData.cursor = [[NSCursor class] performSelector:@selector(_windowResizeNorthSouthCursor)]; + } break; + case MP_MOUSE_CURSOR_RESIZE_45: + { + __mpAppData.cursor = [[NSCursor class] performSelector:@selector(_windowResizeNorthEastSouthWestCursor)]; + } break; + case MP_MOUSE_CURSOR_RESIZE_135: + { + __mpAppData.cursor = [[NSCursor class] performSelector:@selector(_windowResizeNorthWestSouthEastCursor)]; + } break; + case MP_MOUSE_CURSOR_TEXT: + { + __mpAppData.cursor = [NSCursor IBeamCursor]; + } break; + } + [__mpAppData.cursor set]; +} + +void mp_clipboard_clear() +{@autoreleasepool{ + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + [pb clearContents]; +}} + +void mp_clipboard_set_string(str8 string) +{@autoreleasepool{ + + NSString* nsString = [[NSString alloc] initWithBytes:string.ptr length:string.len encoding:NSUTF8StringEncoding]; + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + [pb writeObjects:[[NSArray alloc] initWithObjects:nsString, nil]]; +}} + +str8 mp_clipboard_copy_string(str8 backing) +{@autoreleasepool{ + //WARN(martin): maxSize includes space for a null terminator + + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + NSString* nsString = [pb stringForType:NSPasteboardTypeString]; + const char* cString = [nsString UTF8String]; + u32 len = minimum(backing.len-1, strlen(cString)); //length without null terminator + strncpy(backing.ptr, cString, backing.len-1); + backing.ptr[len] = '\0'; + + str8 result = str8_slice(backing, 0, len); + return(result); +}} + +str8 mp_clipboard_get_string(mem_arena* arena) +{@autoreleasepool{ + //WARN(martin): maxSize includes space for a null terminator + + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + NSString* nsString = [pb stringForType:NSPasteboardTypeString]; + const char* cString = [nsString UTF8String]; + str8 result = str8_push_cstring(arena, cString); + return(result); +}} + +bool mp_clipboard_has_tag(const char* tag) +{@autoreleasepool{ + + NSString* tagString = [[NSString alloc] initWithUTF8String: tag]; + NSArray* tagArray = [NSArray arrayWithObjects: tagString, nil]; + + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + NSString* available = [pb availableTypeFromArray: tagArray]; + + return(available != nil); +}} + +void mp_clipboard_set_data_for_tag(const char* tag, str8 string) +{@autoreleasepool{ + + NSString* tagString = [[NSString alloc] initWithUTF8String: tag]; + NSArray* tagArray = [NSArray arrayWithObjects: tagString, nil]; + NSData* nsData = [NSData dataWithBytes:string.ptr length:string.len]; + + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + [pb addTypes: tagArray owner:nil]; + [pb setData: nsData forType: tagString]; +}} + +str8 mp_clipboard_get_data_for_tag(mem_arena* arena, const char* tag) +{@autoreleasepool{ + + NSString* tagString = [[NSString alloc] initWithUTF8String: tag]; + + NSPasteboard* pb = [NSPasteboard generalPasteboard]; + NSData* nsData = [pb dataForType: tagString]; + str8 result = str8_push_buffer(arena, [nsData length], (char*)[nsData bytes]); + return(result); +}} + + +//--------------------------------------------------------------- +// Window public API +//--------------------------------------------------------------- + +bool mp_window_handle_is_null(mp_window window) +{ + return(window.h == 0); +} + +mp_window mp_window_null_handle() +{ + return((mp_window){.h = 0}); +} + +/* +//TODO(martin): review include scheme +extern "C" { + mp_graphics_surface mp_metal_surface_create_for_window_ptr(mp_window_data* window); + mp_graphics_surface mp_graphics_surface_null_handle(); + mp_graphics_surface mp_graphics_surface_handle_from_ptr(mp_graphics_surface_data* surface); +} +*/ + +mp_window mp_window_create(mp_rect contentRect, const char* title, mp_window_style style) +{@autoreleasepool{ + mp_window_data* window = mp_window_alloc(); + if(!window) + { + LOG_ERROR("Could not allocate window data\n"); + return((mp_window){0}); + } + + window->style = style; + window->shouldClose = false; + window->hidden = true; + + u32 styleMask = mp_osx_get_window_style_mask(style); + + NSRect screenRect = [[NSScreen mainScreen] frame]; + NSRect rect = NSMakeRect(contentRect.x, + screenRect.size.height - contentRect.y - contentRect.h, + contentRect.w, + contentRect.h); + + window->nsWindow = [[MPNativeWindow alloc] initWithMPWindow: window contentRect:rect styleMask:styleMask]; + window->nsWindowDelegate = [[MPNativeWindowDelegate alloc] initWithMPWindow:window]; + + [window->nsWindow setDelegate:(id)window->nsWindowDelegate]; + [window->nsWindow setTitle:[NSString stringWithUTF8String:title]]; + + if(style & MP_WINDOW_STYLE_NO_TITLE) + { + [window->nsWindow setOpaque:NO]; + [window->nsWindow setBackgroundColor:[NSColor clearColor]]; + [window->nsWindow setHasShadow:YES]; + } + if(style & MP_WINDOW_STYLE_FLOAT) + { + [window->nsWindow setLevel:NSFloatingWindowLevel]; + [window->nsWindow setHidesOnDeactivate:YES]; + } + if(style & MP_WINDOW_STYLE_NO_BUTTONS) + { + [[window->nsWindow standardWindowButton:NSWindowCloseButton] setHidden:YES]; + [[window->nsWindow standardWindowButton:NSWindowMiniaturizeButton] setHidden:YES]; + [[window->nsWindow standardWindowButton:NSWindowZoomButton] setHidden:YES]; + } + + MPNativeView* view = [[MPNativeView alloc] initWithMPWindow:window]; + + [window->nsWindow setContentView:view]; + [window->nsWindow makeFirstResponder:view]; + [window->nsWindow setAcceptsMouseMovedEvents:YES]; + + mp_window_update_rect_cache(window); + + mp_window windowHandle = mp_window_handle_from_ptr(window); + + mp_rect mainViewFrame = {0, 0, contentRect.w, contentRect.h}; + window->mainView = mp_view_create(windowHandle, mainViewFrame); + + return(windowHandle); +}//autoreleasepool +} + +void mp_window_destroy(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + [windowData->nsWindow orderOut:nil]; + + [windowData->nsWindow setDelegate:nil]; + [windowData->nsWindowDelegate release]; + windowData->nsWindowDelegate = nil; + + [windowData->nsView release]; + windowData->nsView = nil; + + [windowData->nsWindow close]; //also release the window + + mp_window_recycle_ptr(windowData); + } +} // autoreleasepool +} + +bool mp_window_should_close(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + return(windowData->shouldClose); + } + else + { + return(false); + } +} + +void mp_window_cancel_close(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + windowData->shouldClose = false; + } +} + +void mp_window_request_close(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + [windowData->nsWindow close]; + //NOTE(martin): this will call our window delegate willClose method + } +} + +void* mp_window_native_pointer(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + return((__bridge void*)windowData->nsWindow); + } + else + { + return(0); + } +} + +void mp_window_center(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + [windowData->nsWindow center]; + } +}} + +bool mp_window_is_hidden(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + return(windowData->hidden); + } + else + { + return(false); + } +} + +bool mp_window_is_focused(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + return([windowData->nsWindow isKeyWindow]); + } + else + { + return(false); + } +}} + +void mp_window_hide(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + windowData->hidden = true; + [windowData->nsWindow orderOut:nil]; + } +}} + +void mp_window_focus(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + [windowData->nsWindow makeKeyWindow]; + } +}} + +void mp_window_send_to_back(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + [windowData->nsWindow orderBack:nil]; + } +}} + +void mp_window_bring_to_front(mp_window window) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + windowData->hidden = false; + [windowData->nsWindow orderFront:nil]; + } +}} + +void mp_window_bring_to_front_and_focus(mp_window window) +{ + mp_window_bring_to_front(window); + mp_window_focus(window); +} + + +mp_rect mp_window_content_rect_for_frame_rect(mp_rect frameRect, mp_window_style style) +{@autoreleasepool{ + u32 mask = mp_osx_get_window_style_mask(style); + mp_rect nativeFrame = mp_user_to_osx_screen_rect(frameRect); + NSRect frame = NSMakeRect(nativeFrame.x, nativeFrame.y, nativeFrame.w, nativeFrame.h); + NSRect content = [NSWindow contentRectForFrameRect:frame styleMask:mask]; + mp_rect result = {content.origin.x, content.origin.y, content.size.width, content.size.height}; + result = mp_osx_to_user_screen_rect(result); + return(result); +}} + +mp_rect mp_window_frame_rect_for_content_rect(mp_rect contentRect, mp_window_style style) +{@autoreleasepool{ + uint32 mask = mp_osx_get_window_style_mask(style); + mp_rect nativeContent = mp_user_to_osx_screen_rect(contentRect); + NSRect content = NSMakeRect(nativeContent.x, nativeContent.y, nativeContent.w, nativeContent.h); + NSRect frame = [NSWindow frameRectForContentRect:content styleMask:mask]; + mp_rect result = {frame.origin.x, frame.origin.y, frame.size.width, frame.size.height}; + result = mp_osx_to_user_screen_rect(result); + return(result); +}} + +mp_rect mp_window_get_content_rect(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + return(windowData->contentRect); + } + else + { + return((mp_rect){}); + } +} + +mp_rect mp_window_get_absolute_content_rect(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + mp_rect rect = windowData->contentRect; + rect.x += windowData->frameRect.x; + rect.y += windowData->frameRect.y; + return(rect); + } + else + { + return((mp_rect){}); + } +} + +mp_rect mp_window_get_frame_rect(mp_window window) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + return(windowData->frameRect); + } + else + { + return((mp_rect){}); + } +} + +void mp_window_set_content_rect(mp_window window, mp_rect contentRect) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + u32 mask = mp_osx_get_window_style_mask(windowData->style); + + mp_rect nativeRect = mp_user_to_osx_screen_rect(contentRect); + NSRect content = NSMakeRect(nativeRect.x, nativeRect.y, nativeRect.w, nativeRect.h); + NSRect frame = [NSWindow frameRectForContentRect:content styleMask:mask]; + + [windowData->nsWindow setFrame:frame display:YES]; + + mp_window_update_rect_cache(windowData); + } +}} +void mp_window_set_frame_rect(mp_window window, mp_rect frameRect) +{@autoreleasepool{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + mp_rect nativeRect = mp_user_to_osx_screen_rect(frameRect); + NSRect frame = NSMakeRect(nativeRect.x, nativeRect.y, nativeRect.w, nativeRect.h); + [windowData->nsWindow setFrame:frame display:YES]; + + mp_window_update_rect_cache(windowData); + NSRect contentRect = [[windowData->nsWindow contentView] frame]; + } +}} + +void mp_window_set_frame_size(mp_window window, int width, int height) +{ + mp_rect frame = mp_window_get_frame_rect(window); + frame.w = width; + frame.h = height; + mp_window_set_frame_rect(window, frame); +} + +void mp_window_set_content_size(mp_window window, int width, int height) +{ + mp_window_data* windowData = mp_window_ptr_from_handle(window); + if(windowData) + { + mp_rect frame = windowData->frameRect; + mp_rect content = mp_window_content_rect_for_frame_rect(frame, windowData->style); + content.w = width; + content.h = height; + frame = mp_window_frame_rect_for_content_rect(content, windowData->style); + mp_window_set_frame_rect(window, frame); + } +} + +//-------------------------------------------------------------------- +// view management +//-------------------------------------------------------------------- + +mp_view mp_view_nil() +{ + return((mp_view){0}); +} + +bool mp_view_is_nil(mp_view view) +{ + return(!view.h); +} + +mp_view mp_view_create(mp_window windowHandle, mp_rect frame) +{@autoreleasepool{ + mp_window_data* window = mp_window_ptr_from_handle(windowHandle); + if(!window) + { + LOG_ERROR("Can't create view for nil window\n"); + return(mp_view_nil()); + } + + mp_view_data* view = mp_view_alloc(); + if(!view) + { + LOG_ERROR("Could not allocate view data\n"); + return(mp_view_nil()); + } + + view->window = windowHandle; + + NSRect nsFrame = {{frame.x, frame.y}, {frame.w, frame.h}}; + view->nsView = [[NSView alloc] initWithFrame: nsFrame]; + + [[window->nsWindow contentView] addSubview: view->nsView]; + + return(mp_view_handle_from_ptr(view)); +}} + +void mp_view_destroy(mp_view viewHandle) +{@autoreleasepool{ + mp_view_data* view = mp_view_ptr_from_handle(viewHandle); + if(!view) + { + return; + } + + mp_window_data* window = mp_window_ptr_from_handle(view->window); + if(!window) + { + return; + } + + [view->nsView removeFromSuperview]; + + mp_view_recycle_ptr(view); +}} + +void mp_view_set_frame(mp_view viewHandle, mp_rect frame) +{ + mp_view_data* view = mp_view_ptr_from_handle(viewHandle); + if(!view) + { + return; + } + + NSRect nsFrame = {{frame.x, frame.y}, {frame.w, frame.h}}; + view->nsView = [[NSView alloc] initWithFrame: nsFrame]; + + mg_surface_resize(view->surface, frame.w, frame.h); +} + +//-------------------------------------------------------------------- +// Main loop throttle +//-------------------------------------------------------------------- +/* +//////////////////////////////////////////////////////////// +//TODO +//////////////////////////////////////////////////////////// + +void mp_set_target_fps(u32 fps) +{ + __mpAppData.frameStats.targetFramePeriod = 1./(f64)fps; + __mpAppData.frameStats.workTime = 0; + __mpAppData.frameStats.remainingTime = 0; + + if(__mpAppData.frameTimer) + { + [__mpAppData.frameTimer invalidate]; + } + + __mpAppData.frameTimer = [NSTimer timerWithTimeInterval: __mpAppData.frameStats.targetFramePeriod + target: [NSApp delegate] + selector:@selector(timerElapsed:) + userInfo:nil + repeats:YES]; + + [[NSRunLoop currentRunLoop] addTimer:__mpAppData.frameTimer forMode:NSRunLoopCommonModes]; +} + +void mp_begin_frame() +{ + __mpAppData.frameStats.start = mp_get_elapsed_seconds(); + + LOG_DEBUG("workTime = %.6f (%.6f fps), remaining = %.6f\n", + __mpAppData.frameStats.workTime, + 1/__mpAppData.frameStats.workTime, + __mpAppData.frameStats.remainingTime); + +} + +void mp_end_frame() +{ + __mpAppData.frameStats.workTime = mp_get_elapsed_seconds() - __mpAppData.frameStats.start; + __mpAppData.frameStats.remainingTime = __mpAppData.frameStats.targetFramePeriod - __mpAppData.frameStats.workTime; + + while(__mpAppData.frameStats.remainingTime > 100e-9) + { + if(__mpAppData.frameStats.remainingTime > 10e-6) + { + mp_sleep_nanoseconds(__mpAppData.frameStats.remainingTime*0.8*1e9); + } + __mpAppData.frameStats.workTime = mp_get_elapsed_seconds() - __mpAppData.frameStats.start; + __mpAppData.frameStats.remainingTime = __mpAppData.frameStats.targetFramePeriod - __mpAppData.frameStats.workTime; + } +} +*/ + +//-------------------------------------------------------------------- +// Events handling +//-------------------------------------------------------------------- + +void mp_pump_events(f64 timeout) +{ + __mpAppData.inputState.frameCounter++; + + @autoreleasepool + { + bool accumulate = false; + NSDate* date = 0; + if(timeout > 0) + { + date = [NSDate dateWithTimeIntervalSinceNow: (double) timeout]; + } + else if(timeout == 0) + { + date = [NSDate distantPast]; + accumulate = true; + } + else + { + date = [NSDate distantFuture]; + } + + for(;;) + { + NSEvent* event = [NSApp nextEventMatchingMask: NSEventMaskAny + untilDate:date + inMode: NSDefaultRunLoopMode + dequeue: YES]; + + if(event != nil) + { + [NSApp sendEvent:event]; + + if(!accumulate) + { + break; + } + } + else + { + break; + } + } + } +} + +bool mp_next_event(mp_event* event) +{ + //NOTE pop and return event from queue + if(ringbuffer_read_available(&__mpAppData.eventQueue) >= sizeof(mp_event)) + { + u64 read = ringbuffer_read(&__mpAppData.eventQueue, sizeof(mp_event), (u8*)event); + DEBUG_ASSERT(read == sizeof(mp_event)); + return(true); + } + else + { + return(false); + } +} + +//-------------------------------------------------------------------- +// Input state polling +//-------------------------------------------------------------------- + +mp_key_state mp_input_get_key_state(mp_key_code key) +{ + if(key <= MP_KEY_COUNT) + { + return(__mpAppData.inputState.keyboard.keys[key]); + } + else + { + return((mp_key_state){0}); + } +} +mp_key_state mp_input_get_mouse_button_state(mp_mouse_button button) +{ + if(button <= MP_MOUSE_BUTTON_COUNT) + { + return(__mpAppData.inputState.mouse.buttons[button]); + } + else + { + return((mp_key_state){0}); + } +} + +bool mp_input_check_key_transition(mp_key_state* key, bool pressed) +{ + bool res = ( (key->lastUpdate == __mpAppData.inputState.frameCounter) + && key->transitionCounter + &&(key->down == pressed || key->transitionCounter > 1)); + return(res); +} + +bool mp_input_key_down(mp_key_code key) +{ + mp_key_state state = mp_input_get_key_state(key); + return(state.down); +} +bool mp_input_key_pressed(mp_key_code key) +{ + mp_key_state state = mp_input_get_key_state(key); + bool res = mp_input_check_key_transition(&state, true); + return(res); +} + +bool mp_input_key_released(mp_key_code key) +{ + mp_key_state state = mp_input_get_key_state(key); + bool res = mp_input_check_key_transition(&state, false); + return(res); +} + +bool mp_input_mouse_down(mp_mouse_button button) +{ + mp_key_state state = mp_input_get_mouse_button_state(button); + return(state.down); +} + +bool mp_input_mouse_pressed(mp_mouse_button button) +{ + mp_key_state state = mp_input_get_mouse_button_state(button); + bool res = mp_input_check_key_transition(&state, true); + return(res); +} + +bool mp_input_mouse_released(mp_mouse_button button) +{ + mp_key_state state = mp_input_get_mouse_button_state(button); + bool res = mp_input_check_key_transition(&state, false); + return(res); +} + +bool mp_input_mouse_clicked(mp_mouse_button button) +{ + mp_key_state state = mp_input_get_mouse_button_state(button); + return(state.clicked); +} + +bool mp_input_mouse_double_clicked(mp_mouse_button button) +{ + mp_key_state state = mp_input_get_mouse_button_state(button); + return(state.doubleClicked); +} + +mp_key_mods mp_input_key_mods() +{ + return(__mpAppData.inputState.keyboard.mods); +} + +vec2 mp_input_mouse_position() +{ + return(__mpAppData.inputState.mouse.pos); +} + +vec2 mp_input_mouse_delta() +{ + if(__mpAppData.inputState.mouse.lastUpdate == __mpAppData.inputState.frameCounter) + { + return(__mpAppData.inputState.mouse.delta); + } + else + { + return((vec2){0, 0}); + } +} + +vec2 mp_input_mouse_wheel() +{ + if(__mpAppData.inputState.mouse.lastUpdate == __mpAppData.inputState.frameCounter) + { + return(__mpAppData.inputState.mouse.wheel); + } + else + { + return((vec2){0, 0}); + } +} + +str32 mp_input_text_utf32(mem_arena* arena) +{ + str32 res = {0}; + if(__mpAppData.inputState.text.lastUpdate == __mpAppData.inputState.frameCounter) + { + res = str32_push_copy(arena, __mpAppData.inputState.text.codePoints); + } + return(res); +} + +str8 mp_input_text_utf8(mem_arena* arena) +{ + str8 res = {0}; + if(__mpAppData.inputState.text.lastUpdate == __mpAppData.inputState.frameCounter) + { + res = utf8_push_from_codepoints(arena, __mpAppData.inputState.text.codePoints); + } + return(res); +} + +//-------------------------------------------------------------------- +// app resources +//-------------------------------------------------------------------- + +#import +#include + +int mp_app_get_resource_path(const char* name, char** result) +{ + @autoreleasepool + { + NSBundle* mainBundle = [NSBundle mainBundle]; + if(!mainBundle) + { + //NOTE(martin): we assume we are running from the command line in debug mode + char* currentPath = getcwd(0, 0); + char* buffer = (char*)malloc(strlen(currentPath) + strlen(name) + 2); + strcpy(buffer, currentPath); + strcat(buffer, "/"); + strcat(buffer, name); + *result = realpath(buffer, 0); + free(currentPath); + free(buffer); + return(1); + } + else + { + NSString* nsName = [[NSString alloc] initWithUTF8String:name]; + NSString* nsPath = [mainBundle executablePath]; + + const char* utf8Path = [nsPath UTF8String]; + const char* dir = dirname((char*)utf8Path); + char* buffer = (char*)malloc(strlen(dir) + strlen(name) + 2); + strcpy(buffer, dir); + strcat(buffer, "/"); + strcat(buffer, name); + *result = realpath(buffer, 0); + free(buffer); + + return(-1); + } + } +} + +#include +str8 mp_app_get_executable_path(mem_arena* arena) +{@autoreleasepool{ + str8 result = {}; + u32 size = 0; + _NSGetExecutablePath(0, &size); + result.len = size; + result.ptr = mem_arena_alloc_array(arena, char, result.len); + _NSGetExecutablePath(result.ptr, &size); + return(result); +}} + +//-------------------------------------------------------------------- +// system dialogs windows +//-------------------------------------------------------------------- + +str8 mp_open_dialog(mem_arena* arena, + const char* title, + const char* defaultPath, + int filterCount, + const char** filters, + bool directory) +{ + @autoreleasepool + { + NSWindow *keyWindow = [NSApp keyWindow]; + + NSOpenPanel* dialog = [NSOpenPanel openPanel] ; + [dialog setLevel:CGShieldingWindowLevel()]; + + if(filterCount) + { + NSMutableArray * fileTypesArray = [NSMutableArray array]; + for(int i=0; i < filterCount; i++) + { + NSString * filt = [NSString stringWithUTF8String:filters[i]]; + [fileTypesArray addObject:filt]; + } + [dialog setAllowedFileTypes:fileTypesArray]; + } + // Enable options in the dialog. + if(directory) + { + [dialog setCanChooseDirectories:YES]; + } + else + { + [dialog setCanChooseFiles:YES]; + } + + + [dialog setAllowsMultipleSelection:FALSE]; + NSString* nsPath = [[NSString stringWithUTF8String:defaultPath?defaultPath:"~"] stringByExpandingTildeInPath]; + [dialog setDirectoryURL:[NSURL fileURLWithPath:nsPath]]; + + // Display the dialog box. If the OK pressed, + // process the files. + + if( [dialog runModal] == NSModalResponseOK ) + { + // Gets list of all files selected + NSArray *files = [dialog URLs]; + //TODO: Loop through the files and process them. + + const char* result = [[[files objectAtIndex:0] path] UTF8String]; + + str8 path = str8_push_cstring(arena, result); + [keyWindow makeKeyWindow]; + + return(path); + } + else + { + [keyWindow makeKeyWindow]; + return((str8){0, 0}); + } + } +} + + +str8 mp_save_dialog(mem_arena* arena, + const char* title, + const char* defaultPath, + int filterCount, + const char** filters) +{ + @autoreleasepool + { + NSWindow *keyWindow = [NSApp keyWindow]; + + NSSavePanel* dialog = [NSSavePanel savePanel] ; + [dialog setLevel:CGShieldingWindowLevel()]; + + if(filterCount) + { + NSMutableArray * fileTypesArray = [NSMutableArray array]; + for(int i=0; i < filterCount; i++) + { + NSString * filt = [NSString stringWithUTF8String:filters[i]]; + [fileTypesArray addObject:filt]; + } + + // Enable options in the dialog. + [dialog setAllowedFileTypes:fileTypesArray]; + } + NSString* nsPath = [[NSString stringWithUTF8String:defaultPath?defaultPath:"~"] stringByExpandingTildeInPath]; + [dialog setDirectoryURL:[NSURL fileURLWithPath:nsPath]]; + + // Display the dialog box. If the OK pressed, + // process the files. + + if( [dialog runModal] == NSModalResponseOK ) + { + // Gets list of all files selected + NSURL *files = [dialog URL]; + // Loop through the files and process them. + + const char* result = [[files path] UTF8String]; + + str8 path = str8_push_cstring(arena, result); + [keyWindow makeKeyWindow]; + return(path); + } + else + { + [keyWindow makeKeyWindow]; + return((str8){0, 0}); + } + } +} + +int mp_alert_popup(const char* title, + const char* message, + uint32 count, + const char** options) +{ + @autoreleasepool + { + NSWindow *keyWindow = [NSApp keyWindow]; + + NSAlert* alert = [[NSAlert alloc] init]; + NSString* string; + for(int i=count-1;i>=0;i--) + { + string = [[NSString alloc] initWithUTF8String:options[i]]; + [alert addButtonWithTitle:string]; + [string release]; + } + string = [[NSString alloc] initWithUTF8String:message]; + [alert setMessageText:string]; + [string release]; + + [alert setAlertStyle:NSAlertStyleWarning]; + int result = count - ([alert runModal]-NSAlertFirstButtonReturn) - 1; + printf("result = %i, NSAlertFirstButtonReturn = %li\n", result, (long)NSAlertFirstButtonReturn); + [keyWindow makeKeyWindow]; + return(result); + } +} + + +#undef LOG_SUBSYSTEM diff --git a/src/platform/linux_clock.c b/src/platform/linux_clock.c new file mode 100644 index 0000000..d6c1367 --- /dev/null +++ b/src/platform/linux_clock.c @@ -0,0 +1,193 @@ +/************************************************************//** +* +* @file: linux_clock.c +* @author: Martin Fouilleul +* @date: 20/04/2020 +* @revision: +* +*****************************************************************/ + +#include //fabs() +#include +#include // gettimeofday() + +#include +#include + +#include // nanosleep() + +#include"platform_rng.h" +#include"platform_clock.h" + +extern "C" { + +//TODO(martin): measure the actual values of these constants +const f64 SYSTEM_FUZZ = 25e-9, // minimum time to read the clock (s) + SYSTEM_TICK = 25e-9; // minimum step between two clock readings (s) + +static u64 __initialTimestamp__ = 0; +static u64 __initialMonotonicNanoseconds__ = 0; + +static inline u64 LinuxGetMonotonicNanoseconds() +{ + timespec ts; + clock_gettime(CLOCK_MONOTONIC_RAW, &ts); + //WARN(martin): do not multiply ts.tv_sec directly (implicit conversion will overflow) + u64 r = ts.tv_sec; + r *= 1000000000; + r += ts.tv_nsec; + return(r); +} + +static inline u64 LinuxGetUptimeNanoseconds() +{ + timespec ts; + clock_gettime(CLOCK_MONOTONIC, &ts); + //WARN(martin): do not multiply ts.tv_sec directly (implicit conversion will overflow) + u64 r = ts.tv_sec; + r *= 1000000000; + r += ts.tv_nsec; + return(r); +} + + +void ClockSystemInit() +{ + //NOTE(martin): we don't know of a widely supported way of getting the boot time on linux, so + // we fallback to our second best choice, which is taking an initial timestamp and + // the initial monotonic time now + + timeval tv; + gettimeofday(&tv, 0); + __initialMonotonicNanoseconds__ = LinuxGetMonotonicNanoseconds(); + + //NOTE(martin): convert boot date to timestamp + __initialTimestamp__ = (((u64)tv.tv_sec + JAN_1970) << 32) + + (u64)(tv.tv_usec * 1e-6 * TIMESTAMPS_PER_SECOND); + + //TODO(martin): maybe get a state vector for exclusive clock usage ? + RandomSeedFromDevice(); +} + +fx_timestamp ClockGetTimestamp(clock_kind clock) +{ + fx_timestamp ts = {0}; + switch(clock) + { + case SYS_CLOCK_MONOTONIC: + { + //NOTE(martin): compute monotonic offset and add it to bootup timestamp + u64 noff = LinuxGetMonotonicNanoseconds() - __initialMonotonicNanoseconds__; + u64 foff = (u64)(noff * 1e-9 * TIMESTAMPS_PER_SECOND); + ts.ts = __initialTimestamp__ + foff; + } break; + + case SYS_CLOCK_UPTIME: + { + //TODO(martin): maybe we should warn that this date is inconsistent after a sleep ? + //NOTE(martin): compute uptime offset and add it to bootup timestamp + u64 noff = LinuxGetUptimeNanoseconds() - __initialMonotonicNanoseconds__ ; + u64 foff = (u64)(noff * 1e-9 * TIMESTAMPS_PER_SECOND); + ts.ts = __initialTimestamp__ + foff; + } break; + + case SYS_CLOCK_DATE: + { + //NOTE(martin): get system date and convert it to a fixed-point timestamp + timeval tv; + gettimeofday(&tv, 0); + ts.ts = (((u64)tv.tv_sec + JAN_1970) << 32) + + (u64)(tv.tv_usec * 1e-6 * TIMESTAMPS_PER_SECOND); + } break; + } + + //NOTE(martin): add a random fuzz between 0 and 1 times the system fuzz + f64 fuzz = RandomU32()/(f64)(~(0UL)) * SYSTEM_FUZZ; + fx_timediff tdfuzz = TimediffFromSeconds(fuzz); + ts = TimestampAdd(ts, tdfuzz); + //TODO(martin): ensure that we always return a value greater than the last value + + return(ts); +} + + +f64 ClockGetTime(clock_kind clock) +{ + switch(clock) + { + case SYS_CLOCK_MONOTONIC: + { + //NOTE(martin): compute monotonic offset and add it to bootup timestamp + u64 noff = LinuxGetMonotonicNanoseconds(); + return((f64)noff * 1e-9); + } break; + + case SYS_CLOCK_UPTIME: + { + //TODO(martin): maybe we should warn that this date is inconsistent after a sleep ? + //NOTE(martin): compute uptime offset and add it to bootup timestamp + u64 noff = LinuxGetUptimeNanoseconds(); + return((f64)noff * 1e-9); + } break; + + case SYS_CLOCK_DATE: + { + //TODO(martin): maybe warn about precision loss ? + // could also change the epoch since we only promise to return a relative time + //NOTE(martin): get system date and convert it to seconds + timeval tv; + gettimeofday(&tv, 0); + return(((f64)tv.tv_sec + JAN_1970) + ((f64)tv.tv_usec * 1e-6)); + } break; + } + return(0); +} + +void ClockSleepNanoseconds(u64 nanoseconds) +{ + timespec rqtp; + rqtp.tv_sec = nanoseconds / 1000000000; + rqtp.tv_nsec = nanoseconds - rqtp.tv_sec * 1000000000; + nanosleep(&rqtp, 0); +} + + + +//////////////////////////////////////////////////////////////////// +//TODO: update these functions for various clocks besides monotonic +//////////////////////////////////////////////////////////////////// +f64 ClockGetGranularity(clock_kind clock) +{ + u64 minDiff = ~(0ULL); + + const int GRANULARITY_NUM_ITERATION = 100000; + for(int i=0; i //fabs() +#include +#include // gettimeofday() +#include +#include +#include +#include // availability macros + +#include +#include + +#include // nanosleep() + +#include"platform_clock.h" + + +typedef struct timeval timeval; +typedef struct timespec timespec; + +#define LOG_SUBSYSTEM "Platform" + +//TODO(martin): measure the actual values of these constants +const f64 SYSTEM_FUZZ = 25e-9, // minimum time to read the clock (s) + SYSTEM_TICK = 25e-9; // minimum step between two clock readings (s) + +static mach_timebase_info_data_t __machTimeBase__ = {1,1}; +static u64 __initialTimestamp__ = 0; + +static inline u64 OSXGetUptimeNanoseconds() +{ + //NOTE(martin): according to the documentation, mach_absolute_time() does not + // increment when the system is asleep + u64 now = mach_absolute_time(); + now *= __machTimeBase__.numer; + now /= __machTimeBase__.denom; + return(now); +} + +static inline u64 OSXGetMonotonicNanoseconds() +{ + //NOTE(martin): according to the documentation, MP_CLOCK_MONOTONIC increment monotonically + // on systems where MP_CLOCK_MONOTONIC_RAW is present, we may want to use that instead, + // because MP_CLOCK_MONOTONIC seems to be subject to frequency changes ? + + #if __MAC_OS_X_VERSION_MIN_REQUIRED >= 101200 + #ifndef CLOCK_MONOTONIC_RAW + #error "CLOCK_MONOTONIC_RAW not found. Please verify that is included from the MacOSX SDK rather than /usr/local/include" + #else + return(clock_gettime_nsec_np(CLOCK_MONOTONIC_RAW)); + #endif + #else + //TODO(martin): quick and dirty hack is to fallback to uptime, + // but we should either only support macos version >= 10.12, or find a proper solution + return(OSXGetUptimeNanoseconds()); + #endif +} + +static const f64 CLK_TIMESTAMPS_PER_SECOND = 4294967296.; // 2^32 as a double +static const u64 CLK_JAN_1970 = 2208988800ULL; // seconds from january 1900 to january 1970 + +void mp_clock_init() +{ + mach_timebase_info(&__machTimeBase__); + + //NOTE(martin): get the date of system boot time + timeval tv = {0, 0}; + + int mib[2] = {CTL_KERN, + KERN_BOOTTIME}; + size_t size = sizeof(tv); + + if(sysctl(mib, 2, &tv, &size, 0, 0) == -1) + { + LOG_ERROR("can't read boot time\n"); + } + //NOTE(martin): convert boot date to timestamp + __initialTimestamp__ = (((u64)tv.tv_sec + CLK_JAN_1970) << 32) + + (u64)(tv.tv_usec * 1e-6 * CLK_TIMESTAMPS_PER_SECOND); + + //TODO(martin): maybe get a state vector for exclusive clock usage ? + //RandomSeedFromDevice(); +} + +u64 mp_get_timestamp(mp_clock_kind clock) +{ + u64 ts = 0; + switch(clock) + { + case MP_CLOCK_MONOTONIC: + { + //NOTE(martin): compute monotonic offset and add it to bootup timestamp + u64 noff = OSXGetMonotonicNanoseconds() ; + u64 foff = (u64)(noff * 1e-9 * CLK_TIMESTAMPS_PER_SECOND); + ts = __initialTimestamp__ + foff; + } break; + + case MP_CLOCK_UPTIME: + { + //TODO(martin): maybe we should warn that this date is inconsistent after a sleep ? + //NOTE(martin): compute uptime offset and add it to bootup timestamp + u64 noff = OSXGetUptimeNanoseconds() ; + u64 foff = (u64)(noff * 1e-9 * CLK_TIMESTAMPS_PER_SECOND); + ts = __initialTimestamp__ + foff; + } break; + + case MP_CLOCK_DATE: + { + //NOTE(martin): get system date and convert it to a fixed-point timestamp + timeval tv; + gettimeofday(&tv, 0); + ts = (((u64)tv.tv_sec + CLK_JAN_1970) << 32) + + (u64)(tv.tv_usec * 1e-6 * CLK_TIMESTAMPS_PER_SECOND); + } break; + } + + /* + //NOTE(martin): add a random fuzz between 0 and 1 times the system fuzz + //TODO(martin): ensure that we always return a value greater than the last value + f64 fuzz = RandomU32()/(f64)(~(0UL)) * SYSTEM_FUZZ; + ts_timediff tdfuzz = TimediffFromSeconds(fuzz); + ts = TimestampAdd(ts, tdfuzz); + */ + + return(ts); +} + + +f64 mp_get_time(mp_clock_kind clock) +{ + switch(clock) + { + case MP_CLOCK_MONOTONIC: + { + //NOTE(martin): compute monotonic offset and add it to bootup timestamp + u64 noff = OSXGetMonotonicNanoseconds(); + return((f64)noff * 1e-9); + } break; + + case MP_CLOCK_UPTIME: + { + //TODO(martin): maybe we should warn that this date is inconsistent after a sleep ? + //NOTE(martin): compute uptime offset and add it to bootup timestamp + u64 noff = OSXGetUptimeNanoseconds(); + return((f64)noff * 1e-9); + } break; + + case MP_CLOCK_DATE: + { + //TODO(martin): maybe warn about precision loss ? + // could also change the epoch since we only promise to return a relative time + //NOTE(martin): get system date and convert it to seconds + timeval tv; + gettimeofday(&tv, 0); + return(((f64)tv.tv_sec + CLK_JAN_1970) + ((f64)tv.tv_usec * 1e-6)); + } break; + } +} + +void mp_sleep_nanoseconds(u64 nanoseconds) +{ + timespec rqtp; + rqtp.tv_sec = nanoseconds / 1000000000; + rqtp.tv_nsec = nanoseconds - rqtp.tv_sec * 1000000000; + nanosleep(&rqtp, 0); +} + +#undef LOG_SUBSYSTEM diff --git a/src/platform/platform_base_allocator.h b/src/platform/platform_base_allocator.h new file mode 100644 index 0000000..975879d --- /dev/null +++ b/src/platform/platform_base_allocator.h @@ -0,0 +1,28 @@ +/************************************************************//** +* +* @file: platform_base_allocator.h +* @author: Martin Fouilleul +* @date: 10/09/2021 +* @revision: +* +*****************************************************************/ +#ifndef __PLATFORM_BASE_ALLOCATOR_H_ +#define __PLATFORM_BASE_ALLOCATOR_H_ + +#include"typedefs.h" +#include"memory.h" + +#ifdef __cplusplus +extern "C" { +#endif + +void* mem_base_reserve_mmap(void* context, u64 size); +void mem_base_release_mmap(void* context, void* ptr, u64 size); + +mem_base_allocator* mem_base_allocator_default(); + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__PLATFORM_BASE_ALLOCATOR_H_ diff --git a/src/platform/platform_clock.h b/src/platform/platform_clock.h new file mode 100644 index 0000000..29d982d --- /dev/null +++ b/src/platform/platform_clock.h @@ -0,0 +1,34 @@ +/************************************************************//** +* +* @file: platform_clock.h +* @author: Martin Fouilleul +* @date: 07/03/2019 +* @revision: +* +*****************************************************************/ +#ifndef __PLATFORM_CLOCK_H_ +#define __PLATFORM_CLOCK_H_ + +#include"typedefs.h" + +#ifdef __cplusplus +extern "C" { +#endif // __cplusplus + +typedef enum { + MP_CLOCK_MONOTONIC, // clock that increment monotonically + MP_CLOCK_UPTIME, // clock that increment monotonically during uptime + MP_CLOCK_DATE // clock that is driven by the platform time +} mp_clock_kind; + +void mp_clock_init(); // initialize the clock subsystem +u64 mp_get_timestamp(mp_clock_kind clock); +f64 mp_get_time(mp_clock_kind clock); +void mp_sleep_nanoseconds(u64 nanoseconds); // sleep for a given number of nanoseconds + +#ifdef __cplusplus +} // extern "C" +#endif // __cplusplus + + +#endif //__PLATFORM_CLOCK_H_ diff --git a/src/platform/platform_rng.h b/src/platform/platform_rng.h new file mode 100644 index 0000000..b4726b6 --- /dev/null +++ b/src/platform/platform_rng.h @@ -0,0 +1,18 @@ +/************************************************************//** +* +* @file: platform_rng.h +* @author: Martin Fouilleul +* @date: 06/03/2020 +* @revision: +* +*****************************************************************/ +#ifndef __PLATFORM_RANDOM_H_ +#define __PLATFORM_RANDOM_H_ + +#include"typedefs.h" + +int RandomSeedFromDevice(); +u32 RandomU32(); +u64 RandomU64(); + +#endif //__PLATFORM_RANDOM_H_ diff --git a/src/platform/platform_socket.h b/src/platform/platform_socket.h new file mode 100644 index 0000000..8585cd0 --- /dev/null +++ b/src/platform/platform_socket.h @@ -0,0 +1,161 @@ +/************************************************************//** +* +* @file: platform_socket.h +* @author: Martin Fouilleul +* @date: 22/03/2019 +* @revision: +* +*****************************************************************/ +#ifndef __PLATFORM_SOCKET_H_ +#define __PLATFORM_SOCKET_H_ + +#include // timeval +#include"typedefs.h" + +#ifdef __cplusplus +extern "C" { +#else +typedef struct timeval timeval; +#endif //__cplusplus + + +//---------------------------------------------------------------------------------- +// Errors +//---------------------------------------------------------------------------------- + +//TODO(martin): extend these error codes +const int SOCK_ERR_OK = 0, + SOCK_ERR_UNKNOWN = -1, + SOCK_ERR_ACCESS = -2, + SOCK_ERR_MEM = -3, + SOCK_ERR_INTR = -4, + SOCK_ERR_USED = -5, + SOCK_ERR_BADF = -6, + SOCK_ERR_ABORT = -7, + SOCK_ERR_NBLOCK = -8; + +int SocketGetLastError(); +const char* SocketGetLastErrorMessage(); + +//---------------------------------------------------------------------------------- +// Addresses +//---------------------------------------------------------------------------------- + +//NOTE(martin): net_ip and net_port are stored in network byte order +// host_ip and host_port are stored in host byte order +typedef uint32 net_ip; +typedef uint16 net_port; +typedef uint32 host_ip; +typedef uint16 host_port; + +typedef struct +{ + net_ip ip; + net_port port; +} socket_address; + +//NOTE(martin): these are in host byte order ! +const host_ip SOCK_IP_LOOPBACK = 0x7f000001; +const host_ip SOCK_IP_ANY = 0; +const host_port SOCK_PORT_ANY = 0; + +net_ip StringToNetIP(const char* addr); +const char* NetIPToString(net_ip ip); + +host_ip StringToHostIP(const char* addr); +const char* HostIPToString(host_ip ip); + +net_ip HostToNetIP(host_ip ip); +net_port HostToNetPort(host_port port); +host_ip NetToHostIP(net_ip ip); +host_port NetToHostPort(net_port port); + + +int SocketGetIFAddresses(int* count, net_ip* ips); +net_ip SocketGetDefaultExternalIP(); + +//---------------------------------------------------------------------------------- +// Socket API +//---------------------------------------------------------------------------------- + +typedef struct platform_socket platform_socket; + +typedef enum { SOCK_UDP, SOCK_TCP } socket_transport; + +const int SOCK_MSG_OOB = 0x01, + SOCK_MSG_PEEK = 0x02, + SOCK_MSG_DONTROUTE = 0x04, + SOCK_MSG_WAITALL = 0x40; + +platform_socket* SocketOpen(socket_transport transport); +int SocketClose(platform_socket* socket); + +int SocketBind(platform_socket* socket, socket_address* addr); +int SocketListen(platform_socket* socket, int backlog); +platform_socket* SocketAccept(platform_socket* socket, socket_address* from); + +int SocketConnect(platform_socket* socket, socket_address* addr); + +int64 SocketReceive(platform_socket* socket, void* buffer, uint64 size, int flags); +int64 SocketReceiveFrom(platform_socket* socket, void* buffer, uint64 size, int flags, socket_address* from); + +int64 SocketSend(platform_socket* socket, void* buffer, uint64 size, int flags); +int64 SocketSendTo(platform_socket* socket, void* buffer, uint64 size, int flags, socket_address* to); + +int SocketGetAddress(platform_socket* socket, socket_address* addr); + +//---------------------------------------------------------------------------------- +// Multiplexing +//---------------------------------------------------------------------------------- +const uint8 SOCK_ACTIVITY_IN = 1<<0, + SOCK_ACTIVITY_OUT = 1<<2, + SOCK_ACTIVITY_ERR = 1<<3; + +typedef struct +{ + platform_socket* sock; + uint8 watch; + uint8 set; +} socket_activity; + +int SocketSelect(uint32 count, socket_activity* set, double timeout); + +//---------------------------------------------------------------------------------- +// Socket Options +//---------------------------------------------------------------------------------- + +int SocketSetReceiveTimeout(platform_socket* socket, timeval* tv); +int SocketSetSendTimeout(platform_socket* socket, timeval* tv); +int SocketSetBroadcast(platform_socket* sock, bool enable); +int SocketSetReuseAddress(platform_socket* sock, bool enable); +int SocketSetReusePort(platform_socket* sock, bool enable); +int SocketSetReceiveTimestamping(platform_socket* socket, bool enable); + +//---------------------------------------------------------------------------------- +// Multicast +//---------------------------------------------------------------------------------- + +int SocketSetMulticastLoop(platform_socket* sock, bool enable); +int SocketJoinMulticastGroup(platform_socket* socket, host_ip group, host_ip interface); +int SocketLeaveMulticastGroup(platform_socket* socket, host_ip group, host_ip interface); + +//---------------------------------------------------------------------------------- +//Ancillary data API +//---------------------------------------------------------------------------------- +typedef struct +{ + u64 messageBufferSize; + char* messageBuffer; + + u64 controlBufferSize; + char* controlBuffer; +} socket_msg; + +int SocketReceiveMessage(platform_socket* socket, socket_msg* msg, socket_address* from); + +#ifdef __cplusplus +} // extern "C" +#endif + + +#endif //__PLATFORM_SOCKET_H_ diff --git a/src/platform/platform_thread.h b/src/platform/platform_thread.h new file mode 100644 index 0000000..0740748 --- /dev/null +++ b/src/platform/platform_thread.h @@ -0,0 +1,88 @@ +/************************************************************//** +* +* @file: platform_thread.h +* @author: Martin Fouilleul +* @date: 21/03/2019 +* @revision: +* +*****************************************************************/ +#ifndef __PLATFORM_THREAD_H_ +#define __PLATFORM_THREAD_H_ + + +#ifdef __cplusplus + #include + #define _Atomic(T) std::atomic +#else + #include +#endif + +#ifdef __cplusplus +extern "C" { +#endif // __cplusplus + +//--------------------------------------------------------------- +// Platform Thread API +//--------------------------------------------------------------- + +typedef struct platform_thread platform_thread; + +typedef void* (*ThreadStartFunction)(void* userPointer); + +platform_thread* ThreadCreate(ThreadStartFunction start, void* userPointer); +platform_thread* ThreadCreateWithName(ThreadStartFunction start, void* userPointer, const char* name); + +const char* ThreadGetName(platform_thread* thread); + +u64 ThreadSelfID(); +u64 ThreadUniqueID(platform_thread* thread); + +int ThreadSignal(platform_thread* thread, int sig); +void ThreadCancel(platform_thread* thread); +int ThreadJoin(platform_thread* thread, void** ret); +int ThreadDetach(platform_thread* thread); + +//--------------------------------------------------------------- +// Platform Mutex API +//--------------------------------------------------------------- + +typedef struct platform_mutex platform_mutex; + +platform_mutex* MutexCreate(); +int MutexDestroy(platform_mutex* mutex); +int MutexLock(platform_mutex* mutex); +int MutexUnlock(platform_mutex* mutex); + + +//--------------------------------------------------------------- +// Lightweight ticket mutex API +//--------------------------------------------------------------- + +typedef struct ticket_spin_mutex +{ + volatile _Atomic(u64) nextTicket; + volatile _Atomic(u64) serving; +} ticket_spin_mutex; + +void TicketSpinMutexInit(ticket_spin_mutex* mutex); +void TicketSpinMutexLock(ticket_spin_mutex* mutex); +void TicketSpinMutexUnlock(ticket_spin_mutex* mutex); + +//--------------------------------------------------------------- +// Platform condition variable API +//--------------------------------------------------------------- + +typedef struct platform_condition platform_condition; + +platform_condition* ConditionCreate(); +int ConditionDestroy(platform_condition* cond); +int ConditionWait(platform_condition* cond, platform_mutex* mutex); +int ConditionTimedWait(platform_condition* cond, platform_mutex* mutex, f64 seconds); +int ConditionSignal(platform_condition* cond); +int ConditionBroadcast(platform_condition* cond); + +#ifdef __cplusplus +} // extern "C" +#endif // __cplusplus + +#endif //__PLATFORM_THREAD_H_ diff --git a/src/platform/posix_socket.c b/src/platform/posix_socket.c new file mode 100644 index 0000000..81cb29b --- /dev/null +++ b/src/platform/posix_socket.c @@ -0,0 +1,515 @@ +/************************************************************//** +* +* @file: posix_socket.c +* @author: Martin Fouilleul +* @date: 22/03/2019 +* @revision: +* +*****************************************************************/ + +#include // socket() +#include // socaddr_in +#include // inet_addr() +#include // getifaddrs() / freeifaddrs() +#include // close() +#include // strerror() +#include // errno +#include // malloc()/free() + +#include"platform_socket.h" +#include"debug_log.h" + +typedef struct in_addr in_addr; +typedef struct sockaddr_in sockaddr_in; +typedef struct sockaddr sockaddr; +typedef struct msghdr msghdr; +typedef struct cmsghdr cmsghdr; +typedef struct ip_mreq ip_mreq; +typedef struct iovec iovec; + +#define LOG_SUBSYSTEM "Platform" + +net_ip StringToNetIP(const char* addr) +{ + return(inet_addr(addr)); +} +const char* NetIPToString(net_ip ip) +{ + in_addr in; + in.s_addr = ip; + return(inet_ntoa(in)); +} + +host_ip StringToHostIP(const char* addr) +{ + return(NetToHostIP(StringToNetIP(addr))); +} +const char* HostIPToString(host_ip ip) +{ + return(NetIPToString(HostToNetIP(ip))); +} + +net_ip HostToNetIP(uint32 ip) +{ + return(htonl(ip)); +} +net_port HostToNetPort(uint16 port) +{ + return(htons(port)); +} +uint32 NetToHostIP(net_ip ip) +{ + return(ntohl(ip)); +} +uint16 NetToHostPort(net_port port) +{ + return(ntohs(port)); +} + +static int PlatformToSocketFlags(int flags) +{ + int sflags = 0; + if(flags & SOCK_MSG_OOB) + { + sflags |= MSG_OOB; + } + if(flags & SOCK_MSG_PEEK) + { + sflags |= MSG_PEEK; + } + if(flags & SOCK_MSG_DONTROUTE) + { + sflags |= MSG_DONTROUTE; + } + if(flags & SOCK_MSG_WAITALL) + { + sflags |= MSG_WAITALL; + } + + return(sflags); +} + +static int ErrnoToSocketError(int err) +{ + //TODO(martin): extend these error codes + switch(err) + { + case 0: return(SOCK_ERR_OK); + case EACCES: return(SOCK_ERR_ACCESS); + case ENOBUFS: + case ENOMEM: return(SOCK_ERR_MEM); + case EINTR: return(SOCK_ERR_INTR); + case EADDRINUSE: return(SOCK_ERR_USED); + case EBADF: return(SOCK_ERR_BADF); + case ECONNABORTED: return(SOCK_ERR_ABORT); + + case EWOULDBLOCK: + #if EAGAIN != EWOULDBLOCK + case EAGAIN: + #endif + return(SOCK_ERR_NBLOCK); + + default: + return(SOCK_ERR_UNKNOWN); + } +} + +int SocketGetLastError() +{ + return(ErrnoToSocketError(errno)); +} + +const char* SocketGetLastErrorMessage() +{ + return(strerror(errno)); +} + +struct platform_socket +{ + int sd; +}; + +platform_socket* SocketOpen(socket_transport transport) +{ + int sd = 0; + switch(transport) + { + case SOCK_UDP: + sd = socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP); + break; + + case SOCK_TCP: + sd = socket(PF_INET, SOCK_STREAM, IPPROTO_TCP); + break; + } + if(!sd) + { + return(0); + } + + platform_socket* sock = (platform_socket*)malloc(sizeof(platform_socket)); + if(!sock) + { + close(sd); + return(0); + } + sock->sd = sd; + return(sock); +} + +int SocketClose(platform_socket* sock) +{ + if(!sock) + { + return(-1); + } + int res = close(sock->sd); + free(sock); + return(res); +} + +int SocketBind(platform_socket* sock, socket_address* addr) +{ + sockaddr_in saddr; + saddr.sin_addr.s_addr = addr->ip; + saddr.sin_port = addr->port; + saddr.sin_family = AF_INET; + + return(bind(sock->sd, (sockaddr*)&saddr, sizeof(saddr))); +} + +int SocketListen(platform_socket* sock, int backlog) +{ + return(listen(sock->sd, backlog)); +} +platform_socket* SocketAccept(platform_socket* sock, socket_address* from) +{ + sockaddr_in saddr; + socklen_t saddrSize = sizeof(saddr); + int sd = accept(sock->sd, (sockaddr*)&saddr, &saddrSize); + + from->ip = saddr.sin_addr.s_addr; + from->port = saddr.sin_port; + + if(sd <= 0) + { + return(0); + } + else + { + platform_socket* client = (platform_socket*)malloc(sizeof(platform_socket)); + if(!client) + { + close(sd); + return(0); + } + client->sd = sd; + return(client); + + } +} + +int SocketConnect(platform_socket* sock, socket_address* addr) +{ + sockaddr_in saddr; + saddr.sin_addr.s_addr = addr->ip; + saddr.sin_port = addr->port; + saddr.sin_family = AF_INET; + + return(connect(sock->sd, (sockaddr*)&saddr, sizeof(saddr))); +} + +int64 SocketReceive(platform_socket* sock, void* buffer, uint64 size, int flags) +{ + return(recv(sock->sd, buffer, size, PlatformToSocketFlags(flags))); +} + +int64 SocketReceiveFrom(platform_socket* sock, void* buffer, uint64 size, int flags, socket_address* from) +{ + sockaddr_in saddr; + socklen_t saddrSize = sizeof(saddr); + + int res = recvfrom(sock->sd, buffer, size, PlatformToSocketFlags(flags), (sockaddr*)&saddr, &saddrSize); + + from->ip = saddr.sin_addr.s_addr; + from->port = saddr.sin_port; + return(res); +} + +int64 SocketSend(platform_socket* sock, void* buffer, uint64 size, int flags) +{ + return(send(sock->sd, buffer, size, PlatformToSocketFlags(flags))); +} +int64 SocketSendTo(platform_socket* sock, void* buffer, uint64 size, int flags, socket_address* to) +{ + sockaddr_in saddr; + saddr.sin_addr.s_addr = to->ip; + saddr.sin_port = to->port; + saddr.sin_family = AF_INET; + + return(sendto(sock->sd, buffer, size, PlatformToSocketFlags(flags), (sockaddr*)&saddr, sizeof(saddr))); +} + + +int SocketSetReceiveTimeout(platform_socket* sock, timeval* tv) +{ + DEBUG_ASSERT(sock); + DEBUG_ASSERT(sock->sd); + return(setsockopt(sock->sd, SOL_SOCKET, SO_RCVTIMEO, tv, sizeof(timeval))); +} + +int SocketSetSendTimeout(platform_socket* sock, timeval* tv) +{ + DEBUG_ASSERT(sock); + DEBUG_ASSERT(sock->sd); + return(setsockopt(sock->sd, SOL_SOCKET, SO_SNDTIMEO, tv, sizeof(timeval))); +} + +int SocketSetReceiveTimestamping(platform_socket* socket, bool enable) +{ + int opt = enable ? 1 : 0; + socklen_t len = sizeof(int); + return(setsockopt(socket->sd, SOL_SOCKET, SO_TIMESTAMP, &enable, len)); +} + +int SocketSetBroadcast(platform_socket* sock, bool enable) +{ + DEBUG_ASSERT(sock); + DEBUG_ASSERT(sock->sd); + + int opt = enable ? 1 : 0; + return(setsockopt(sock->sd, SOL_SOCKET, SO_BROADCAST, &opt, sizeof(int))); +} + + +int SocketSelect(uint32 count, socket_activity* set, double timeout) +{ + fd_set fdInSet; + fd_set fdOutSet; + fd_set fdErrSet; + FD_ZERO(&fdInSet); + FD_ZERO(&fdOutSet); + FD_ZERO(&fdErrSet); + + int maxSd = -1; + for(int i=0; isock) + { + item->set = 0; + if(item->watch & SOCK_ACTIVITY_IN) + { + FD_SET(item->sock->sd, &fdInSet); + } + if(item->watch & SOCK_ACTIVITY_OUT) + { + FD_SET(item->sock->sd, &fdOutSet); + } + if(item->watch & SOCK_ACTIVITY_ERR) + { + FD_SET(item->sock->sd, &fdErrSet); + } + + if(item->watch && (item->sock->sd > maxSd)) + { + maxSd = item->sock->sd; + } + } + } + + if(maxSd <= 0) + { + return(0); + } + timeval tv; + tv.tv_sec = (time_t)timeout; + tv.tv_usec = (suseconds_t)((timeout - tv.tv_sec)*1000000); + + timeval* ptv = timeout >= 0 ? &tv : 0; + + int activity = select(maxSd+1, &fdInSet, &fdOutSet, &fdErrSet, ptv); + if(activity < 0) + { + return(-1); + } + + int processed = 0; + for(int i=0; i= activity) + { + break; + } + socket_activity* item = &(set[i]); + if(item->sock) + { + if(FD_ISSET(item->sock->sd, &fdInSet)) + { + item->set |= SOCK_ACTIVITY_IN; + } + if(FD_ISSET(item->sock->sd, &fdOutSet)) + { + item->set |= SOCK_ACTIVITY_OUT; + } + if(FD_ISSET(item->sock->sd, &fdErrSet)) + { + item->set |= SOCK_ACTIVITY_ERR; + } + if(item->set) + { + processed++; + } + } + } + return(activity); +} + + +int SocketGetAddress(platform_socket* sock, socket_address* addr) +{ + sockaddr_in saddr; + socklen_t sockLen = sizeof(saddr); + if(getsockname(sock->sd, (sockaddr*)&saddr, &sockLen)) + { + return(-1); + } + addr->ip = saddr.sin_addr.s_addr; + addr->port = saddr.sin_port; + + return(0); +} + + +int SocketGetIFAddresses(int* count, net_ip* ips) +{ + struct ifaddrs* ifaList = 0; + if(getifaddrs(&ifaList)) + { + return(-1); + } + + int maxCount = *count; + int i = 0; + + for(struct ifaddrs* ifa = ifaList; ifa != 0 ; ifa = ifa->ifa_next) + { + if(i >= maxCount) + { + freeifaddrs(ifaList); + *count = i; + return(-1); + } + if(ifa->ifa_addr->sa_family == AF_INET) + { + struct in_addr in = ((sockaddr_in*)ifa->ifa_addr)->sin_addr; + ips[i] = in.s_addr; + i++; + } + } + freeifaddrs(ifaList); + *count = i; + return(0); +} + + +int SocketSetReuseAddress(platform_socket* sock, bool enable) +{ + int reuse = enable ? 1 : 0; + return(setsockopt(sock->sd, SOL_SOCKET, SO_REUSEADDR, (char*)&reuse, sizeof(reuse))); +} + +int SocketSetReusePort(platform_socket* sock, bool enable) +{ + int reuse = enable ? 1 : 0; + return(setsockopt(sock->sd, SOL_SOCKET, SO_REUSEPORT, (char*)&reuse, sizeof(reuse))); +} + +int SocketSetMulticastLoop(platform_socket* sock, bool enable) +{ + int on = enable ? 1 : 0; + return(setsockopt(sock->sd, IPPROTO_IP, IP_MULTICAST_LOOP, (char*)&on, sizeof(on))); +} + +int SocketJoinMulticastGroup(platform_socket* sock, host_ip group, host_ip interface) +{ + ip_mreq mreq; + mreq.imr_multiaddr.s_addr = HostToNetIP(group); + mreq.imr_interface.s_addr = HostToNetIP(interface); + return(setsockopt(sock->sd, IPPROTO_IP, IP_ADD_MEMBERSHIP, (char*)&mreq, sizeof(mreq))); +} + +int SocketLeaveMulticastGroup(platform_socket* sock, host_ip group, host_ip interface) +{ + ip_mreq mreq; + mreq.imr_multiaddr.s_addr = HostToNetIP(group); + mreq.imr_interface.s_addr = HostToNetIP(interface); + return(setsockopt(sock->sd, IPPROTO_IP, IP_DROP_MEMBERSHIP, (char*)&mreq, sizeof(mreq))); +} + + +net_ip SocketGetDefaultExternalIP() +{ + //NOTE(martin): get the default local ip. This is a dumb way to do this + // 'cause I can't be bothered to think of a better way right now :( + + socket_address addr = {.ip = StringToNetIP("8.8.8.8"), + .port = HostToNetPort(5001)}; + + platform_socket* sock = SocketOpen(SOCK_UDP); + if(!sock) + { + LOG_ERROR("can't create socket"); + return(0); + } + + if(SocketConnect(sock, &addr) != 0) + { + LOG_ERROR("can't connect socket: %s\n", SocketGetLastErrorMessage()); + LOG_WARNING("try loopback interface\n"); + + addr.ip = HostToNetIP(SOCK_IP_LOOPBACK); + if(SocketConnect(sock, &addr) != 0) + { + LOG_ERROR("can't connect socket: %s\n", SocketGetLastErrorMessage()); + SocketClose(sock); + return(0); + } + } + SocketGetAddress(sock, &addr); + SocketClose(sock); + return(addr.ip); +} + +int SocketReceiveMessage(platform_socket* socket, socket_msg* msg, socket_address* from) +{ + sockaddr_in saddr; + + iovec iov; + iov.iov_base = msg->messageBuffer; + iov.iov_len = msg->messageBufferSize; + + msghdr hdr; + hdr.msg_name = &saddr; + hdr.msg_namelen = sizeof(saddr); + hdr.msg_iov = &iov; + hdr.msg_iovlen = 1; + hdr.msg_control = msg->controlBuffer; + hdr.msg_controllen = msg->controlBufferSize; + + int size = recvmsg(socket->sd, &hdr, 0); + if(size <= 0) + { + return(size); + } + + from->ip = saddr.sin_addr.s_addr; + from->port = saddr.sin_port; + + msg->controlBufferSize = hdr.msg_controllen; + msg->messageBufferSize = iov.iov_len; + + return(size); +} + +#undef LOG_SUBSYSTEM diff --git a/src/platform/posix_thread.c b/src/platform/posix_thread.c new file mode 100644 index 0000000..166f514 --- /dev/null +++ b/src/platform/posix_thread.c @@ -0,0 +1,241 @@ +/************************************************************//** +* +* @file: posix_thread.c +* @author: Martin Fouilleul +* @date: 21/03/2019 +* @revision: +* +*****************************************************************/ +#include +#include +#include //needed for pthread_kill() on linux +#include +#include + +#include"platform_thread.h" + +const u32 PLATFORM_THREAD_NAME_MAX_SIZE = 64; // including null terminator + +struct platform_thread +{ + bool valid; + pthread_t pthread; + ThreadStartFunction start; + void* userPointer; + char name[PLATFORM_THREAD_NAME_MAX_SIZE]; +}; + +void* platform_thread_bootstrap(void* data) +{ + platform_thread* thread = (platform_thread*)data; + if(strlen(thread->name)) + { + pthread_setname_np(thread->name); + } + return(thread->start(thread->userPointer)); +} + +platform_thread* ThreadCreateWithName(ThreadStartFunction start, void* userPointer, const char* name) +{ + platform_thread* thread = (platform_thread*)malloc(sizeof(platform_thread)); + if(!thread) + { + return(0); + } + + if(name) + { + char* end = stpncpy(thread->name, name, PLATFORM_THREAD_NAME_MAX_SIZE-1); + *end = '\0'; + } + else + { + thread->name[0] = '\0'; + } + thread->start = start; + thread->userPointer = userPointer; + + if(pthread_create(&thread->pthread, 0, platform_thread_bootstrap, thread) != 0) + { + free(thread); + return(0); + } + else + { + thread->valid = true; + return(thread); + } +} + +platform_thread* ThreadCreate(ThreadStartFunction start, void* userPointer) +{ + return(ThreadCreateWithName(start, userPointer, 0)); +} + +void ThreadCancel(platform_thread* thread) +{ + pthread_cancel(thread->pthread); +} + +const char* ThreadGetName(platform_thread* thread) +{ + return(thread->name); +} + + +u64 ThreadUniqueID(platform_thread* thread) +{ + u64 id; + pthread_threadid_np(thread->pthread, &id); + return(id); +} + +u64 ThreadSelfID() +{ + pthread_t thread = pthread_self(); + u64 id; + pthread_threadid_np(thread, &id); + return(id); +} + +int ThreadSignal(platform_thread* thread, int sig) +{ + return(pthread_kill(thread->pthread, sig)); +} + +int ThreadJoin(platform_thread* thread, void** ret) +{ + if(pthread_join(thread->pthread, ret)) + { + return(-1); + } + free(thread); + return(0); +} + +int ThreadDetach(platform_thread* thread) +{ + if(pthread_detach(thread->pthread)) + { + return(-1); + } + free(thread); + return(0); +} + + +struct platform_mutex +{ + pthread_mutex_t pmutex; +}; + +platform_mutex* MutexCreate() +{ + platform_mutex* mutex = (platform_mutex*)malloc(sizeof(platform_mutex)); + if(!mutex) + { + return(0); + } + if(pthread_mutex_init(&mutex->pmutex, 0) != 0) + { + free(mutex); + return(0); + } + return(mutex); +} +int MutexDestroy(platform_mutex* mutex) +{ + if(pthread_mutex_destroy(&mutex->pmutex) != 0) + { + return(-1); + } + free(mutex); + return(0); +} + +int MutexLock(platform_mutex* mutex) +{ + return(pthread_mutex_lock(&mutex->pmutex)); +} + +int MutexUnlock(platform_mutex* mutex) +{ + return(pthread_mutex_unlock(&mutex->pmutex)); +} + +void TicketSpinMutexInit(ticket_spin_mutex* mutex) +{ + mutex->nextTicket = 0; + mutex->serving = 0; +} + +void TicketSpinMutexLock(ticket_spin_mutex* mutex) +{ + u64 ticket = atomic_fetch_add(&mutex->nextTicket, 1ULL); + while(ticket != mutex->serving); //spin +} + +void TicketSpinMutexUnlock(ticket_spin_mutex* mutex) +{ + atomic_fetch_add(&mutex->serving, 1ULL); +} + +struct platform_condition +{ + pthread_cond_t pcond; +}; + +platform_condition* ConditionCreate() +{ + platform_condition* cond = (platform_condition*)malloc(sizeof(platform_condition)); + if(!cond) + { + return(0); + } + if(pthread_cond_init(&cond->pcond, 0) != 0) + { + free(cond); + return(0); + } + return(cond); +} +int ConditionDestroy(platform_condition* cond) +{ + if(pthread_cond_destroy(&cond->pcond) != 0) + { + return(-1); + } + free(cond); + return(0); +} +int ConditionWait(platform_condition* cond, platform_mutex* mutex) +{ + return(pthread_cond_wait(&cond->pcond, &mutex->pmutex)); +} + +int ConditionTimedWait(platform_condition* cond, platform_mutex* mutex, f64 seconds) +{ + struct timeval tv; + gettimeofday(&tv, 0); + + i64 iSeconds = (i64)seconds; + f64 fracSeconds = seconds - (f64)iSeconds; + + struct timespec ts; + ts.tv_sec = tv.tv_sec + iSeconds; + ts.tv_nsec = tv.tv_usec * 1000 + (i32)(fracSeconds*1e9); + ts.tv_sec += ts.tv_nsec / 1000000000; + ts.tv_nsec = ts.tv_nsec % 1000000000; + + return(pthread_cond_timedwait(&cond->pcond, &mutex->pmutex, &ts)); +} + +int ConditionSignal(platform_condition* cond) +{ + return(pthread_cond_signal(&cond->pcond)); +} + +int ConditionBroadcast(platform_condition* cond) +{ + return(pthread_cond_broadcast(&cond->pcond)); +} diff --git a/src/platform/unix_base_allocator.c b/src/platform/unix_base_allocator.c new file mode 100644 index 0000000..839dae7 --- /dev/null +++ b/src/platform/unix_base_allocator.c @@ -0,0 +1,38 @@ +/************************************************************//** +* +* @file: unix_base_allocator.c +* @author: Martin Fouilleul +* @date: 10/09/2021 +* @revision: +* +*****************************************************************/ +#include +#include"platform_base_allocator.h" + +/*NOTE(martin): + Linux and MacOS don't make a distinction between reserved and committed memory, contrary to Windows +*/ +void mem_base_nop(void* context, void* ptr, u64 size) {} + +void* mem_base_reserve_mmap(void* context, u64 size) +{ + return(mmap(0, size, PROT_READ|PROT_WRITE, MAP_ANON|MAP_PRIVATE, 0, 0)); +} + +void mem_base_release_mmap(void* context, void* ptr, u64 size) +{ + munmap(ptr, size); +} + +mem_base_allocator* mem_base_allocator_default() +{ + static mem_base_allocator base = {}; + if(base.reserve == 0) + { + base.reserve = mem_base_reserve_mmap; + base.commit = mem_base_nop; + base.decommit = mem_base_nop; + base.release = mem_base_release_mmap; + } + return(&base); +} diff --git a/src/platform/unix_rng.c b/src/platform/unix_rng.c new file mode 100644 index 0000000..ba06283 --- /dev/null +++ b/src/platform/unix_rng.c @@ -0,0 +1,60 @@ +/************************************************************//** +* +* @file: unix_rng.c +* @author: Martin Fouilleul +* @date: 06/03/2020 +* @revision: +* +*****************************************************************/ + +#include +#include + +#include"debug_log.h" +#include"typedefs.h" + +#define LOG_SUBSYSTEM "Platform" + +int RandomSeedFromDevice() +{ + FILE* urandom = fopen("/dev/urandom", "r"); + if(!urandom) + { + LOG_ERROR("can't open /dev/urandom\n"); + return(-1); + } + + union + { + u32 u; + char buff[4]; + } seed; + + int size = fread(seed.buff, 1, 4, urandom); + if(size != 4) + { + LOG_ERROR("couldn't read from /dev/urandom\n"); + return(-1); + } + + fclose(urandom); + srandom(seed.u); + return(0); +} + +u32 RandomU32() +{ + u32 u1 = (u32)random(); + u32 u2 = (u32)random(); + return((u1<<1) | (u2 & 0x01)); +} + +u64 RandomU64() +{ + u64 u1 = (u64)random(); + u64 u2 = (u64)random(); + u64 u3 = (u64)random(); + return((u1<<33) | (u2<<2) | (u3 & 0x03)); +} + +#undef LOG_SUBSYSTEM diff --git a/src/stb_image.h b/src/stb_image.h new file mode 100644 index 0000000..81b2fa2 --- /dev/null +++ b/src/stb_image.h @@ -0,0 +1,7440 @@ +/* stb_image - v2.19 - public domain image loader - http://nothings.org/stb + no warranty implied; use at your own risk + Do this: + #define STB_IMAGE_IMPLEMENTATION + before you include this file in *one* C or C++ file to create the implementation. + // i.e. it should look like this: + #include ... + #include ... + #include ... + #define STB_IMAGE_IMPLEMENTATION + #include "stb_image.h" + You can #define STBI_ASSERT(x) before the #include to avoid using assert.h. + And #define STBI_MALLOC, STBI_REALLOC, and STBI_FREE to avoid using malloc,realloc,free + QUICK NOTES: + Primarily of interest to game developers and other people who can + avoid problematic images and only need the trivial interface + JPEG baseline & progressive (12 bpc/arithmetic not supported, same as stock IJG lib) + PNG 1/2/4/8/16-bit-per-channel + TGA (not sure what subset, if a subset) + BMP non-1bpp, non-RLE + PSD (composited view only, no extra channels, 8/16 bit-per-channel) + GIF (*comp always reports as 4-channel) + HDR (radiance rgbE format) + PIC (Softimage PIC) + PNM (PPM and PGM binary only) + Animated GIF still needs a proper API, but here's one way to do it: + http://gist.github.com/urraka/685d9a6340b26b830d49 + - decode from memory or through FILE (define STBI_NO_STDIO to remove code) + - decode from arbitrary I/O callbacks + - SIMD acceleration on x86/x64 (SSE2) and ARM (NEON) + Full documentation under "DOCUMENTATION" below. +LICENSE + See end of file for license information. +RECENT REVISION HISTORY: + 2.19 (2018-02-11) fix warning + 2.18 (2018-01-30) fix warnings + 2.17 (2018-01-29) bugfix, 1-bit BMP, 16-bitness query, fix warnings + 2.16 (2017-07-23) all functions have 16-bit variants; optimizations; bugfixes + 2.15 (2017-03-18) fix png-1,2,4; all Imagenet JPGs; no runtime SSE detection on GCC + 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs + 2.13 (2016-12-04) experimental 16-bit API, only for PNG so far; fixes + 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes + 2.11 (2016-04-02) 16-bit PNGS; enable SSE2 in non-gcc x64 + RGB-format JPEG; remove white matting in PSD; + allocate large structures on the stack; + correct channel count for PNG & BMP + 2.10 (2016-01-22) avoid warning introduced in 2.09 + 2.09 (2016-01-16) 16-bit TGA; comments in PNM files; STBI_REALLOC_SIZED + See end of file for full revision history. + ============================ Contributors ========================= + Image formats Extensions, features + Sean Barrett (jpeg, png, bmp) Jetro Lauha (stbi_info) + Nicolas Schulz (hdr, psd) Martin "SpartanJ" Golini (stbi_info) + Jonathan Dummer (tga) James "moose2000" Brown (iPhone PNG) + Jean-Marc Lienher (gif) Ben "Disch" Wenger (io callbacks) + Tom Seddon (pic) Omar Cornut (1/2/4-bit PNG) + Thatcher Ulrich (psd) Nicolas Guillemot (vertical flip) + Ken Miller (pgm, ppm) Richard Mitton (16-bit PSD) + github:urraka (animated gif) Junggon Kim (PNM comments) + Christopher Forseth (animated gif) Daniel Gibson (16-bit TGA) + socks-the-fox (16-bit PNG) + Jeremy Sawicki (handle all ImageNet JPGs) + Optimizations & bugfixes Mikhail Morozov (1-bit BMP) + Fabian "ryg" Giesen Anael Seghezzi (is-16-bit query) + Arseny Kapoulkine + John-Mark Allen + Bug & warning fixes + Marc LeBlanc David Woo Guillaume George Martins Mozeiko + Christpher Lloyd Jerry Jansson Joseph Thomson Phil Jordan + Dave Moore Roy Eltham Hayaki Saito Nathan Reed + Won Chun Luke Graham Johan Duparc Nick Verigakis + the Horde3D community Thomas Ruf Ronny Chevalier github:rlyeh + Janez Zemva John Bartholomew Michal Cichon github:romigrou + Jonathan Blow Ken Hamada Tero Hanninen github:svdijk + Laurent Gomila Cort Stratton Sergio Gonzalez github:snagar + Aruelien Pocheville Thibault Reuille Cass Everitt github:Zelex + Ryamond Barbiero Paul Du Bois Engin Manap github:grim210 + Aldo Culquicondor Philipp Wiesemann Dale Weiler github:sammyhw + Oriol Ferrer Mesia Josh Tobin Matthew Gregan github:phprus + Julian Raschke Gregory Mullen Baldur Karlsson github:poppolopoppo + Christian Floisand Kevin Schmidt github:darealshinji + Blazej Dariusz Roszkowski github:Michaelangel007 +*/ + +#ifndef STBI_INCLUDE_STB_IMAGE_H +#define STBI_INCLUDE_STB_IMAGE_H + +// DOCUMENTATION +// +// Limitations: +// - no 12-bit-per-channel JPEG +// - no JPEGs with arithmetic coding +// - GIF always returns *comp=4 +// +// Basic usage (see HDR discussion below for HDR usage): +// int x,y,n; +// unsigned char *data = stbi_load(filename, &x, &y, &n, 0); +// // ... process data if not NULL ... +// // ... x = width, y = height, n = # 8-bit components per pixel ... +// // ... replace '0' with '1'..'4' to force that many components per pixel +// // ... but 'n' will always be the number that it would have been if you said 0 +// stbi_image_free(data) +// +// Standard parameters: +// int *x -- outputs image width in pixels +// int *y -- outputs image height in pixels +// int *channels_in_file -- outputs # of image components in image file +// int desired_channels -- if non-zero, # of image components requested in result +// +// The return value from an image loader is an 'unsigned char *' which points +// to the pixel data, or NULL on an allocation failure or if the image is +// corrupt or invalid. The pixel data consists of *y scanlines of *x pixels, +// with each pixel consisting of N interleaved 8-bit components; the first +// pixel pointed to is top-left-most in the image. There is no padding between +// image scanlines or between pixels, regardless of format. The number of +// components N is 'desired_channels' if desired_channels is non-zero, or +// *channels_in_file otherwise. If desired_channels is non-zero, +// *channels_in_file has the number of components that _would_ have been +// output otherwise. E.g. if you set desired_channels to 4, you will always +// get RGBA output, but you can check *channels_in_file to see if it's trivially +// opaque because e.g. there were only 3 channels in the source image. +// +// An output image with N components has the following components interleaved +// in this order in each pixel: +// +// N=#comp components +// 1 grey +// 2 grey, alpha +// 3 red, green, blue +// 4 red, green, blue, alpha +// +// If image loading fails for any reason, the return value will be NULL, +// and *x, *y, *channels_in_file will be unchanged. The function +// stbi_failure_reason() can be queried for an extremely brief, end-user +// unfriendly explanation of why the load failed. Define STBI_NO_FAILURE_STRINGS +// to avoid compiling these strings at all, and STBI_FAILURE_USERMSG to get slightly +// more user-friendly ones. +// +// Paletted PNG, BMP, GIF, and PIC images are automatically depalettized. +// +// =========================================================================== +// +// Philosophy +// +// stb libraries are designed with the following priorities: +// +// 1. easy to use +// 2. easy to maintain +// 3. good performance +// +// Sometimes I let "good performance" creep up in priority over "easy to maintain", +// and for best performance I may provide less-easy-to-use APIs that give higher +// performance, in addition to the easy to use ones. Nevertheless, it's important +// to keep in mind that from the standpoint of you, a client of this library, +// all you care about is #1 and #3, and stb libraries DO NOT emphasize #3 above all. +// +// Some secondary priorities arise directly from the first two, some of which +// make more explicit reasons why performance can't be emphasized. +// +// - Portable ("ease of use") +// - Small source code footprint ("easy to maintain") +// - No dependencies ("ease of use") +// +// =========================================================================== +// +// I/O callbacks +// +// I/O callbacks allow you to read from arbitrary sources, like packaged +// files or some other source. Data read from callbacks are processed +// through a small internal buffer (currently 128 bytes) to try to reduce +// overhead. +// +// The three functions you must define are "read" (reads some bytes of data), +// "skip" (skips some bytes of data), "eof" (reports if the stream is at the end). +// +// =========================================================================== +// +// SIMD support +// +// The JPEG decoder will try to automatically use SIMD kernels on x86 when +// supported by the compiler. For ARM Neon support, you must explicitly +// request it. +// +// (The old do-it-yourself SIMD API is no longer supported in the current +// code.) +// +// On x86, SSE2 will automatically be used when available based on a run-time +// test; if not, the generic C versions are used as a fall-back. On ARM targets, +// the typical path is to have separate builds for NEON and non-NEON devices +// (at least this is true for iOS and Android). Therefore, the NEON support is +// toggled by a build flag: define STBI_NEON to get NEON loops. +// +// If for some reason you do not want to use any of SIMD code, or if +// you have issues compiling it, you can disable it entirely by +// defining STBI_NO_SIMD. +// +// =========================================================================== +// +// HDR image support (disable by defining STBI_NO_HDR) +// +// stb_image now supports loading HDR images in general, and currently +// the Radiance .HDR file format, although the support is provided +// generically. You can still load any file through the existing interface; +// if you attempt to load an HDR file, it will be automatically remapped to +// LDR, assuming gamma 2.2 and an arbitrary scale factor defaulting to 1; +// both of these constants can be reconfigured through this interface: +// +// stbi_hdr_to_ldr_gamma(2.2f); +// stbi_hdr_to_ldr_scale(1.0f); +// +// (note, do not use _inverse_ constants; stbi_image will invert them +// appropriately). +// +// Additionally, there is a new, parallel interface for loading files as +// (linear) floats to preserve the full dynamic range: +// +// float *data = stbi_loadf(filename, &x, &y, &n, 0); +// +// If you load LDR images through this interface, those images will +// be promoted to floating point values, run through the inverse of +// constants corresponding to the above: +// +// stbi_ldr_to_hdr_scale(1.0f); +// stbi_ldr_to_hdr_gamma(2.2f); +// +// Finally, given a filename (or an open file or memory block--see header +// file for details) containing image data, you can query for the "most +// appropriate" interface to use (that is, whether the image is HDR or +// not), using: +// +// stbi_is_hdr(char *filename); +// +// =========================================================================== +// +// iPhone PNG support: +// +// By default we convert iphone-formatted PNGs back to RGB, even though +// they are internally encoded differently. You can disable this conversion +// by by calling stbi_convert_iphone_png_to_rgb(0), in which case +// you will always just get the native iphone "format" through (which +// is BGR stored in RGB). +// +// Call stbi_set_unpremultiply_on_load(1) as well to force a divide per +// pixel to remove any premultiplied alpha *only* if the image file explicitly +// says there's premultiplied data (currently only happens in iPhone images, +// and only if iPhone convert-to-rgb processing is on). +// +// =========================================================================== +// +// ADDITIONAL CONFIGURATION +// +// - You can suppress implementation of any of the decoders to reduce +// your code footprint by #defining one or more of the following +// symbols before creating the implementation. +// +// STBI_NO_JPEG +// STBI_NO_PNG +// STBI_NO_BMP +// STBI_NO_PSD +// STBI_NO_TGA +// STBI_NO_GIF +// STBI_NO_HDR +// STBI_NO_PIC +// STBI_NO_PNM (.ppm and .pgm) +// +// - You can request *only* certain decoders and suppress all other ones +// (this will be more forward-compatible, as addition of new decoders +// doesn't require you to disable them explicitly): +// +// STBI_ONLY_JPEG +// STBI_ONLY_PNG +// STBI_ONLY_BMP +// STBI_ONLY_PSD +// STBI_ONLY_TGA +// STBI_ONLY_GIF +// STBI_ONLY_HDR +// STBI_ONLY_PIC +// STBI_ONLY_PNM (.ppm and .pgm) +// +// - If you use STBI_NO_PNG (or _ONLY_ without PNG), and you still +// want the zlib decoder to be available, #define STBI_SUPPORT_ZLIB +// + + +#ifndef STBI_NO_STDIO +#include +#endif // STBI_NO_STDIO + +#define STBI_VERSION 1 + +enum +{ + STBI_default = 0, // only used for desired_channels + + STBI_grey = 1, + STBI_grey_alpha = 2, + STBI_rgb = 3, + STBI_rgb_alpha = 4 +}; + +typedef unsigned char stbi_uc; +typedef unsigned short stbi_us; + +#ifdef __cplusplus +extern "C" { +#endif + +#ifdef STB_IMAGE_STATIC +#define STBIDEF static +#else +#define STBIDEF extern +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// PRIMARY API - works on images of any type +// + +// +// load image by filename, open file, or memory buffer +// + +typedef struct +{ + int (*read) (void *user,char *data,int size); // fill 'data' with 'size' bytes. return number of bytes actually read + void (*skip) (void *user,int n); // skip the next 'n' bytes, or 'unget' the last -n bytes if negative + int (*eof) (void *user); // returns nonzero if we are at end of file/data +} stbi_io_callbacks; + +//////////////////////////////////// +// +// 8-bits-per-channel interface +// + +STBIDEF stbi_uc *stbi_load_from_memory (stbi_uc const *buffer, int len , int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk , void *user, int *x, int *y, int *channels_in_file, int desired_channels); +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +#endif + + +#ifndef STBI_NO_STDIO +STBIDEF stbi_uc *stbi_load (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +// for stbi_load_from_file, file pointer is left pointing immediately after image +#endif + +//////////////////////////////////// +// +// 16-bits-per-channel interface +// + +STBIDEF stbi_us *stbi_load_16_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + +#ifndef STBI_NO_STDIO +STBIDEF stbi_us *stbi_load_16 (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_from_file_16(FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +#endif + +//////////////////////////////////// +// +// float-per-channel interface +// +#ifndef STBI_NO_LINEAR + STBIDEF float *stbi_loadf_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); + STBIDEF float *stbi_loadf_from_callbacks (stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + + #ifndef STBI_NO_STDIO + STBIDEF float *stbi_loadf (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); + STBIDEF float *stbi_loadf_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); + #endif +#endif + +#ifndef STBI_NO_HDR + STBIDEF void stbi_hdr_to_ldr_gamma(float gamma); + STBIDEF void stbi_hdr_to_ldr_scale(float scale); +#endif // STBI_NO_HDR + +#ifndef STBI_NO_LINEAR + STBIDEF void stbi_ldr_to_hdr_gamma(float gamma); + STBIDEF void stbi_ldr_to_hdr_scale(float scale); +#endif // STBI_NO_LINEAR + +// stbi_is_hdr is always defined, but always returns false if STBI_NO_HDR +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user); +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len); +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename); +STBIDEF int stbi_is_hdr_from_file(FILE *f); +#endif // STBI_NO_STDIO + + +// get a VERY brief reason for failure +// NOT THREADSAFE +STBIDEF const char *stbi_failure_reason (void); + +// free the loaded image -- this is just free() +STBIDEF void stbi_image_free (void *retval_from_stbi_load); + +// get image dimensions & components without fully decoding +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len); +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *clbk, void *user); + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info (char const *filename, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_file (FILE *f, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit (char const *filename); +STBIDEF int stbi_is_16_bit_from_file(FILE *f); +#endif + + + +// for image formats that explicitly notate that they have premultiplied alpha, +// we just return the colors as stored in the file. set this flag to force +// unpremultiplication. results are undefined if the unpremultiply overflow. +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply); + +// indicate whether we should process iphone images back to canonical format, +// or just pass them through "as-is" +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert); + +// flip the image vertically, so the first pixel in the output array is the bottom left +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip); + +// ZLIB client - used by PNG, available for other purposes + +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen); +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header); +STBIDEF char *stbi_zlib_decode_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); + +STBIDEF char *stbi_zlib_decode_noheader_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); + + +#ifdef __cplusplus +} +#endif + +// +// +//// end header file ///////////////////////////////////////////////////// +#endif // STBI_INCLUDE_STB_IMAGE_H + +#ifdef STB_IMAGE_IMPLEMENTATION + +#if defined(STBI_ONLY_JPEG) || defined(STBI_ONLY_PNG) || defined(STBI_ONLY_BMP) \ + || defined(STBI_ONLY_TGA) || defined(STBI_ONLY_GIF) || defined(STBI_ONLY_PSD) \ + || defined(STBI_ONLY_HDR) || defined(STBI_ONLY_PIC) || defined(STBI_ONLY_PNM) \ + || defined(STBI_ONLY_ZLIB) + #ifndef STBI_ONLY_JPEG + #define STBI_NO_JPEG + #endif + #ifndef STBI_ONLY_PNG + #define STBI_NO_PNG + #endif + #ifndef STBI_ONLY_BMP + #define STBI_NO_BMP + #endif + #ifndef STBI_ONLY_PSD + #define STBI_NO_PSD + #endif + #ifndef STBI_ONLY_TGA + #define STBI_NO_TGA + #endif + #ifndef STBI_ONLY_GIF + #define STBI_NO_GIF + #endif + #ifndef STBI_ONLY_HDR + #define STBI_NO_HDR + #endif + #ifndef STBI_ONLY_PIC + #define STBI_NO_PIC + #endif + #ifndef STBI_ONLY_PNM + #define STBI_NO_PNM + #endif +#endif + +#if defined(STBI_NO_PNG) && !defined(STBI_SUPPORT_ZLIB) && !defined(STBI_NO_ZLIB) +#define STBI_NO_ZLIB +#endif + + +#include +#include // ptrdiff_t on osx +#include +#include +#include + +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +#include // ldexp, pow +#endif + +#ifndef STBI_NO_STDIO +#include +#endif + +#ifndef STBI_ASSERT +#include +#define STBI_ASSERT(x) assert(x) +#endif + + +#ifndef _MSC_VER + #ifdef __cplusplus + #define stbi_inline inline + #else + #define stbi_inline + #endif +#else + #define stbi_inline __forceinline +#endif + + +#ifdef _MSC_VER +typedef unsigned short stbi__uint16; +typedef signed short stbi__int16; +typedef unsigned int stbi__uint32; +typedef signed int stbi__int32; +#else +#include +typedef uint16_t stbi__uint16; +typedef int16_t stbi__int16; +typedef uint32_t stbi__uint32; +typedef int32_t stbi__int32; +#endif + +// should produce compiler error if size is wrong +typedef unsigned char validate_uint32[sizeof(stbi__uint32)==4 ? 1 : -1]; + +#ifdef _MSC_VER +#define STBI_NOTUSED(v) (void)(v) +#else +#define STBI_NOTUSED(v) (void)sizeof(v) +#endif + +#ifdef _MSC_VER +#define STBI_HAS_LROTL +#endif + +#ifdef STBI_HAS_LROTL + #define stbi_lrot(x,y) _lrotl(x,y) +#else + #define stbi_lrot(x,y) (((x) << (y)) | ((x) >> (32 - (y)))) +#endif + +#if defined(STBI_MALLOC) && defined(STBI_FREE) && (defined(STBI_REALLOC) || defined(STBI_REALLOC_SIZED)) +// ok +#elif !defined(STBI_MALLOC) && !defined(STBI_FREE) && !defined(STBI_REALLOC) && !defined(STBI_REALLOC_SIZED) +// ok +#else +#error "Must define all or none of STBI_MALLOC, STBI_FREE, and STBI_REALLOC (or STBI_REALLOC_SIZED)." +#endif + +#ifndef STBI_MALLOC +#define STBI_MALLOC(sz) malloc(sz) +#define STBI_REALLOC(p,newsz) realloc(p,newsz) +#define STBI_FREE(p) free(p) +#endif + +#ifndef STBI_REALLOC_SIZED +#define STBI_REALLOC_SIZED(p,oldsz,newsz) STBI_REALLOC(p,newsz) +#endif + +// x86/x64 detection +#if defined(__x86_64__) || defined(_M_X64) +#define STBI__X64_TARGET +#elif defined(__i386) || defined(_M_IX86) +#define STBI__X86_TARGET +#endif + +#if defined(__GNUC__) && defined(STBI__X86_TARGET) && !defined(__SSE2__) && !defined(STBI_NO_SIMD) +// gcc doesn't support sse2 intrinsics unless you compile with -msse2, +// which in turn means it gets to use SSE2 everywhere. This is unfortunate, +// but previous attempts to provide the SSE2 functions with runtime +// detection caused numerous issues. The way architecture extensions are +// exposed in GCC/Clang is, sadly, not really suited for one-file libs. +// New behavior: if compiled with -msse2, we use SSE2 without any +// detection; if not, we don't use it at all. +#define STBI_NO_SIMD +#endif + +#if defined(__MINGW32__) && defined(STBI__X86_TARGET) && !defined(STBI_MINGW_ENABLE_SSE2) && !defined(STBI_NO_SIMD) +// Note that __MINGW32__ doesn't actually mean 32-bit, so we have to avoid STBI__X64_TARGET +// +// 32-bit MinGW wants ESP to be 16-byte aligned, but this is not in the +// Windows ABI and VC++ as well as Windows DLLs don't maintain that invariant. +// As a result, enabling SSE2 on 32-bit MinGW is dangerous when not +// simultaneously enabling "-mstackrealign". +// +// See https://github.com/nothings/stb/issues/81 for more information. +// +// So default to no SSE2 on 32-bit MinGW. If you've read this far and added +// -mstackrealign to your build settings, feel free to #define STBI_MINGW_ENABLE_SSE2. +#define STBI_NO_SIMD +#endif + +#if !defined(STBI_NO_SIMD) && (defined(STBI__X86_TARGET) || defined(STBI__X64_TARGET)) +#define STBI_SSE2 +#include + +#ifdef _MSC_VER + +#if _MSC_VER >= 1400 // not VC6 +#include // __cpuid +static int stbi__cpuid3(void) +{ + int info[4]; + __cpuid(info,1); + return info[3]; +} +#else +static int stbi__cpuid3(void) +{ + int res; + __asm { + mov eax,1 + cpuid + mov res,edx + } + return res; +} +#endif + +#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name + +static int stbi__sse2_available(void) +{ + int info3 = stbi__cpuid3(); + return ((info3 >> 26) & 1) != 0; +} +#else // assume GCC-style if not VC++ +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) + +static int stbi__sse2_available(void) +{ + // If we're even attempting to compile this on GCC/Clang, that means + // -msse2 is on, which means the compiler is allowed to use SSE2 + // instructions at will, and so are we. + return 1; +} +#endif +#endif + +// ARM NEON +#if defined(STBI_NO_SIMD) && defined(STBI_NEON) +#undef STBI_NEON +#endif + +#ifdef STBI_NEON +#include +// assume GCC or Clang on ARM targets +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) +#endif + +#ifndef STBI_SIMD_ALIGN +#define STBI_SIMD_ALIGN(type, name) type name +#endif + +/////////////////////////////////////////////// +// +// stbi__context struct and start_xxx functions + +// stbi__context structure is our basic context used by all images, so it +// contains all the IO context, plus some basic image information +typedef struct +{ + stbi__uint32 img_x, img_y; + int img_n, img_out_n; + + stbi_io_callbacks io; + void *io_user_data; + + int read_from_callbacks; + int buflen; + stbi_uc buffer_start[128]; + + stbi_uc *img_buffer, *img_buffer_end; + stbi_uc *img_buffer_original, *img_buffer_original_end; +} stbi__context; + + +static void stbi__refill_buffer(stbi__context *s); + +// initialize a memory-decode context +static void stbi__start_mem(stbi__context *s, stbi_uc const *buffer, int len) +{ + s->io.read = NULL; + s->read_from_callbacks = 0; + s->img_buffer = s->img_buffer_original = (stbi_uc *) buffer; + s->img_buffer_end = s->img_buffer_original_end = (stbi_uc *) buffer+len; +} + +// initialize a callback-based context +static void stbi__start_callbacks(stbi__context *s, stbi_io_callbacks *c, void *user) +{ + s->io = *c; + s->io_user_data = user; + s->buflen = sizeof(s->buffer_start); + s->read_from_callbacks = 1; + s->img_buffer_original = s->buffer_start; + stbi__refill_buffer(s); + s->img_buffer_original_end = s->img_buffer_end; +} + +#ifndef STBI_NO_STDIO + +static int stbi__stdio_read(void *user, char *data, int size) +{ + return (int) fread(data,1,size,(FILE*) user); +} + +static void stbi__stdio_skip(void *user, int n) +{ + fseek((FILE*) user, n, SEEK_CUR); +} + +static int stbi__stdio_eof(void *user) +{ + return feof((FILE*) user); +} + +static stbi_io_callbacks stbi__stdio_callbacks = +{ + stbi__stdio_read, + stbi__stdio_skip, + stbi__stdio_eof, +}; + +static void stbi__start_file(stbi__context *s, FILE *f) +{ + stbi__start_callbacks(s, &stbi__stdio_callbacks, (void *) f); +} + +//static void stop_file(stbi__context *s) { } + +#endif // !STBI_NO_STDIO + +static void stbi__rewind(stbi__context *s) +{ + // conceptually rewind SHOULD rewind to the beginning of the stream, + // but we just rewind to the beginning of the initial buffer, because + // we only use it after doing 'test', which only ever looks at at most 92 bytes + s->img_buffer = s->img_buffer_original; + s->img_buffer_end = s->img_buffer_original_end; +} + +enum +{ + STBI_ORDER_RGB, + STBI_ORDER_BGR +}; + +typedef struct +{ + int bits_per_channel; + int num_channels; + int channel_order; +} stbi__result_info; + +#ifndef STBI_NO_JPEG +static int stbi__jpeg_test(stbi__context *s); +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PNG +static int stbi__png_test(stbi__context *s); +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__png_is16(stbi__context *s); +#endif + +#ifndef STBI_NO_BMP +static int stbi__bmp_test(stbi__context *s); +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_TGA +static int stbi__tga_test(stbi__context *s); +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s); +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc); +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__psd_is16(stbi__context *s); +#endif + +#ifndef STBI_NO_HDR +static int stbi__hdr_test(stbi__context *s); +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PIC +static int stbi__pic_test(stbi__context *s); +static void *stbi__pic_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_GIF +static int stbi__gif_test(stbi__context *s); +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PNM +static int stbi__pnm_test(stbi__context *s); +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +// this is not threadsafe +static const char *stbi__g_failure_reason; + +STBIDEF const char *stbi_failure_reason(void) +{ + return stbi__g_failure_reason; +} + +static int stbi__err(const char *str) +{ + stbi__g_failure_reason = str; + return 0; +} + +static void *stbi__malloc(size_t size) +{ + return STBI_MALLOC(size); +} + +// stb_image uses ints pervasively, including for offset calculations. +// therefore the largest decoded image size we can support with the +// current code, even on 64-bit targets, is INT_MAX. this is not a +// significant limitation for the intended use case. +// +// we do, however, need to make sure our size calculations don't +// overflow. hence a few helper functions for size calculations that +// multiply integers together, making sure that they're non-negative +// and no overflow occurs. + +// return 1 if the sum is valid, 0 on overflow. +// negative terms are considered invalid. +static int stbi__addsizes_valid(int a, int b) +{ + if (b < 0) return 0; + // now 0 <= b <= INT_MAX, hence also + // 0 <= INT_MAX - b <= INTMAX. + // And "a + b <= INT_MAX" (which might overflow) is the + // same as a <= INT_MAX - b (no overflow) + return a <= INT_MAX - b; +} + +// returns 1 if the product is valid, 0 on overflow. +// negative factors are considered invalid. +static int stbi__mul2sizes_valid(int a, int b) +{ + if (a < 0 || b < 0) return 0; + if (b == 0) return 1; // mul-by-0 is always safe + // portable way to check for no overflows in a*b + return a <= INT_MAX/b; +} + +// returns 1 if "a*b + add" has no negative terms/factors and doesn't overflow +static int stbi__mad2sizes_valid(int a, int b, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__addsizes_valid(a*b, add); +} + +// returns 1 if "a*b*c + add" has no negative terms/factors and doesn't overflow +static int stbi__mad3sizes_valid(int a, int b, int c, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__addsizes_valid(a*b*c, add); +} + +// returns 1 if "a*b*c*d + add" has no negative terms/factors and doesn't overflow +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +static int stbi__mad4sizes_valid(int a, int b, int c, int d, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__mul2sizes_valid(a*b*c, d) && stbi__addsizes_valid(a*b*c*d, add); +} +#endif + +// mallocs with size overflow checking +static void *stbi__malloc_mad2(int a, int b, int add) +{ + if (!stbi__mad2sizes_valid(a, b, add)) return NULL; + return stbi__malloc(a*b + add); +} + +static void *stbi__malloc_mad3(int a, int b, int c, int add) +{ + if (!stbi__mad3sizes_valid(a, b, c, add)) return NULL; + return stbi__malloc(a*b*c + add); +} + +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +static void *stbi__malloc_mad4(int a, int b, int c, int d, int add) +{ + if (!stbi__mad4sizes_valid(a, b, c, d, add)) return NULL; + return stbi__malloc(a*b*c*d + add); +} +#endif + +// stbi__err - error +// stbi__errpf - error returning pointer to float +// stbi__errpuc - error returning pointer to unsigned char + +#ifdef STBI_NO_FAILURE_STRINGS + #define stbi__err(x,y) 0 +#elif defined(STBI_FAILURE_USERMSG) + #define stbi__err(x,y) stbi__err(y) +#else + #define stbi__err(x,y) stbi__err(x) +#endif + +#define stbi__errpf(x,y) ((float *)(size_t) (stbi__err(x,y)?NULL:NULL)) +#define stbi__errpuc(x,y) ((unsigned char *)(size_t) (stbi__err(x,y)?NULL:NULL)) + +STBIDEF void stbi_image_free(void *retval_from_stbi_load) +{ + STBI_FREE(retval_from_stbi_load); +} + +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp); +#endif + +#ifndef STBI_NO_HDR +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp); +#endif + +static int stbi__vertically_flip_on_load = 0; + +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip) +{ + stbi__vertically_flip_on_load = flag_true_if_should_flip; +} + +static void *stbi__load_main(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + memset(ri, 0, sizeof(*ri)); // make sure it's initialized if we add new fields + ri->bits_per_channel = 8; // default is 8 so most paths don't have to be changed + ri->channel_order = STBI_ORDER_RGB; // all current input & output are this, but this is here so we can add BGR order + ri->num_channels = 0; + + #ifndef STBI_NO_JPEG + if (stbi__jpeg_test(s)) return stbi__jpeg_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNG + if (stbi__png_test(s)) return stbi__png_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_BMP + if (stbi__bmp_test(s)) return stbi__bmp_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_GIF + if (stbi__gif_test(s)) return stbi__gif_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PSD + if (stbi__psd_test(s)) return stbi__psd_load(s,x,y,comp,req_comp, ri, bpc); + #endif + #ifndef STBI_NO_PIC + if (stbi__pic_test(s)) return stbi__pic_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNM + if (stbi__pnm_test(s)) return stbi__pnm_load(s,x,y,comp,req_comp, ri); + #endif + + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + float *hdr = stbi__hdr_load(s, x,y,comp,req_comp, ri); + return stbi__hdr_to_ldr(hdr, *x, *y, req_comp ? req_comp : *comp); + } + #endif + + #ifndef STBI_NO_TGA + // test tga last because it's a crappy test! + if (stbi__tga_test(s)) + return stbi__tga_load(s,x,y,comp,req_comp, ri); + #endif + + return stbi__errpuc("unknown image type", "Image not of any known type, or corrupt"); +} + +static stbi_uc *stbi__convert_16_to_8(stbi__uint16 *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi_uc *reduced; + + reduced = (stbi_uc *) stbi__malloc(img_len); + if (reduced == NULL) return stbi__errpuc("outofmem", "Out of memory"); + + for (i = 0; i < img_len; ++i) + reduced[i] = (stbi_uc)((orig[i] >> 8) & 0xFF); // top half of each byte is sufficient approx of 16->8 bit scaling + + STBI_FREE(orig); + return reduced; +} + +static stbi__uint16 *stbi__convert_8_to_16(stbi_uc *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi__uint16 *enlarged; + + enlarged = (stbi__uint16 *) stbi__malloc(img_len*2); + if (enlarged == NULL) return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + + for (i = 0; i < img_len; ++i) + enlarged[i] = (stbi__uint16)((orig[i] << 8) + orig[i]); // replicate to high and low byte, maps 0->0, 255->0xffff + + STBI_FREE(orig); + return enlarged; +} + +static void stbi__vertical_flip(void *image, int w, int h, int bytes_per_pixel) +{ + int row; + size_t bytes_per_row = (size_t)w * bytes_per_pixel; + stbi_uc temp[2048]; + stbi_uc *bytes = (stbi_uc *)image; + + for (row = 0; row < (h>>1); row++) { + stbi_uc *row0 = bytes + row*bytes_per_row; + stbi_uc *row1 = bytes + (h - row - 1)*bytes_per_row; + // swap row0 with row1 + size_t bytes_left = bytes_per_row; + while (bytes_left) { + size_t bytes_copy = (bytes_left < sizeof(temp)) ? bytes_left : sizeof(temp); + memcpy(temp, row0, bytes_copy); + memcpy(row0, row1, bytes_copy); + memcpy(row1, temp, bytes_copy); + row0 += bytes_copy; + row1 += bytes_copy; + bytes_left -= bytes_copy; + } + } +} + +static void stbi__vertical_flip_slices(void *image, int w, int h, int z, int bytes_per_pixel) +{ + int slice; + int slice_size = w * h * bytes_per_pixel; + + stbi_uc *bytes = (stbi_uc *)image; + for (slice = 0; slice < z; ++slice) { + stbi__vertical_flip(bytes, w, h, bytes_per_pixel); + bytes += slice_size; + } +} + +static unsigned char *stbi__load_and_postprocess_8bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 8); + + if (result == NULL) + return NULL; + + if (ri.bits_per_channel != 8) { + STBI_ASSERT(ri.bits_per_channel == 16); + result = stbi__convert_16_to_8((stbi__uint16 *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 8; + } + + // @TODO: move stbi__convert_format to here + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi_uc)); + } + + return (unsigned char *) result; +} + +static stbi__uint16 *stbi__load_and_postprocess_16bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 16); + + if (result == NULL) + return NULL; + + if (ri.bits_per_channel != 16) { + STBI_ASSERT(ri.bits_per_channel == 8); + result = stbi__convert_8_to_16((stbi_uc *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 16; + } + + // @TODO: move stbi__convert_format16 to here + // @TODO: special case RGB-to-Y (and RGBA-to-YA) for 8-bit-to-16-bit case to keep more precision + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi__uint16)); + } + + return (stbi__uint16 *) result; +} + +#if !defined(STBI_NO_HDR) || !defined(STBI_NO_LINEAR) +static void stbi__float_postprocess(float *result, int *x, int *y, int *comp, int req_comp) +{ + if (stbi__vertically_flip_on_load && result != NULL) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(float)); + } +} +#endif + +#ifndef STBI_NO_STDIO + +static FILE *stbi__fopen(char const *filename, char const *mode) +{ + FILE *f; +#if defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != fopen_s(&f, filename, mode)) + f=0; +#else + f = fopen(filename, mode); +#endif + return f; +} + + +STBIDEF stbi_uc *stbi_load(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + unsigned char *result; + if (!f) return stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} + +STBIDEF stbi_uc *stbi_load_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); + if (result) { + // need to 'unget' all the characters in the IO buffer + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} + +STBIDEF stbi__uint16 *stbi_load_from_file_16(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__uint16 *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_16bit(&s,x,y,comp,req_comp); + if (result) { + // need to 'unget' all the characters in the IO buffer + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} + +STBIDEF stbi_us *stbi_load_16(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + stbi__uint16 *result; + if (!f) return (stbi_us *) stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file_16(f,x,y,comp,req_comp); + fclose(f); + return result; +} + + +#endif //!STBI_NO_STDIO + +STBIDEF stbi_us *stbi_load_16_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} + +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *)clbk, user); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} + +STBIDEF stbi_uc *stbi_load_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} + +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} + +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_mem(&s,buffer,len); + + result = (unsigned char*) stbi__load_gif_main(&s, delays, x, y, z, comp, req_comp); + if (stbi__vertically_flip_on_load) { + stbi__vertical_flip_slices( result, *x, *y, *z, *comp ); + } + + return result; +} +#endif + +#ifndef STBI_NO_LINEAR +static float *stbi__loadf_main(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *data; + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + stbi__result_info ri; + float *hdr_data = stbi__hdr_load(s,x,y,comp,req_comp, &ri); + if (hdr_data) + stbi__float_postprocess(hdr_data,x,y,comp,req_comp); + return hdr_data; + } + #endif + data = stbi__load_and_postprocess_8bit(s, x, y, comp, req_comp); + if (data) + return stbi__ldr_to_hdr(data, *x, *y, req_comp ? req_comp : *comp); + return stbi__errpf("unknown image type", "Image not of any known type, or corrupt"); +} + +STBIDEF float *stbi_loadf_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} + +STBIDEF float *stbi_loadf_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} + +#ifndef STBI_NO_STDIO +STBIDEF float *stbi_loadf(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + float *result; + FILE *f = stbi__fopen(filename, "rb"); + if (!f) return stbi__errpf("can't fopen", "Unable to open file"); + result = stbi_loadf_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} + +STBIDEF float *stbi_loadf_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_file(&s,f); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} +#endif // !STBI_NO_STDIO + +#endif // !STBI_NO_LINEAR + +// these is-hdr-or-not is defined independent of whether STBI_NO_LINEAR is +// defined, for API simplicity; if STBI_NO_LINEAR is defined, it always +// reports false! + +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(buffer); + STBI_NOTUSED(len); + return 0; + #endif +} + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result=0; + if (f) { + result = stbi_is_hdr_from_file(f); + fclose(f); + } + return result; +} + +STBIDEF int stbi_is_hdr_from_file(FILE *f) +{ + #ifndef STBI_NO_HDR + long pos = ftell(f); + int res; + stbi__context s; + stbi__start_file(&s,f); + res = stbi__hdr_test(&s); + fseek(f, pos, SEEK_SET); + return res; + #else + STBI_NOTUSED(f); + return 0; + #endif +} +#endif // !STBI_NO_STDIO + +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(clbk); + STBI_NOTUSED(user); + return 0; + #endif +} + +#ifndef STBI_NO_LINEAR +static float stbi__l2h_gamma=2.2f, stbi__l2h_scale=1.0f; + +STBIDEF void stbi_ldr_to_hdr_gamma(float gamma) { stbi__l2h_gamma = gamma; } +STBIDEF void stbi_ldr_to_hdr_scale(float scale) { stbi__l2h_scale = scale; } +#endif + +static float stbi__h2l_gamma_i=1.0f/2.2f, stbi__h2l_scale_i=1.0f; + +STBIDEF void stbi_hdr_to_ldr_gamma(float gamma) { stbi__h2l_gamma_i = 1/gamma; } +STBIDEF void stbi_hdr_to_ldr_scale(float scale) { stbi__h2l_scale_i = 1/scale; } + + +////////////////////////////////////////////////////////////////////////////// +// +// Common code used by all image loaders +// + +enum +{ + STBI__SCAN_load=0, + STBI__SCAN_type, + STBI__SCAN_header +}; + +static void stbi__refill_buffer(stbi__context *s) +{ + int n = (s->io.read)(s->io_user_data,(char*)s->buffer_start,s->buflen); + if (n == 0) { + // at end of file, treat same as if from memory, but need to handle case + // where s->img_buffer isn't pointing to safe memory, e.g. 0-byte file + s->read_from_callbacks = 0; + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start+1; + *s->img_buffer = 0; + } else { + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start + n; + } +} + +stbi_inline static stbi_uc stbi__get8(stbi__context *s) +{ + if (s->img_buffer < s->img_buffer_end) + return *s->img_buffer++; + if (s->read_from_callbacks) { + stbi__refill_buffer(s); + return *s->img_buffer++; + } + return 0; +} + +stbi_inline static int stbi__at_eof(stbi__context *s) +{ + if (s->io.read) { + if (!(s->io.eof)(s->io_user_data)) return 0; + // if feof() is true, check if buffer = end + // special case: we've only got the special 0 character at the end + if (s->read_from_callbacks == 0) return 1; + } + + return s->img_buffer >= s->img_buffer_end; +} + +static void stbi__skip(stbi__context *s, int n) +{ + if (n < 0) { + s->img_buffer = s->img_buffer_end; + return; + } + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + s->img_buffer = s->img_buffer_end; + (s->io.skip)(s->io_user_data, n - blen); + return; + } + } + s->img_buffer += n; +} + +static int stbi__getn(stbi__context *s, stbi_uc *buffer, int n) +{ + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + int res, count; + + memcpy(buffer, s->img_buffer, blen); + + count = (s->io.read)(s->io_user_data, (char*) buffer + blen, n - blen); + res = (count == (n-blen)); + s->img_buffer = s->img_buffer_end; + return res; + } + } + + if (s->img_buffer+n <= s->img_buffer_end) { + memcpy(buffer, s->img_buffer, n); + s->img_buffer += n; + return 1; + } else + return 0; +} + +static int stbi__get16be(stbi__context *s) +{ + int z = stbi__get8(s); + return (z << 8) + stbi__get8(s); +} + +static stbi__uint32 stbi__get32be(stbi__context *s) +{ + stbi__uint32 z = stbi__get16be(s); + return (z << 16) + stbi__get16be(s); +} + +#if defined(STBI_NO_BMP) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) +// nothing +#else +static int stbi__get16le(stbi__context *s) +{ + int z = stbi__get8(s); + return z + (stbi__get8(s) << 8); +} +#endif + +#ifndef STBI_NO_BMP +static stbi__uint32 stbi__get32le(stbi__context *s) +{ + stbi__uint32 z = stbi__get16le(s); + return z + (stbi__get16le(s) << 16); +} +#endif + +#define STBI__BYTECAST(x) ((stbi_uc) ((x) & 255)) // truncate int to byte without warnings + + +////////////////////////////////////////////////////////////////////////////// +// +// generic converter from built-in img_n to req_comp +// individual types do this automatically as much as possible (e.g. jpeg +// does all cases internally since it needs to colorspace convert anyway, +// and it never has alpha, so very few cases ). png can automatically +// interleave an alpha=255 channel, but falls back to this for other cases +// +// assume data buffer is malloced, so malloc a new one and free that one +// only failure mode is malloc failing + +static stbi_uc stbi__compute_y(int r, int g, int b) +{ + return (stbi_uc) (((r*77) + (g*150) + (29*b)) >> 8); +} + +static unsigned char *stbi__convert_format(unsigned char *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + unsigned char *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (unsigned char *) stbi__malloc_mad3(req_comp, x, y, 0); + if (good == NULL) { + STBI_FREE(data); + return stbi__errpuc("outofmem", "Out of memory"); + } + + for (j=0; j < (int) y; ++j) { + unsigned char *src = data + j * x * img_n ; + unsigned char *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + // convert source image with img_n components to one with req_comp components; + // avoid switch per pixel, so use switch per scanline and massive macros + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0], dest[1]=255; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=255; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2],dest[3]=255; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]), dest[1] = 255; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]), dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} + +static stbi__uint16 stbi__compute_y_16(int r, int g, int b) +{ + return (stbi__uint16) (((r*77) + (g*150) + (29*b)) >> 8); +} + +static stbi__uint16 *stbi__convert_format16(stbi__uint16 *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + stbi__uint16 *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (stbi__uint16 *) stbi__malloc(req_comp * x * y * 2); + if (good == NULL) { + STBI_FREE(data); + return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + } + + for (j=0; j < (int) y; ++j) { + stbi__uint16 *src = data + j * x * img_n ; + stbi__uint16 *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + // convert source image with img_n components to one with req_comp components; + // avoid switch per pixel, so use switch per scanline and massive macros + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0], dest[1]=0xffff; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=0xffff; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2],dest[3]=0xffff; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]), dest[1] = 0xffff; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]), dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} + +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp) +{ + int i,k,n; + float *output; + if (!data) return NULL; + output = (float *) stbi__malloc_mad4(x, y, comp, sizeof(float), 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpf("outofmem", "Out of memory"); } + // compute number of non-alpha components + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + output[i*comp + k] = (float) (pow(data[i*comp+k]/255.0f, stbi__l2h_gamma) * stbi__l2h_scale); + } + if (k < comp) output[i*comp + k] = data[i*comp+k]/255.0f; + } + STBI_FREE(data); + return output; +} +#endif + +#ifndef STBI_NO_HDR +#define stbi__float2int(x) ((int) (x)) +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp) +{ + int i,k,n; + stbi_uc *output; + if (!data) return NULL; + output = (stbi_uc *) stbi__malloc_mad3(x, y, comp, 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpuc("outofmem", "Out of memory"); } + // compute number of non-alpha components + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + float z = (float) pow(data[i*comp+k]*stbi__h2l_scale_i, stbi__h2l_gamma_i) * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + if (k < comp) { + float z = data[i*comp+k] * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + } + STBI_FREE(data); + return output; +} +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// "baseline" JPEG/JFIF decoder +// +// simple implementation +// - doesn't support delayed output of y-dimension +// - simple interface (only one output format: 8-bit interleaved RGB) +// - doesn't try to recover corrupt jpegs +// - doesn't allow partial loading, loading multiple at once +// - still fast on x86 (copying globals into locals doesn't help x86) +// - allocates lots of intermediate memory (full size of all components) +// - non-interleaved case requires this anyway +// - allows good upsampling (see next) +// high-quality +// - upsampled channels are bilinearly interpolated, even across blocks +// - quality integer IDCT derived from IJG's 'slow' +// performance +// - fast huffman; reasonable integer IDCT +// - some SIMD kernels for common paths on targets with SSE2/NEON +// - uses a lot of intermediate memory, could cache poorly + +#ifndef STBI_NO_JPEG + +// huffman decoding acceleration +#define FAST_BITS 9 // larger handles more cases; smaller stomps less cache + +typedef struct +{ + stbi_uc fast[1 << FAST_BITS]; + // weirdly, repacking this into AoS is a 10% speed loss, instead of a win + stbi__uint16 code[256]; + stbi_uc values[256]; + stbi_uc size[257]; + unsigned int maxcode[18]; + int delta[17]; // old 'firstsymbol' - old 'firstcode' +} stbi__huffman; + +typedef struct +{ + stbi__context *s; + stbi__huffman huff_dc[4]; + stbi__huffman huff_ac[4]; + stbi__uint16 dequant[4][64]; + stbi__int16 fast_ac[4][1 << FAST_BITS]; + +// sizes for components, interleaved MCUs + int img_h_max, img_v_max; + int img_mcu_x, img_mcu_y; + int img_mcu_w, img_mcu_h; + +// definition of jpeg image component + struct + { + int id; + int h,v; + int tq; + int hd,ha; + int dc_pred; + + int x,y,w2,h2; + stbi_uc *data; + void *raw_data, *raw_coeff; + stbi_uc *linebuf; + short *coeff; // progressive only + int coeff_w, coeff_h; // number of 8x8 coefficient blocks + } img_comp[4]; + + stbi__uint32 code_buffer; // jpeg entropy-coded buffer + int code_bits; // number of valid bits + unsigned char marker; // marker seen while filling entropy buffer + int nomore; // flag if we saw a marker so must stop + + int progressive; + int spec_start; + int spec_end; + int succ_high; + int succ_low; + int eob_run; + int jfif; + int app14_color_transform; // Adobe APP14 tag + int rgb; + + int scan_n, order[4]; + int restart_interval, todo; + +// kernels + void (*idct_block_kernel)(stbi_uc *out, int out_stride, short data[64]); + void (*YCbCr_to_RGB_kernel)(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step); + stbi_uc *(*resample_row_hv_2_kernel)(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs); +} stbi__jpeg; + +static int stbi__build_huffman(stbi__huffman *h, int *count) +{ + int i,j,k=0; + unsigned int code; + // build size list for each symbol (from JPEG spec) + for (i=0; i < 16; ++i) + for (j=0; j < count[i]; ++j) + h->size[k++] = (stbi_uc) (i+1); + h->size[k] = 0; + + // compute actual symbols (from jpeg spec) + code = 0; + k = 0; + for(j=1; j <= 16; ++j) { + // compute delta to add to code to compute symbol id + h->delta[j] = k - code; + if (h->size[k] == j) { + while (h->size[k] == j) + h->code[k++] = (stbi__uint16) (code++); + if (code-1 >= (1u << j)) return stbi__err("bad code lengths","Corrupt JPEG"); + } + // compute largest code + 1 for this size, preshifted as needed later + h->maxcode[j] = code << (16-j); + code <<= 1; + } + h->maxcode[j] = 0xffffffff; + + // build non-spec acceleration table; 255 is flag for not-accelerated + memset(h->fast, 255, 1 << FAST_BITS); + for (i=0; i < k; ++i) { + int s = h->size[i]; + if (s <= FAST_BITS) { + int c = h->code[i] << (FAST_BITS-s); + int m = 1 << (FAST_BITS-s); + for (j=0; j < m; ++j) { + h->fast[c+j] = (stbi_uc) i; + } + } + } + return 1; +} + +// build a table that decodes both magnitude and value of small ACs in +// one go. +static void stbi__build_fast_ac(stbi__int16 *fast_ac, stbi__huffman *h) +{ + int i; + for (i=0; i < (1 << FAST_BITS); ++i) { + stbi_uc fast = h->fast[i]; + fast_ac[i] = 0; + if (fast < 255) { + int rs = h->values[fast]; + int run = (rs >> 4) & 15; + int magbits = rs & 15; + int len = h->size[fast]; + + if (magbits && len + magbits <= FAST_BITS) { + // magnitude code followed by receive_extend code + int k = ((i << len) & ((1 << FAST_BITS) - 1)) >> (FAST_BITS - magbits); + int m = 1 << (magbits - 1); + if (k < m) k += (~0U << magbits) + 1; + // if the result is small enough, we can fit it in fast_ac table + if (k >= -128 && k <= 127) + fast_ac[i] = (stbi__int16) ((k * 256) + (run * 16) + (len + magbits)); + } + } + } +} + +static void stbi__grow_buffer_unsafe(stbi__jpeg *j) +{ + do { + unsigned int b = j->nomore ? 0 : stbi__get8(j->s); + if (b == 0xff) { + int c = stbi__get8(j->s); + while (c == 0xff) c = stbi__get8(j->s); // consume fill bytes + if (c != 0) { + j->marker = (unsigned char) c; + j->nomore = 1; + return; + } + } + j->code_buffer |= b << (24 - j->code_bits); + j->code_bits += 8; + } while (j->code_bits <= 24); +} + +// (1 << n) - 1 +static const stbi__uint32 stbi__bmask[17]={0,1,3,7,15,31,63,127,255,511,1023,2047,4095,8191,16383,32767,65535}; + +// decode a jpeg huffman value from the bitstream +stbi_inline static int stbi__jpeg_huff_decode(stbi__jpeg *j, stbi__huffman *h) +{ + unsigned int temp; + int c,k; + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + // look at the top FAST_BITS and determine what symbol ID it is, + // if the code is <= FAST_BITS + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + k = h->fast[c]; + if (k < 255) { + int s = h->size[k]; + if (s > j->code_bits) + return -1; + j->code_buffer <<= s; + j->code_bits -= s; + return h->values[k]; + } + + // naive test is to shift the code_buffer down so k bits are + // valid, then test against maxcode. To speed this up, we've + // preshifted maxcode left so that it has (16-k) 0s at the + // end; in other words, regardless of the number of bits, it + // wants to be compared against something shifted to have 16; + // that way we don't need to shift inside the loop. + temp = j->code_buffer >> 16; + for (k=FAST_BITS+1 ; ; ++k) + if (temp < h->maxcode[k]) + break; + if (k == 17) { + // error! code not found + j->code_bits -= 16; + return -1; + } + + if (k > j->code_bits) + return -1; + + // convert the huffman code to the symbol id + c = ((j->code_buffer >> (32 - k)) & stbi__bmask[k]) + h->delta[k]; + STBI_ASSERT((((j->code_buffer) >> (32 - h->size[c])) & stbi__bmask[h->size[c]]) == h->code[c]); + + // convert the id to a symbol + j->code_bits -= k; + j->code_buffer <<= k; + return h->values[c]; +} + +// bias[n] = (-1<code_bits < n) stbi__grow_buffer_unsafe(j); + + sgn = (stbi__int32)j->code_buffer >> 31; // sign bit is always in MSB + k = stbi_lrot(j->code_buffer, n); + STBI_ASSERT(n >= 0 && n < (int) (sizeof(stbi__bmask)/sizeof(*stbi__bmask))); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k + (stbi__jbias[n] & ~sgn); +} + +// get some unsigned bits +stbi_inline static int stbi__jpeg_get_bits(stbi__jpeg *j, int n) +{ + unsigned int k; + if (j->code_bits < n) stbi__grow_buffer_unsafe(j); + k = stbi_lrot(j->code_buffer, n); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k; +} + +stbi_inline static int stbi__jpeg_get_bit(stbi__jpeg *j) +{ + unsigned int k; + if (j->code_bits < 1) stbi__grow_buffer_unsafe(j); + k = j->code_buffer; + j->code_buffer <<= 1; + --j->code_bits; + return k & 0x80000000; +} + +// given a value that's at position X in the zigzag stream, +// where does it appear in the 8x8 matrix coded as row-major? +static const stbi_uc stbi__jpeg_dezigzag[64+15] = +{ + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 48, 41, 34, + 27, 20, 13, 6, 7, 14, 21, 28, + 35, 42, 49, 56, 57, 50, 43, 36, + 29, 22, 15, 23, 30, 37, 44, 51, + 58, 59, 52, 45, 38, 31, 39, 46, + 53, 60, 61, 54, 47, 55, 62, 63, + // let corrupt input sample past end + 63, 63, 63, 63, 63, 63, 63, 63, + 63, 63, 63, 63, 63, 63, 63 +}; + +// decode one 64-entry block-- +static int stbi__jpeg_decode_block(stbi__jpeg *j, short data[64], stbi__huffman *hdc, stbi__huffman *hac, stbi__int16 *fac, int b, stbi__uint16 *dequant) +{ + int diff,dc,k; + int t; + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + t = stbi__jpeg_huff_decode(j, hdc); + if (t < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + + // 0 all the ac values now so we can do it 32-bits at a time + memset(data,0,64*sizeof(data[0])); + + diff = t ? stbi__extend_receive(j, t) : 0; + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + data[0] = (short) (dc * dequant[0]); + + // decode AC components, see JPEG spec + k = 1; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { // fast-AC path + k += (r >> 4) & 15; // run + s = r & 15; // combined length + j->code_buffer <<= s; + j->code_bits -= s; + // decode into unzigzag'd location + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) * dequant[zig]); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (rs != 0xf0) break; // end block + k += 16; + } else { + k += r; + // decode into unzigzag'd location + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) * dequant[zig]); + } + } + } while (k < 64); + return 1; +} + +static int stbi__jpeg_decode_block_prog_dc(stbi__jpeg *j, short data[64], stbi__huffman *hdc, int b) +{ + int diff,dc; + int t; + if (j->spec_end != 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + if (j->succ_high == 0) { + // first scan for DC coefficient, must be first + memset(data,0,64*sizeof(data[0])); // 0 all the ac values now + t = stbi__jpeg_huff_decode(j, hdc); + diff = t ? stbi__extend_receive(j, t) : 0; + + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + data[0] = (short) (dc << j->succ_low); + } else { + // refinement scan for DC coefficient + if (stbi__jpeg_get_bit(j)) + data[0] += (short) (1 << j->succ_low); + } + return 1; +} + +// @OPTIMIZE: store non-zigzagged during the decode passes, +// and only de-zigzag when dequantizing +static int stbi__jpeg_decode_block_prog_ac(stbi__jpeg *j, short data[64], stbi__huffman *hac, stbi__int16 *fac) +{ + int k; + if (j->spec_start == 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + + if (j->succ_high == 0) { + int shift = j->succ_low; + + if (j->eob_run) { + --j->eob_run; + return 1; + } + + k = j->spec_start; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { // fast-AC path + k += (r >> 4) & 15; // run + s = r & 15; // combined length + j->code_buffer <<= s; + j->code_bits -= s; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) << shift); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r); + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + --j->eob_run; + break; + } + k += 16; + } else { + k += r; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) << shift); + } + } + } while (k <= j->spec_end); + } else { + // refinement scan for these AC coefficients + + short bit = (short) (1 << j->succ_low); + + if (j->eob_run) { + --j->eob_run; + for (k = j->spec_start; k <= j->spec_end; ++k) { + short *p = &data[stbi__jpeg_dezigzag[k]]; + if (*p != 0) + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } + } else { + k = j->spec_start; + do { + int r,s; + int rs = stbi__jpeg_huff_decode(j, hac); // @OPTIMIZE see if we can use the fast path here, advance-by-r is so slow, eh + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r) - 1; + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + r = 64; // force end of block + } else { + // r=15 s=0 should write 16 0s, so we just do + // a run of 15 0s and then write s (which is 0), + // so we don't have to do anything special here + } + } else { + if (s != 1) return stbi__err("bad huffman code", "Corrupt JPEG"); + // sign bit + if (stbi__jpeg_get_bit(j)) + s = bit; + else + s = -bit; + } + + // advance by r + while (k <= j->spec_end) { + short *p = &data[stbi__jpeg_dezigzag[k++]]; + if (*p != 0) { + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } else { + if (r == 0) { + *p = (short) s; + break; + } + --r; + } + } + } while (k <= j->spec_end); + } + } + return 1; +} + +// take a -128..127 value and stbi__clamp it and convert to 0..255 +stbi_inline static stbi_uc stbi__clamp(int x) +{ + // trick to use a single test to catch both cases + if ((unsigned int) x > 255) { + if (x < 0) return 0; + if (x > 255) return 255; + } + return (stbi_uc) x; +} + +#define stbi__f2f(x) ((int) (((x) * 4096 + 0.5))) +#define stbi__fsh(x) ((x) * 4096) + +// derived from jidctint -- DCT_ISLOW +#define STBI__IDCT_1D(s0,s1,s2,s3,s4,s5,s6,s7) \ + int t0,t1,t2,t3,p1,p2,p3,p4,p5,x0,x1,x2,x3; \ + p2 = s2; \ + p3 = s6; \ + p1 = (p2+p3) * stbi__f2f(0.5411961f); \ + t2 = p1 + p3*stbi__f2f(-1.847759065f); \ + t3 = p1 + p2*stbi__f2f( 0.765366865f); \ + p2 = s0; \ + p3 = s4; \ + t0 = stbi__fsh(p2+p3); \ + t1 = stbi__fsh(p2-p3); \ + x0 = t0+t3; \ + x3 = t0-t3; \ + x1 = t1+t2; \ + x2 = t1-t2; \ + t0 = s7; \ + t1 = s5; \ + t2 = s3; \ + t3 = s1; \ + p3 = t0+t2; \ + p4 = t1+t3; \ + p1 = t0+t3; \ + p2 = t1+t2; \ + p5 = (p3+p4)*stbi__f2f( 1.175875602f); \ + t0 = t0*stbi__f2f( 0.298631336f); \ + t1 = t1*stbi__f2f( 2.053119869f); \ + t2 = t2*stbi__f2f( 3.072711026f); \ + t3 = t3*stbi__f2f( 1.501321110f); \ + p1 = p5 + p1*stbi__f2f(-0.899976223f); \ + p2 = p5 + p2*stbi__f2f(-2.562915447f); \ + p3 = p3*stbi__f2f(-1.961570560f); \ + p4 = p4*stbi__f2f(-0.390180644f); \ + t3 += p1+p4; \ + t2 += p2+p3; \ + t1 += p2+p4; \ + t0 += p1+p3; + +static void stbi__idct_block(stbi_uc *out, int out_stride, short data[64]) +{ + int i,val[64],*v=val; + stbi_uc *o; + short *d = data; + + // columns + for (i=0; i < 8; ++i,++d, ++v) { + // if all zeroes, shortcut -- this avoids dequantizing 0s and IDCTing + if (d[ 8]==0 && d[16]==0 && d[24]==0 && d[32]==0 + && d[40]==0 && d[48]==0 && d[56]==0) { + // no shortcut 0 seconds + // (1|2|3|4|5|6|7)==0 0 seconds + // all separate -0.047 seconds + // 1 && 2|3 && 4|5 && 6|7: -0.047 seconds + int dcterm = d[0]*4; + v[0] = v[8] = v[16] = v[24] = v[32] = v[40] = v[48] = v[56] = dcterm; + } else { + STBI__IDCT_1D(d[ 0],d[ 8],d[16],d[24],d[32],d[40],d[48],d[56]) + // constants scaled things up by 1<<12; let's bring them back + // down, but keep 2 extra bits of precision + x0 += 512; x1 += 512; x2 += 512; x3 += 512; + v[ 0] = (x0+t3) >> 10; + v[56] = (x0-t3) >> 10; + v[ 8] = (x1+t2) >> 10; + v[48] = (x1-t2) >> 10; + v[16] = (x2+t1) >> 10; + v[40] = (x2-t1) >> 10; + v[24] = (x3+t0) >> 10; + v[32] = (x3-t0) >> 10; + } + } + + for (i=0, v=val, o=out; i < 8; ++i,v+=8,o+=out_stride) { + // no fast case since the first 1D IDCT spread components out + STBI__IDCT_1D(v[0],v[1],v[2],v[3],v[4],v[5],v[6],v[7]) + // constants scaled things up by 1<<12, plus we had 1<<2 from first + // loop, plus horizontal and vertical each scale by sqrt(8) so together + // we've got an extra 1<<3, so 1<<17 total we need to remove. + // so we want to round that, which means adding 0.5 * 1<<17, + // aka 65536. Also, we'll end up with -128 to 127 that we want + // to encode as 0..255 by adding 128, so we'll add that before the shift + x0 += 65536 + (128<<17); + x1 += 65536 + (128<<17); + x2 += 65536 + (128<<17); + x3 += 65536 + (128<<17); + // tried computing the shifts into temps, or'ing the temps to see + // if any were out of range, but that was slower + o[0] = stbi__clamp((x0+t3) >> 17); + o[7] = stbi__clamp((x0-t3) >> 17); + o[1] = stbi__clamp((x1+t2) >> 17); + o[6] = stbi__clamp((x1-t2) >> 17); + o[2] = stbi__clamp((x2+t1) >> 17); + o[5] = stbi__clamp((x2-t1) >> 17); + o[3] = stbi__clamp((x3+t0) >> 17); + o[4] = stbi__clamp((x3-t0) >> 17); + } +} + +#ifdef STBI_SSE2 +// sse2 integer IDCT. not the fastest possible implementation but it +// produces bit-identical results to the generic C version so it's +// fully "transparent". +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + // This is constructed to match our regular (generic) integer IDCT exactly. + __m128i row0, row1, row2, row3, row4, row5, row6, row7; + __m128i tmp; + + // dot product constant: even elems=x, odd elems=y + #define dct_const(x,y) _mm_setr_epi16((x),(y),(x),(y),(x),(y),(x),(y)) + + // out(0) = c0[even]*x + c0[odd]*y (c0, x, y 16-bit, out 32-bit) + // out(1) = c1[even]*x + c1[odd]*y + #define dct_rot(out0,out1, x,y,c0,c1) \ + __m128i c0##lo = _mm_unpacklo_epi16((x),(y)); \ + __m128i c0##hi = _mm_unpackhi_epi16((x),(y)); \ + __m128i out0##_l = _mm_madd_epi16(c0##lo, c0); \ + __m128i out0##_h = _mm_madd_epi16(c0##hi, c0); \ + __m128i out1##_l = _mm_madd_epi16(c0##lo, c1); \ + __m128i out1##_h = _mm_madd_epi16(c0##hi, c1) + + // out = in << 12 (in 16-bit, out 32-bit) + #define dct_widen(out, in) \ + __m128i out##_l = _mm_srai_epi32(_mm_unpacklo_epi16(_mm_setzero_si128(), (in)), 4); \ + __m128i out##_h = _mm_srai_epi32(_mm_unpackhi_epi16(_mm_setzero_si128(), (in)), 4) + + // wide add + #define dct_wadd(out, a, b) \ + __m128i out##_l = _mm_add_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_add_epi32(a##_h, b##_h) + + // wide sub + #define dct_wsub(out, a, b) \ + __m128i out##_l = _mm_sub_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_sub_epi32(a##_h, b##_h) + + // butterfly a/b, add bias, then shift by "s" and pack + #define dct_bfly32o(out0, out1, a,b,bias,s) \ + { \ + __m128i abiased_l = _mm_add_epi32(a##_l, bias); \ + __m128i abiased_h = _mm_add_epi32(a##_h, bias); \ + dct_wadd(sum, abiased, b); \ + dct_wsub(dif, abiased, b); \ + out0 = _mm_packs_epi32(_mm_srai_epi32(sum_l, s), _mm_srai_epi32(sum_h, s)); \ + out1 = _mm_packs_epi32(_mm_srai_epi32(dif_l, s), _mm_srai_epi32(dif_h, s)); \ + } + + // 8-bit interleave step (for transposes) + #define dct_interleave8(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi8(a, b); \ + b = _mm_unpackhi_epi8(tmp, b) + + // 16-bit interleave step (for transposes) + #define dct_interleave16(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi16(a, b); \ + b = _mm_unpackhi_epi16(tmp, b) + + #define dct_pass(bias,shift) \ + { \ + /* even part */ \ + dct_rot(t2e,t3e, row2,row6, rot0_0,rot0_1); \ + __m128i sum04 = _mm_add_epi16(row0, row4); \ + __m128i dif04 = _mm_sub_epi16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + /* odd part */ \ + dct_rot(y0o,y2o, row7,row3, rot2_0,rot2_1); \ + dct_rot(y1o,y3o, row5,row1, rot3_0,rot3_1); \ + __m128i sum17 = _mm_add_epi16(row1, row7); \ + __m128i sum35 = _mm_add_epi16(row3, row5); \ + dct_rot(y4o,y5o, sum17,sum35, rot1_0,rot1_1); \ + dct_wadd(x4, y0o, y4o); \ + dct_wadd(x5, y1o, y5o); \ + dct_wadd(x6, y2o, y5o); \ + dct_wadd(x7, y3o, y4o); \ + dct_bfly32o(row0,row7, x0,x7,bias,shift); \ + dct_bfly32o(row1,row6, x1,x6,bias,shift); \ + dct_bfly32o(row2,row5, x2,x5,bias,shift); \ + dct_bfly32o(row3,row4, x3,x4,bias,shift); \ + } + + __m128i rot0_0 = dct_const(stbi__f2f(0.5411961f), stbi__f2f(0.5411961f) + stbi__f2f(-1.847759065f)); + __m128i rot0_1 = dct_const(stbi__f2f(0.5411961f) + stbi__f2f( 0.765366865f), stbi__f2f(0.5411961f)); + __m128i rot1_0 = dct_const(stbi__f2f(1.175875602f) + stbi__f2f(-0.899976223f), stbi__f2f(1.175875602f)); + __m128i rot1_1 = dct_const(stbi__f2f(1.175875602f), stbi__f2f(1.175875602f) + stbi__f2f(-2.562915447f)); + __m128i rot2_0 = dct_const(stbi__f2f(-1.961570560f) + stbi__f2f( 0.298631336f), stbi__f2f(-1.961570560f)); + __m128i rot2_1 = dct_const(stbi__f2f(-1.961570560f), stbi__f2f(-1.961570560f) + stbi__f2f( 3.072711026f)); + __m128i rot3_0 = dct_const(stbi__f2f(-0.390180644f) + stbi__f2f( 2.053119869f), stbi__f2f(-0.390180644f)); + __m128i rot3_1 = dct_const(stbi__f2f(-0.390180644f), stbi__f2f(-0.390180644f) + stbi__f2f( 1.501321110f)); + + // rounding biases in column/row passes, see stbi__idct_block for explanation. + __m128i bias_0 = _mm_set1_epi32(512); + __m128i bias_1 = _mm_set1_epi32(65536 + (128<<17)); + + // load + row0 = _mm_load_si128((const __m128i *) (data + 0*8)); + row1 = _mm_load_si128((const __m128i *) (data + 1*8)); + row2 = _mm_load_si128((const __m128i *) (data + 2*8)); + row3 = _mm_load_si128((const __m128i *) (data + 3*8)); + row4 = _mm_load_si128((const __m128i *) (data + 4*8)); + row5 = _mm_load_si128((const __m128i *) (data + 5*8)); + row6 = _mm_load_si128((const __m128i *) (data + 6*8)); + row7 = _mm_load_si128((const __m128i *) (data + 7*8)); + + // column pass + dct_pass(bias_0, 10); + + { + // 16bit 8x8 transpose pass 1 + dct_interleave16(row0, row4); + dct_interleave16(row1, row5); + dct_interleave16(row2, row6); + dct_interleave16(row3, row7); + + // transpose pass 2 + dct_interleave16(row0, row2); + dct_interleave16(row1, row3); + dct_interleave16(row4, row6); + dct_interleave16(row5, row7); + + // transpose pass 3 + dct_interleave16(row0, row1); + dct_interleave16(row2, row3); + dct_interleave16(row4, row5); + dct_interleave16(row6, row7); + } + + // row pass + dct_pass(bias_1, 17); + + { + // pack + __m128i p0 = _mm_packus_epi16(row0, row1); // a0a1a2a3...a7b0b1b2b3...b7 + __m128i p1 = _mm_packus_epi16(row2, row3); + __m128i p2 = _mm_packus_epi16(row4, row5); + __m128i p3 = _mm_packus_epi16(row6, row7); + + // 8bit 8x8 transpose pass 1 + dct_interleave8(p0, p2); // a0e0a1e1... + dct_interleave8(p1, p3); // c0g0c1g1... + + // transpose pass 2 + dct_interleave8(p0, p1); // a0c0e0g0... + dct_interleave8(p2, p3); // b0d0f0h0... + + // transpose pass 3 + dct_interleave8(p0, p2); // a0b0c0d0... + dct_interleave8(p1, p3); // a4b4c4d4... + + // store + _mm_storel_epi64((__m128i *) out, p0); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p0, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p2); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p2, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p1); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p1, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p3); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p3, 0x4e)); + } + +#undef dct_const +#undef dct_rot +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_interleave8 +#undef dct_interleave16 +#undef dct_pass +} + +#endif // STBI_SSE2 + +#ifdef STBI_NEON + +// NEON integer IDCT. should produce bit-identical +// results to the generic C version. +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + int16x8_t row0, row1, row2, row3, row4, row5, row6, row7; + + int16x4_t rot0_0 = vdup_n_s16(stbi__f2f(0.5411961f)); + int16x4_t rot0_1 = vdup_n_s16(stbi__f2f(-1.847759065f)); + int16x4_t rot0_2 = vdup_n_s16(stbi__f2f( 0.765366865f)); + int16x4_t rot1_0 = vdup_n_s16(stbi__f2f( 1.175875602f)); + int16x4_t rot1_1 = vdup_n_s16(stbi__f2f(-0.899976223f)); + int16x4_t rot1_2 = vdup_n_s16(stbi__f2f(-2.562915447f)); + int16x4_t rot2_0 = vdup_n_s16(stbi__f2f(-1.961570560f)); + int16x4_t rot2_1 = vdup_n_s16(stbi__f2f(-0.390180644f)); + int16x4_t rot3_0 = vdup_n_s16(stbi__f2f( 0.298631336f)); + int16x4_t rot3_1 = vdup_n_s16(stbi__f2f( 2.053119869f)); + int16x4_t rot3_2 = vdup_n_s16(stbi__f2f( 3.072711026f)); + int16x4_t rot3_3 = vdup_n_s16(stbi__f2f( 1.501321110f)); + +#define dct_long_mul(out, inq, coeff) \ + int32x4_t out##_l = vmull_s16(vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmull_s16(vget_high_s16(inq), coeff) + +#define dct_long_mac(out, acc, inq, coeff) \ + int32x4_t out##_l = vmlal_s16(acc##_l, vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmlal_s16(acc##_h, vget_high_s16(inq), coeff) + +#define dct_widen(out, inq) \ + int32x4_t out##_l = vshll_n_s16(vget_low_s16(inq), 12); \ + int32x4_t out##_h = vshll_n_s16(vget_high_s16(inq), 12) + +// wide add +#define dct_wadd(out, a, b) \ + int32x4_t out##_l = vaddq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vaddq_s32(a##_h, b##_h) + +// wide sub +#define dct_wsub(out, a, b) \ + int32x4_t out##_l = vsubq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vsubq_s32(a##_h, b##_h) + +// butterfly a/b, then shift using "shiftop" by "s" and pack +#define dct_bfly32o(out0,out1, a,b,shiftop,s) \ + { \ + dct_wadd(sum, a, b); \ + dct_wsub(dif, a, b); \ + out0 = vcombine_s16(shiftop(sum_l, s), shiftop(sum_h, s)); \ + out1 = vcombine_s16(shiftop(dif_l, s), shiftop(dif_h, s)); \ + } + +#define dct_pass(shiftop, shift) \ + { \ + /* even part */ \ + int16x8_t sum26 = vaddq_s16(row2, row6); \ + dct_long_mul(p1e, sum26, rot0_0); \ + dct_long_mac(t2e, p1e, row6, rot0_1); \ + dct_long_mac(t3e, p1e, row2, rot0_2); \ + int16x8_t sum04 = vaddq_s16(row0, row4); \ + int16x8_t dif04 = vsubq_s16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + /* odd part */ \ + int16x8_t sum15 = vaddq_s16(row1, row5); \ + int16x8_t sum17 = vaddq_s16(row1, row7); \ + int16x8_t sum35 = vaddq_s16(row3, row5); \ + int16x8_t sum37 = vaddq_s16(row3, row7); \ + int16x8_t sumodd = vaddq_s16(sum17, sum35); \ + dct_long_mul(p5o, sumodd, rot1_0); \ + dct_long_mac(p1o, p5o, sum17, rot1_1); \ + dct_long_mac(p2o, p5o, sum35, rot1_2); \ + dct_long_mul(p3o, sum37, rot2_0); \ + dct_long_mul(p4o, sum15, rot2_1); \ + dct_wadd(sump13o, p1o, p3o); \ + dct_wadd(sump24o, p2o, p4o); \ + dct_wadd(sump23o, p2o, p3o); \ + dct_wadd(sump14o, p1o, p4o); \ + dct_long_mac(x4, sump13o, row7, rot3_0); \ + dct_long_mac(x5, sump24o, row5, rot3_1); \ + dct_long_mac(x6, sump23o, row3, rot3_2); \ + dct_long_mac(x7, sump14o, row1, rot3_3); \ + dct_bfly32o(row0,row7, x0,x7,shiftop,shift); \ + dct_bfly32o(row1,row6, x1,x6,shiftop,shift); \ + dct_bfly32o(row2,row5, x2,x5,shiftop,shift); \ + dct_bfly32o(row3,row4, x3,x4,shiftop,shift); \ + } + + // load + row0 = vld1q_s16(data + 0*8); + row1 = vld1q_s16(data + 1*8); + row2 = vld1q_s16(data + 2*8); + row3 = vld1q_s16(data + 3*8); + row4 = vld1q_s16(data + 4*8); + row5 = vld1q_s16(data + 5*8); + row6 = vld1q_s16(data + 6*8); + row7 = vld1q_s16(data + 7*8); + + // add DC bias + row0 = vaddq_s16(row0, vsetq_lane_s16(1024, vdupq_n_s16(0), 0)); + + // column pass + dct_pass(vrshrn_n_s32, 10); + + // 16bit 8x8 transpose + { +// these three map to a single VTRN.16, VTRN.32, and VSWP, respectively. +// whether compilers actually get this is another story, sadly. +#define dct_trn16(x, y) { int16x8x2_t t = vtrnq_s16(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn32(x, y) { int32x4x2_t t = vtrnq_s32(vreinterpretq_s32_s16(x), vreinterpretq_s32_s16(y)); x = vreinterpretq_s16_s32(t.val[0]); y = vreinterpretq_s16_s32(t.val[1]); } +#define dct_trn64(x, y) { int16x8_t x0 = x; int16x8_t y0 = y; x = vcombine_s16(vget_low_s16(x0), vget_low_s16(y0)); y = vcombine_s16(vget_high_s16(x0), vget_high_s16(y0)); } + + // pass 1 + dct_trn16(row0, row1); // a0b0a2b2a4b4a6b6 + dct_trn16(row2, row3); + dct_trn16(row4, row5); + dct_trn16(row6, row7); + + // pass 2 + dct_trn32(row0, row2); // a0b0c0d0a4b4c4d4 + dct_trn32(row1, row3); + dct_trn32(row4, row6); + dct_trn32(row5, row7); + + // pass 3 + dct_trn64(row0, row4); // a0b0c0d0e0f0g0h0 + dct_trn64(row1, row5); + dct_trn64(row2, row6); + dct_trn64(row3, row7); + +#undef dct_trn16 +#undef dct_trn32 +#undef dct_trn64 + } + + // row pass + // vrshrn_n_s32 only supports shifts up to 16, we need + // 17. so do a non-rounding shift of 16 first then follow + // up with a rounding shift by 1. + dct_pass(vshrn_n_s32, 16); + + { + // pack and round + uint8x8_t p0 = vqrshrun_n_s16(row0, 1); + uint8x8_t p1 = vqrshrun_n_s16(row1, 1); + uint8x8_t p2 = vqrshrun_n_s16(row2, 1); + uint8x8_t p3 = vqrshrun_n_s16(row3, 1); + uint8x8_t p4 = vqrshrun_n_s16(row4, 1); + uint8x8_t p5 = vqrshrun_n_s16(row5, 1); + uint8x8_t p6 = vqrshrun_n_s16(row6, 1); + uint8x8_t p7 = vqrshrun_n_s16(row7, 1); + + // again, these can translate into one instruction, but often don't. +#define dct_trn8_8(x, y) { uint8x8x2_t t = vtrn_u8(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn8_16(x, y) { uint16x4x2_t t = vtrn_u16(vreinterpret_u16_u8(x), vreinterpret_u16_u8(y)); x = vreinterpret_u8_u16(t.val[0]); y = vreinterpret_u8_u16(t.val[1]); } +#define dct_trn8_32(x, y) { uint32x2x2_t t = vtrn_u32(vreinterpret_u32_u8(x), vreinterpret_u32_u8(y)); x = vreinterpret_u8_u32(t.val[0]); y = vreinterpret_u8_u32(t.val[1]); } + + // sadly can't use interleaved stores here since we only write + // 8 bytes to each scan line! + + // 8x8 8-bit transpose pass 1 + dct_trn8_8(p0, p1); + dct_trn8_8(p2, p3); + dct_trn8_8(p4, p5); + dct_trn8_8(p6, p7); + + // pass 2 + dct_trn8_16(p0, p2); + dct_trn8_16(p1, p3); + dct_trn8_16(p4, p6); + dct_trn8_16(p5, p7); + + // pass 3 + dct_trn8_32(p0, p4); + dct_trn8_32(p1, p5); + dct_trn8_32(p2, p6); + dct_trn8_32(p3, p7); + + // store + vst1_u8(out, p0); out += out_stride; + vst1_u8(out, p1); out += out_stride; + vst1_u8(out, p2); out += out_stride; + vst1_u8(out, p3); out += out_stride; + vst1_u8(out, p4); out += out_stride; + vst1_u8(out, p5); out += out_stride; + vst1_u8(out, p6); out += out_stride; + vst1_u8(out, p7); + +#undef dct_trn8_8 +#undef dct_trn8_16 +#undef dct_trn8_32 + } + +#undef dct_long_mul +#undef dct_long_mac +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_pass +} + +#endif // STBI_NEON + +#define STBI__MARKER_none 0xff +// if there's a pending marker from the entropy stream, return that +// otherwise, fetch from the stream and get a marker. if there's no +// marker, return 0xff, which is never a valid marker value +static stbi_uc stbi__get_marker(stbi__jpeg *j) +{ + stbi_uc x; + if (j->marker != STBI__MARKER_none) { x = j->marker; j->marker = STBI__MARKER_none; return x; } + x = stbi__get8(j->s); + if (x != 0xff) return STBI__MARKER_none; + while (x == 0xff) + x = stbi__get8(j->s); // consume repeated 0xff fill bytes + return x; +} + +// in each scan, we'll have scan_n components, and the order +// of the components is specified by order[] +#define STBI__RESTART(x) ((x) >= 0xd0 && (x) <= 0xd7) + +// after a restart interval, stbi__jpeg_reset the entropy decoder and +// the dc prediction +static void stbi__jpeg_reset(stbi__jpeg *j) +{ + j->code_bits = 0; + j->code_buffer = 0; + j->nomore = 0; + j->img_comp[0].dc_pred = j->img_comp[1].dc_pred = j->img_comp[2].dc_pred = j->img_comp[3].dc_pred = 0; + j->marker = STBI__MARKER_none; + j->todo = j->restart_interval ? j->restart_interval : 0x7fffffff; + j->eob_run = 0; + // no more than 1<<31 MCUs if no restart_interal? that's plenty safe, + // since we don't even allow 1<<30 pixels +} + +static int stbi__parse_entropy_coded_data(stbi__jpeg *z) +{ + stbi__jpeg_reset(z); + if (!z->progressive) { + if (z->scan_n == 1) { + int i,j; + STBI_SIMD_ALIGN(short, data[64]); + int n = z->order[0]; + // non-interleaved data, we just need to process one block at a time, + // in trivial scanline order + // number of blocks to do just depends on how many actual "pixels" this + // component has, independent of interleaved MCU blocking and such + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + // every data block is an MCU, so countdown the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + // if it's NOT a restart, then just bail, so we get corrupt data + // rather than no data + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { // interleaved + int i,j,k,x,y; + STBI_SIMD_ALIGN(short, data[64]); + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + // scan an interleaved mcu... process scan_n components in order + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + // scan out an mcu's worth of this component; that's just determined + // by the basic H and V specified for the component + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x)*8; + int y2 = (j*z->img_comp[n].v + y)*8; + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*y2+x2, z->img_comp[n].w2, data); + } + } + } + // after all interleaved components, that's an interleaved MCU, + // so now count down the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } else { + if (z->scan_n == 1) { + int i,j; + int n = z->order[0]; + // non-interleaved data, we just need to process one block at a time, + // in trivial scanline order + // number of blocks to do just depends on how many actual "pixels" this + // component has, independent of interleaved MCU blocking and such + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + if (z->spec_start == 0) { + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } else { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block_prog_ac(z, data, &z->huff_ac[ha], z->fast_ac[ha])) + return 0; + } + // every data block is an MCU, so countdown the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { // interleaved + int i,j,k,x,y; + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + // scan an interleaved mcu... process scan_n components in order + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + // scan out an mcu's worth of this component; that's just determined + // by the basic H and V specified for the component + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x); + int y2 = (j*z->img_comp[n].v + y); + short *data = z->img_comp[n].coeff + 64 * (x2 + y2 * z->img_comp[n].coeff_w); + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } + } + } + // after all interleaved components, that's an interleaved MCU, + // so now count down the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } +} + +static void stbi__jpeg_dequantize(short *data, stbi__uint16 *dequant) +{ + int i; + for (i=0; i < 64; ++i) + data[i] *= dequant[i]; +} + +static void stbi__jpeg_finish(stbi__jpeg *z) +{ + if (z->progressive) { + // dequantize and idct the data + int i,j,n; + for (n=0; n < z->s->img_n; ++n) { + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + stbi__jpeg_dequantize(data, z->dequant[z->img_comp[n].tq]); + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + } + } + } + } +} + +static int stbi__process_marker(stbi__jpeg *z, int m) +{ + int L; + switch (m) { + case STBI__MARKER_none: // no marker found + return stbi__err("expected marker","Corrupt JPEG"); + + case 0xDD: // DRI - specify restart interval + if (stbi__get16be(z->s) != 4) return stbi__err("bad DRI len","Corrupt JPEG"); + z->restart_interval = stbi__get16be(z->s); + return 1; + + case 0xDB: // DQT - define quantization table + L = stbi__get16be(z->s)-2; + while (L > 0) { + int q = stbi__get8(z->s); + int p = q >> 4, sixteen = (p != 0); + int t = q & 15,i; + if (p != 0 && p != 1) return stbi__err("bad DQT type","Corrupt JPEG"); + if (t > 3) return stbi__err("bad DQT table","Corrupt JPEG"); + + for (i=0; i < 64; ++i) + z->dequant[t][stbi__jpeg_dezigzag[i]] = (stbi__uint16)(sixteen ? stbi__get16be(z->s) : stbi__get8(z->s)); + L -= (sixteen ? 129 : 65); + } + return L==0; + + case 0xC4: // DHT - define huffman table + L = stbi__get16be(z->s)-2; + while (L > 0) { + stbi_uc *v; + int sizes[16],i,n=0; + int q = stbi__get8(z->s); + int tc = q >> 4; + int th = q & 15; + if (tc > 1 || th > 3) return stbi__err("bad DHT header","Corrupt JPEG"); + for (i=0; i < 16; ++i) { + sizes[i] = stbi__get8(z->s); + n += sizes[i]; + } + L -= 17; + if (tc == 0) { + if (!stbi__build_huffman(z->huff_dc+th, sizes)) return 0; + v = z->huff_dc[th].values; + } else { + if (!stbi__build_huffman(z->huff_ac+th, sizes)) return 0; + v = z->huff_ac[th].values; + } + for (i=0; i < n; ++i) + v[i] = stbi__get8(z->s); + if (tc != 0) + stbi__build_fast_ac(z->fast_ac[th], z->huff_ac + th); + L -= n; + } + return L==0; + } + + // check for comment block or APP blocks + if ((m >= 0xE0 && m <= 0xEF) || m == 0xFE) { + L = stbi__get16be(z->s); + if (L < 2) { + if (m == 0xFE) + return stbi__err("bad COM len","Corrupt JPEG"); + else + return stbi__err("bad APP len","Corrupt JPEG"); + } + L -= 2; + + if (m == 0xE0 && L >= 5) { // JFIF APP0 segment + static const unsigned char tag[5] = {'J','F','I','F','\0'}; + int ok = 1; + int i; + for (i=0; i < 5; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 5; + if (ok) + z->jfif = 1; + } else if (m == 0xEE && L >= 12) { // Adobe APP14 segment + static const unsigned char tag[6] = {'A','d','o','b','e','\0'}; + int ok = 1; + int i; + for (i=0; i < 6; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 6; + if (ok) { + stbi__get8(z->s); // version + stbi__get16be(z->s); // flags0 + stbi__get16be(z->s); // flags1 + z->app14_color_transform = stbi__get8(z->s); // color transform + L -= 6; + } + } + + stbi__skip(z->s, L); + return 1; + } + + return stbi__err("unknown marker","Corrupt JPEG"); +} + +// after we see SOS +static int stbi__process_scan_header(stbi__jpeg *z) +{ + int i; + int Ls = stbi__get16be(z->s); + z->scan_n = stbi__get8(z->s); + if (z->scan_n < 1 || z->scan_n > 4 || z->scan_n > (int) z->s->img_n) return stbi__err("bad SOS component count","Corrupt JPEG"); + if (Ls != 6+2*z->scan_n) return stbi__err("bad SOS len","Corrupt JPEG"); + for (i=0; i < z->scan_n; ++i) { + int id = stbi__get8(z->s), which; + int q = stbi__get8(z->s); + for (which = 0; which < z->s->img_n; ++which) + if (z->img_comp[which].id == id) + break; + if (which == z->s->img_n) return 0; // no match + z->img_comp[which].hd = q >> 4; if (z->img_comp[which].hd > 3) return stbi__err("bad DC huff","Corrupt JPEG"); + z->img_comp[which].ha = q & 15; if (z->img_comp[which].ha > 3) return stbi__err("bad AC huff","Corrupt JPEG"); + z->order[i] = which; + } + + { + int aa; + z->spec_start = stbi__get8(z->s); + z->spec_end = stbi__get8(z->s); // should be 63, but might be 0 + aa = stbi__get8(z->s); + z->succ_high = (aa >> 4); + z->succ_low = (aa & 15); + if (z->progressive) { + if (z->spec_start > 63 || z->spec_end > 63 || z->spec_start > z->spec_end || z->succ_high > 13 || z->succ_low > 13) + return stbi__err("bad SOS", "Corrupt JPEG"); + } else { + if (z->spec_start != 0) return stbi__err("bad SOS","Corrupt JPEG"); + if (z->succ_high != 0 || z->succ_low != 0) return stbi__err("bad SOS","Corrupt JPEG"); + z->spec_end = 63; + } + } + + return 1; +} + +static int stbi__free_jpeg_components(stbi__jpeg *z, int ncomp, int why) +{ + int i; + for (i=0; i < ncomp; ++i) { + if (z->img_comp[i].raw_data) { + STBI_FREE(z->img_comp[i].raw_data); + z->img_comp[i].raw_data = NULL; + z->img_comp[i].data = NULL; + } + if (z->img_comp[i].raw_coeff) { + STBI_FREE(z->img_comp[i].raw_coeff); + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].coeff = 0; + } + if (z->img_comp[i].linebuf) { + STBI_FREE(z->img_comp[i].linebuf); + z->img_comp[i].linebuf = NULL; + } + } + return why; +} + +static int stbi__process_frame_header(stbi__jpeg *z, int scan) +{ + stbi__context *s = z->s; + int Lf,p,i,q, h_max=1,v_max=1,c; + Lf = stbi__get16be(s); if (Lf < 11) return stbi__err("bad SOF len","Corrupt JPEG"); // JPEG + p = stbi__get8(s); if (p != 8) return stbi__err("only 8-bit","JPEG format not supported: 8-bit only"); // JPEG baseline + s->img_y = stbi__get16be(s); if (s->img_y == 0) return stbi__err("no header height", "JPEG format not supported: delayed height"); // Legal, but we don't handle it--but neither does IJG + s->img_x = stbi__get16be(s); if (s->img_x == 0) return stbi__err("0 width","Corrupt JPEG"); // JPEG requires + c = stbi__get8(s); + if (c != 3 && c != 1 && c != 4) return stbi__err("bad component count","Corrupt JPEG"); + s->img_n = c; + for (i=0; i < c; ++i) { + z->img_comp[i].data = NULL; + z->img_comp[i].linebuf = NULL; + } + + if (Lf != 8+3*s->img_n) return stbi__err("bad SOF len","Corrupt JPEG"); + + z->rgb = 0; + for (i=0; i < s->img_n; ++i) { + static const unsigned char rgb[3] = { 'R', 'G', 'B' }; + z->img_comp[i].id = stbi__get8(s); + if (s->img_n == 3 && z->img_comp[i].id == rgb[i]) + ++z->rgb; + q = stbi__get8(s); + z->img_comp[i].h = (q >> 4); if (!z->img_comp[i].h || z->img_comp[i].h > 4) return stbi__err("bad H","Corrupt JPEG"); + z->img_comp[i].v = q & 15; if (!z->img_comp[i].v || z->img_comp[i].v > 4) return stbi__err("bad V","Corrupt JPEG"); + z->img_comp[i].tq = stbi__get8(s); if (z->img_comp[i].tq > 3) return stbi__err("bad TQ","Corrupt JPEG"); + } + + if (scan != STBI__SCAN_load) return 1; + + if (!stbi__mad3sizes_valid(s->img_x, s->img_y, s->img_n, 0)) return stbi__err("too large", "Image too large to decode"); + + for (i=0; i < s->img_n; ++i) { + if (z->img_comp[i].h > h_max) h_max = z->img_comp[i].h; + if (z->img_comp[i].v > v_max) v_max = z->img_comp[i].v; + } + + // compute interleaved mcu info + z->img_h_max = h_max; + z->img_v_max = v_max; + z->img_mcu_w = h_max * 8; + z->img_mcu_h = v_max * 8; + // these sizes can't be more than 17 bits + z->img_mcu_x = (s->img_x + z->img_mcu_w-1) / z->img_mcu_w; + z->img_mcu_y = (s->img_y + z->img_mcu_h-1) / z->img_mcu_h; + + for (i=0; i < s->img_n; ++i) { + // number of effective pixels (e.g. for non-interleaved MCU) + z->img_comp[i].x = (s->img_x * z->img_comp[i].h + h_max-1) / h_max; + z->img_comp[i].y = (s->img_y * z->img_comp[i].v + v_max-1) / v_max; + // to simplify generation, we'll allocate enough memory to decode + // the bogus oversized data from using interleaved MCUs and their + // big blocks (e.g. a 16x16 iMCU on an image of width 33); we won't + // discard the extra data until colorspace conversion + // + // img_mcu_x, img_mcu_y: <=17 bits; comp[i].h and .v are <=4 (checked earlier) + // so these muls can't overflow with 32-bit ints (which we require) + z->img_comp[i].w2 = z->img_mcu_x * z->img_comp[i].h * 8; + z->img_comp[i].h2 = z->img_mcu_y * z->img_comp[i].v * 8; + z->img_comp[i].coeff = 0; + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].linebuf = NULL; + z->img_comp[i].raw_data = stbi__malloc_mad2(z->img_comp[i].w2, z->img_comp[i].h2, 15); + if (z->img_comp[i].raw_data == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + // align blocks for idct using mmx/sse + z->img_comp[i].data = (stbi_uc*) (((size_t) z->img_comp[i].raw_data + 15) & ~15); + if (z->progressive) { + // w2, h2 are multiples of 8 (see above) + z->img_comp[i].coeff_w = z->img_comp[i].w2 / 8; + z->img_comp[i].coeff_h = z->img_comp[i].h2 / 8; + z->img_comp[i].raw_coeff = stbi__malloc_mad3(z->img_comp[i].w2, z->img_comp[i].h2, sizeof(short), 15); + if (z->img_comp[i].raw_coeff == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + z->img_comp[i].coeff = (short*) (((size_t) z->img_comp[i].raw_coeff + 15) & ~15); + } + } + + return 1; +} + +// use comparisons since in some cases we handle more than one case (e.g. SOF) +#define stbi__DNL(x) ((x) == 0xdc) +#define stbi__SOI(x) ((x) == 0xd8) +#define stbi__EOI(x) ((x) == 0xd9) +#define stbi__SOF(x) ((x) == 0xc0 || (x) == 0xc1 || (x) == 0xc2) +#define stbi__SOS(x) ((x) == 0xda) + +#define stbi__SOF_progressive(x) ((x) == 0xc2) + +static int stbi__decode_jpeg_header(stbi__jpeg *z, int scan) +{ + int m; + z->jfif = 0; + z->app14_color_transform = -1; // valid values are 0,1,2 + z->marker = STBI__MARKER_none; // initialize cached marker to empty + m = stbi__get_marker(z); + if (!stbi__SOI(m)) return stbi__err("no SOI","Corrupt JPEG"); + if (scan == STBI__SCAN_type) return 1; + m = stbi__get_marker(z); + while (!stbi__SOF(m)) { + if (!stbi__process_marker(z,m)) return 0; + m = stbi__get_marker(z); + while (m == STBI__MARKER_none) { + // some files have extra padding after their blocks, so ok, we'll scan + if (stbi__at_eof(z->s)) return stbi__err("no SOF", "Corrupt JPEG"); + m = stbi__get_marker(z); + } + } + z->progressive = stbi__SOF_progressive(m); + if (!stbi__process_frame_header(z, scan)) return 0; + return 1; +} + +// decode image to YCbCr format +static int stbi__decode_jpeg_image(stbi__jpeg *j) +{ + int m; + for (m = 0; m < 4; m++) { + j->img_comp[m].raw_data = NULL; + j->img_comp[m].raw_coeff = NULL; + } + j->restart_interval = 0; + if (!stbi__decode_jpeg_header(j, STBI__SCAN_load)) return 0; + m = stbi__get_marker(j); + while (!stbi__EOI(m)) { + if (stbi__SOS(m)) { + if (!stbi__process_scan_header(j)) return 0; + if (!stbi__parse_entropy_coded_data(j)) return 0; + if (j->marker == STBI__MARKER_none ) { + // handle 0s at the end of image data from IP Kamera 9060 + while (!stbi__at_eof(j->s)) { + int x = stbi__get8(j->s); + if (x == 255) { + j->marker = stbi__get8(j->s); + break; + } + } + // if we reach eof without hitting a marker, stbi__get_marker() below will fail and we'll eventually return 0 + } + } else if (stbi__DNL(m)) { + int Ld = stbi__get16be(j->s); + stbi__uint32 NL = stbi__get16be(j->s); + if (Ld != 4) return stbi__err("bad DNL len", "Corrupt JPEG"); + if (NL != j->s->img_y) return stbi__err("bad DNL height", "Corrupt JPEG"); + } else { + if (!stbi__process_marker(j, m)) return 0; + } + m = stbi__get_marker(j); + } + if (j->progressive) + stbi__jpeg_finish(j); + return 1; +} + +// static jfif-centered resampling (across block boundaries) + +typedef stbi_uc *(*resample_row_func)(stbi_uc *out, stbi_uc *in0, stbi_uc *in1, + int w, int hs); + +#define stbi__div4(x) ((stbi_uc) ((x) >> 2)) + +static stbi_uc *resample_row_1(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + STBI_NOTUSED(out); + STBI_NOTUSED(in_far); + STBI_NOTUSED(w); + STBI_NOTUSED(hs); + return in_near; +} + +static stbi_uc* stbi__resample_row_v_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate two samples vertically for every one in input + int i; + STBI_NOTUSED(hs); + for (i=0; i < w; ++i) + out[i] = stbi__div4(3*in_near[i] + in_far[i] + 2); + return out; +} + +static stbi_uc* stbi__resample_row_h_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate two samples horizontally for every one in input + int i; + stbi_uc *input = in_near; + + if (w == 1) { + // if only one sample, can't do any interpolation + out[0] = out[1] = input[0]; + return out; + } + + out[0] = input[0]; + out[1] = stbi__div4(input[0]*3 + input[1] + 2); + for (i=1; i < w-1; ++i) { + int n = 3*input[i]+2; + out[i*2+0] = stbi__div4(n+input[i-1]); + out[i*2+1] = stbi__div4(n+input[i+1]); + } + out[i*2+0] = stbi__div4(input[w-2]*3 + input[w-1] + 2); + out[i*2+1] = input[w-1]; + + STBI_NOTUSED(in_far); + STBI_NOTUSED(hs); + + return out; +} + +#define stbi__div16(x) ((stbi_uc) ((x) >> 4)) + +static stbi_uc *stbi__resample_row_hv_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate 2x2 samples for every one in input + int i,t0,t1; + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + + t1 = 3*in_near[0] + in_far[0]; + out[0] = stbi__div4(t1+2); + for (i=1; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + + STBI_NOTUSED(hs); + + return out; +} + +#if defined(STBI_SSE2) || defined(STBI_NEON) +static stbi_uc *stbi__resample_row_hv_2_simd(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate 2x2 samples for every one in input + int i=0,t0,t1; + + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + + t1 = 3*in_near[0] + in_far[0]; + // process groups of 8 pixels for as long as we can. + // note we can't handle the last pixel in a row in this loop + // because we need to handle the filter boundary conditions. + for (; i < ((w-1) & ~7); i += 8) { +#if defined(STBI_SSE2) + // load and perform the vertical filtering pass + // this uses 3*x + y = 4*x + (y - x) + __m128i zero = _mm_setzero_si128(); + __m128i farb = _mm_loadl_epi64((__m128i *) (in_far + i)); + __m128i nearb = _mm_loadl_epi64((__m128i *) (in_near + i)); + __m128i farw = _mm_unpacklo_epi8(farb, zero); + __m128i nearw = _mm_unpacklo_epi8(nearb, zero); + __m128i diff = _mm_sub_epi16(farw, nearw); + __m128i nears = _mm_slli_epi16(nearw, 2); + __m128i curr = _mm_add_epi16(nears, diff); // current row + + // horizontal filter works the same based on shifted vers of current + // row. "prev" is current row shifted right by 1 pixel; we need to + // insert the previous pixel value (from t1). + // "next" is current row shifted left by 1 pixel, with first pixel + // of next block of 8 pixels added in. + __m128i prv0 = _mm_slli_si128(curr, 2); + __m128i nxt0 = _mm_srli_si128(curr, 2); + __m128i prev = _mm_insert_epi16(prv0, t1, 0); + __m128i next = _mm_insert_epi16(nxt0, 3*in_near[i+8] + in_far[i+8], 7); + + // horizontal filter, polyphase implementation since it's convenient: + // even pixels = 3*cur + prev = cur*4 + (prev - cur) + // odd pixels = 3*cur + next = cur*4 + (next - cur) + // note the shared term. + __m128i bias = _mm_set1_epi16(8); + __m128i curs = _mm_slli_epi16(curr, 2); + __m128i prvd = _mm_sub_epi16(prev, curr); + __m128i nxtd = _mm_sub_epi16(next, curr); + __m128i curb = _mm_add_epi16(curs, bias); + __m128i even = _mm_add_epi16(prvd, curb); + __m128i odd = _mm_add_epi16(nxtd, curb); + + // interleave even and odd pixels, then undo scaling. + __m128i int0 = _mm_unpacklo_epi16(even, odd); + __m128i int1 = _mm_unpackhi_epi16(even, odd); + __m128i de0 = _mm_srli_epi16(int0, 4); + __m128i de1 = _mm_srli_epi16(int1, 4); + + // pack and write output + __m128i outv = _mm_packus_epi16(de0, de1); + _mm_storeu_si128((__m128i *) (out + i*2), outv); +#elif defined(STBI_NEON) + // load and perform the vertical filtering pass + // this uses 3*x + y = 4*x + (y - x) + uint8x8_t farb = vld1_u8(in_far + i); + uint8x8_t nearb = vld1_u8(in_near + i); + int16x8_t diff = vreinterpretq_s16_u16(vsubl_u8(farb, nearb)); + int16x8_t nears = vreinterpretq_s16_u16(vshll_n_u8(nearb, 2)); + int16x8_t curr = vaddq_s16(nears, diff); // current row + + // horizontal filter works the same based on shifted vers of current + // row. "prev" is current row shifted right by 1 pixel; we need to + // insert the previous pixel value (from t1). + // "next" is current row shifted left by 1 pixel, with first pixel + // of next block of 8 pixels added in. + int16x8_t prv0 = vextq_s16(curr, curr, 7); + int16x8_t nxt0 = vextq_s16(curr, curr, 1); + int16x8_t prev = vsetq_lane_s16(t1, prv0, 0); + int16x8_t next = vsetq_lane_s16(3*in_near[i+8] + in_far[i+8], nxt0, 7); + + // horizontal filter, polyphase implementation since it's convenient: + // even pixels = 3*cur + prev = cur*4 + (prev - cur) + // odd pixels = 3*cur + next = cur*4 + (next - cur) + // note the shared term. + int16x8_t curs = vshlq_n_s16(curr, 2); + int16x8_t prvd = vsubq_s16(prev, curr); + int16x8_t nxtd = vsubq_s16(next, curr); + int16x8_t even = vaddq_s16(curs, prvd); + int16x8_t odd = vaddq_s16(curs, nxtd); + + // undo scaling and round, then store with even/odd phases interleaved + uint8x8x2_t o; + o.val[0] = vqrshrun_n_s16(even, 4); + o.val[1] = vqrshrun_n_s16(odd, 4); + vst2_u8(out + i*2, o); +#endif + + // "previous" value for next iter + t1 = 3*in_near[i+7] + in_far[i+7]; + } + + t0 = t1; + t1 = 3*in_near[i] + in_far[i]; + out[i*2] = stbi__div16(3*t1 + t0 + 8); + + for (++i; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + + STBI_NOTUSED(hs); + + return out; +} +#endif + +static stbi_uc *stbi__resample_row_generic(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // resample with nearest-neighbor + int i,j; + STBI_NOTUSED(in_far); + for (i=0; i < w; ++i) + for (j=0; j < hs; ++j) + out[i*hs+j] = in_near[i]; + return out; +} + +// this is a reduced-precision calculation of YCbCr-to-RGB introduced +// to make sure the code produces the same results in both SIMD and scalar +#define stbi__float2fixed(x) (((int) ((x) * 4096.0f + 0.5f)) << 8) +static void stbi__YCbCr_to_RGB_row(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step) +{ + int i; + for (i=0; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + (cr*-stbi__float2fixed(0.71414f)) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} + +#if defined(STBI_SSE2) || defined(STBI_NEON) +static void stbi__YCbCr_to_RGB_simd(stbi_uc *out, stbi_uc const *y, stbi_uc const *pcb, stbi_uc const *pcr, int count, int step) +{ + int i = 0; + +#ifdef STBI_SSE2 + // step == 3 is pretty ugly on the final interleave, and i'm not convinced + // it's useful in practice (you wouldn't use it for textures, for example). + // so just accelerate step == 4 case. + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + __m128i signflip = _mm_set1_epi8(-0x80); + __m128i cr_const0 = _mm_set1_epi16( (short) ( 1.40200f*4096.0f+0.5f)); + __m128i cr_const1 = _mm_set1_epi16( - (short) ( 0.71414f*4096.0f+0.5f)); + __m128i cb_const0 = _mm_set1_epi16( - (short) ( 0.34414f*4096.0f+0.5f)); + __m128i cb_const1 = _mm_set1_epi16( (short) ( 1.77200f*4096.0f+0.5f)); + __m128i y_bias = _mm_set1_epi8((char) (unsigned char) 128); + __m128i xw = _mm_set1_epi16(255); // alpha channel + + for (; i+7 < count; i += 8) { + // load + __m128i y_bytes = _mm_loadl_epi64((__m128i *) (y+i)); + __m128i cr_bytes = _mm_loadl_epi64((__m128i *) (pcr+i)); + __m128i cb_bytes = _mm_loadl_epi64((__m128i *) (pcb+i)); + __m128i cr_biased = _mm_xor_si128(cr_bytes, signflip); // -128 + __m128i cb_biased = _mm_xor_si128(cb_bytes, signflip); // -128 + + // unpack to short (and left-shift cr, cb by 8) + __m128i yw = _mm_unpacklo_epi8(y_bias, y_bytes); + __m128i crw = _mm_unpacklo_epi8(_mm_setzero_si128(), cr_biased); + __m128i cbw = _mm_unpacklo_epi8(_mm_setzero_si128(), cb_biased); + + // color transform + __m128i yws = _mm_srli_epi16(yw, 4); + __m128i cr0 = _mm_mulhi_epi16(cr_const0, crw); + __m128i cb0 = _mm_mulhi_epi16(cb_const0, cbw); + __m128i cb1 = _mm_mulhi_epi16(cbw, cb_const1); + __m128i cr1 = _mm_mulhi_epi16(crw, cr_const1); + __m128i rws = _mm_add_epi16(cr0, yws); + __m128i gwt = _mm_add_epi16(cb0, yws); + __m128i bws = _mm_add_epi16(yws, cb1); + __m128i gws = _mm_add_epi16(gwt, cr1); + + // descale + __m128i rw = _mm_srai_epi16(rws, 4); + __m128i bw = _mm_srai_epi16(bws, 4); + __m128i gw = _mm_srai_epi16(gws, 4); + + // back to byte, set up for transpose + __m128i brb = _mm_packus_epi16(rw, bw); + __m128i gxb = _mm_packus_epi16(gw, xw); + + // transpose to interleave channels + __m128i t0 = _mm_unpacklo_epi8(brb, gxb); + __m128i t1 = _mm_unpackhi_epi8(brb, gxb); + __m128i o0 = _mm_unpacklo_epi16(t0, t1); + __m128i o1 = _mm_unpackhi_epi16(t0, t1); + + // store + _mm_storeu_si128((__m128i *) (out + 0), o0); + _mm_storeu_si128((__m128i *) (out + 16), o1); + out += 32; + } + } +#endif + +#ifdef STBI_NEON + // in this version, step=3 support would be easy to add. but is there demand? + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + uint8x8_t signflip = vdup_n_u8(0x80); + int16x8_t cr_const0 = vdupq_n_s16( (short) ( 1.40200f*4096.0f+0.5f)); + int16x8_t cr_const1 = vdupq_n_s16( - (short) ( 0.71414f*4096.0f+0.5f)); + int16x8_t cb_const0 = vdupq_n_s16( - (short) ( 0.34414f*4096.0f+0.5f)); + int16x8_t cb_const1 = vdupq_n_s16( (short) ( 1.77200f*4096.0f+0.5f)); + + for (; i+7 < count; i += 8) { + // load + uint8x8_t y_bytes = vld1_u8(y + i); + uint8x8_t cr_bytes = vld1_u8(pcr + i); + uint8x8_t cb_bytes = vld1_u8(pcb + i); + int8x8_t cr_biased = vreinterpret_s8_u8(vsub_u8(cr_bytes, signflip)); + int8x8_t cb_biased = vreinterpret_s8_u8(vsub_u8(cb_bytes, signflip)); + + // expand to s16 + int16x8_t yws = vreinterpretq_s16_u16(vshll_n_u8(y_bytes, 4)); + int16x8_t crw = vshll_n_s8(cr_biased, 7); + int16x8_t cbw = vshll_n_s8(cb_biased, 7); + + // color transform + int16x8_t cr0 = vqdmulhq_s16(crw, cr_const0); + int16x8_t cb0 = vqdmulhq_s16(cbw, cb_const0); + int16x8_t cr1 = vqdmulhq_s16(crw, cr_const1); + int16x8_t cb1 = vqdmulhq_s16(cbw, cb_const1); + int16x8_t rws = vaddq_s16(yws, cr0); + int16x8_t gws = vaddq_s16(vaddq_s16(yws, cb0), cr1); + int16x8_t bws = vaddq_s16(yws, cb1); + + // undo scaling, round, convert to byte + uint8x8x4_t o; + o.val[0] = vqrshrun_n_s16(rws, 4); + o.val[1] = vqrshrun_n_s16(gws, 4); + o.val[2] = vqrshrun_n_s16(bws, 4); + o.val[3] = vdup_n_u8(255); + + // store, interleaving r/g/b/a + vst4_u8(out, o); + out += 8*4; + } + } +#endif + + for (; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + cr*-stbi__float2fixed(0.71414f) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} +#endif + +// set up the kernels +static void stbi__setup_jpeg(stbi__jpeg *j) +{ + j->idct_block_kernel = stbi__idct_block; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_row; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2; + +#ifdef STBI_SSE2 + if (stbi__sse2_available()) { + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; + } +#endif + +#ifdef STBI_NEON + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; +#endif +} + +// clean up the temporary component buffers +static void stbi__cleanup_jpeg(stbi__jpeg *j) +{ + stbi__free_jpeg_components(j, j->s->img_n, 0); +} + +typedef struct +{ + resample_row_func resample; + stbi_uc *line0,*line1; + int hs,vs; // expansion factor in each axis + int w_lores; // horizontal pixels pre-expansion + int ystep; // how far through vertical expansion we are + int ypos; // which pre-expansion row we're on +} stbi__resample; + +// fast 0..255 * 0..255 => 0..255 rounded multiplication +static stbi_uc stbi__blinn_8x8(stbi_uc x, stbi_uc y) +{ + unsigned int t = x*y + 128; + return (stbi_uc) ((t + (t >>8)) >> 8); +} + +static stbi_uc *load_jpeg_image(stbi__jpeg *z, int *out_x, int *out_y, int *comp, int req_comp) +{ + int n, decode_n, is_rgb; + z->s->img_n = 0; // make stbi__cleanup_jpeg safe + + // validate req_comp + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + + // load a jpeg image from whichever source, but leave in YCbCr format + if (!stbi__decode_jpeg_image(z)) { stbi__cleanup_jpeg(z); return NULL; } + + // determine actual number of components to generate + n = req_comp ? req_comp : z->s->img_n >= 3 ? 3 : 1; + + is_rgb = z->s->img_n == 3 && (z->rgb == 3 || (z->app14_color_transform == 0 && !z->jfif)); + + if (z->s->img_n == 3 && n < 3 && !is_rgb) + decode_n = 1; + else + decode_n = z->s->img_n; + + // resample and color-convert + { + int k; + unsigned int i,j; + stbi_uc *output; + stbi_uc *coutput[4]; + + stbi__resample res_comp[4]; + + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + + // allocate line buffer big enough for upsampling off the edges + // with upsample factor of 4 + z->img_comp[k].linebuf = (stbi_uc *) stbi__malloc(z->s->img_x + 3); + if (!z->img_comp[k].linebuf) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + r->hs = z->img_h_max / z->img_comp[k].h; + r->vs = z->img_v_max / z->img_comp[k].v; + r->ystep = r->vs >> 1; + r->w_lores = (z->s->img_x + r->hs-1) / r->hs; + r->ypos = 0; + r->line0 = r->line1 = z->img_comp[k].data; + + if (r->hs == 1 && r->vs == 1) r->resample = resample_row_1; + else if (r->hs == 1 && r->vs == 2) r->resample = stbi__resample_row_v_2; + else if (r->hs == 2 && r->vs == 1) r->resample = stbi__resample_row_h_2; + else if (r->hs == 2 && r->vs == 2) r->resample = z->resample_row_hv_2_kernel; + else r->resample = stbi__resample_row_generic; + } + + // can't error after this so, this is safe + output = (stbi_uc *) stbi__malloc_mad3(n, z->s->img_x, z->s->img_y, 1); + if (!output) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + // now go ahead and resample + for (j=0; j < z->s->img_y; ++j) { + stbi_uc *out = output + n * z->s->img_x * j; + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + int y_bot = r->ystep >= (r->vs >> 1); + coutput[k] = r->resample(z->img_comp[k].linebuf, + y_bot ? r->line1 : r->line0, + y_bot ? r->line0 : r->line1, + r->w_lores, r->hs); + if (++r->ystep >= r->vs) { + r->ystep = 0; + r->line0 = r->line1; + if (++r->ypos < z->img_comp[k].y) + r->line1 += z->img_comp[k].w2; + } + } + if (n >= 3) { + stbi_uc *y = coutput[0]; + if (z->s->img_n == 3) { + if (is_rgb) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = y[i]; + out[1] = coutput[1][i]; + out[2] = coutput[2][i]; + out[3] = 255; + out += n; + } + } else { + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else if (z->s->img_n == 4) { + if (z->app14_color_transform == 0) { // CMYK + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(coutput[0][i], m); + out[1] = stbi__blinn_8x8(coutput[1][i], m); + out[2] = stbi__blinn_8x8(coutput[2][i], m); + out[3] = 255; + out += n; + } + } else if (z->app14_color_transform == 2) { // YCCK + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(255 - out[0], m); + out[1] = stbi__blinn_8x8(255 - out[1], m); + out[2] = stbi__blinn_8x8(255 - out[2], m); + out += n; + } + } else { // YCbCr + alpha? Ignore the fourth channel for now + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else + for (i=0; i < z->s->img_x; ++i) { + out[0] = out[1] = out[2] = y[i]; + out[3] = 255; // not used if n==3 + out += n; + } + } else { + if (is_rgb) { + if (n == 1) + for (i=0; i < z->s->img_x; ++i) + *out++ = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + else { + for (i=0; i < z->s->img_x; ++i, out += 2) { + out[0] = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + out[1] = 255; + } + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 0) { + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + stbi_uc r = stbi__blinn_8x8(coutput[0][i], m); + stbi_uc g = stbi__blinn_8x8(coutput[1][i], m); + stbi_uc b = stbi__blinn_8x8(coutput[2][i], m); + out[0] = stbi__compute_y(r, g, b); + out[1] = 255; + out += n; + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 2) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = stbi__blinn_8x8(255 - coutput[0][i], coutput[3][i]); + out[1] = 255; + out += n; + } + } else { + stbi_uc *y = coutput[0]; + if (n == 1) + for (i=0; i < z->s->img_x; ++i) out[i] = y[i]; + else + for (i=0; i < z->s->img_x; ++i) *out++ = y[i], *out++ = 255; + } + } + } + stbi__cleanup_jpeg(z); + *out_x = z->s->img_x; + *out_y = z->s->img_y; + if (comp) *comp = z->s->img_n >= 3 ? 3 : 1; // report original components, not output + return output; + } +} + +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + unsigned char* result; + stbi__jpeg* j = (stbi__jpeg*) stbi__malloc(sizeof(stbi__jpeg)); + STBI_NOTUSED(ri); + j->s = s; + stbi__setup_jpeg(j); + result = load_jpeg_image(j, x,y,comp,req_comp); + STBI_FREE(j); + return result; +} + +static int stbi__jpeg_test(stbi__context *s) +{ + int r; + stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg)); + j->s = s; + stbi__setup_jpeg(j); + r = stbi__decode_jpeg_header(j, STBI__SCAN_type); + stbi__rewind(s); + STBI_FREE(j); + return r; +} + +static int stbi__jpeg_info_raw(stbi__jpeg *j, int *x, int *y, int *comp) +{ + if (!stbi__decode_jpeg_header(j, STBI__SCAN_header)) { + stbi__rewind( j->s ); + return 0; + } + if (x) *x = j->s->img_x; + if (y) *y = j->s->img_y; + if (comp) *comp = j->s->img_n >= 3 ? 3 : 1; + return 1; +} + +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp) +{ + int result; + stbi__jpeg* j = (stbi__jpeg*) (stbi__malloc(sizeof(stbi__jpeg))); + j->s = s; + result = stbi__jpeg_info_raw(j, x, y, comp); + STBI_FREE(j); + return result; +} +#endif + +// public domain zlib decode v0.2 Sean Barrett 2006-11-18 +// simple implementation +// - all input must be provided in an upfront buffer +// - all output is written to a single output buffer (can malloc/realloc) +// performance +// - fast huffman + +#ifndef STBI_NO_ZLIB + +// fast-way is faster to check than jpeg huffman, but slow way is slower +#define STBI__ZFAST_BITS 9 // accelerate all cases in default tables +#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1) + +// zlib-style huffman encoding +// (jpegs packs from left, zlib from right, so can't share code) +typedef struct +{ + stbi__uint16 fast[1 << STBI__ZFAST_BITS]; + stbi__uint16 firstcode[16]; + int maxcode[17]; + stbi__uint16 firstsymbol[16]; + stbi_uc size[288]; + stbi__uint16 value[288]; +} stbi__zhuffman; + +stbi_inline static int stbi__bitreverse16(int n) +{ + n = ((n & 0xAAAA) >> 1) | ((n & 0x5555) << 1); + n = ((n & 0xCCCC) >> 2) | ((n & 0x3333) << 2); + n = ((n & 0xF0F0) >> 4) | ((n & 0x0F0F) << 4); + n = ((n & 0xFF00) >> 8) | ((n & 0x00FF) << 8); + return n; +} + +stbi_inline static int stbi__bit_reverse(int v, int bits) +{ + STBI_ASSERT(bits <= 16); + // to bit reverse n bits, reverse 16 and shift + // e.g. 11 bits, bit reverse and shift away 5 + return stbi__bitreverse16(v) >> (16-bits); +} + +static int stbi__zbuild_huffman(stbi__zhuffman *z, const stbi_uc *sizelist, int num) +{ + int i,k=0; + int code, next_code[16], sizes[17]; + + // DEFLATE spec for generating codes + memset(sizes, 0, sizeof(sizes)); + memset(z->fast, 0, sizeof(z->fast)); + for (i=0; i < num; ++i) + ++sizes[sizelist[i]]; + sizes[0] = 0; + for (i=1; i < 16; ++i) + if (sizes[i] > (1 << i)) + return stbi__err("bad sizes", "Corrupt PNG"); + code = 0; + for (i=1; i < 16; ++i) { + next_code[i] = code; + z->firstcode[i] = (stbi__uint16) code; + z->firstsymbol[i] = (stbi__uint16) k; + code = (code + sizes[i]); + if (sizes[i]) + if (code-1 >= (1 << i)) return stbi__err("bad codelengths","Corrupt PNG"); + z->maxcode[i] = code << (16-i); // preshift for inner loop + code <<= 1; + k += sizes[i]; + } + z->maxcode[16] = 0x10000; // sentinel + for (i=0; i < num; ++i) { + int s = sizelist[i]; + if (s) { + int c = next_code[s] - z->firstcode[s] + z->firstsymbol[s]; + stbi__uint16 fastv = (stbi__uint16) ((s << 9) | i); + z->size [c] = (stbi_uc ) s; + z->value[c] = (stbi__uint16) i; + if (s <= STBI__ZFAST_BITS) { + int j = stbi__bit_reverse(next_code[s],s); + while (j < (1 << STBI__ZFAST_BITS)) { + z->fast[j] = fastv; + j += (1 << s); + } + } + ++next_code[s]; + } + } + return 1; +} + +// zlib-from-memory implementation for PNG reading +// because PNG allows splitting the zlib stream arbitrarily, +// and it's annoying structurally to have PNG call ZLIB call PNG, +// we require PNG read all the IDATs and combine them into a single +// memory buffer + +typedef struct +{ + stbi_uc *zbuffer, *zbuffer_end; + int num_bits; + stbi__uint32 code_buffer; + + char *zout; + char *zout_start; + char *zout_end; + int z_expandable; + + stbi__zhuffman z_length, z_distance; +} stbi__zbuf; + +stbi_inline static stbi_uc stbi__zget8(stbi__zbuf *z) +{ + if (z->zbuffer >= z->zbuffer_end) return 0; + return *z->zbuffer++; +} + +static void stbi__fill_bits(stbi__zbuf *z) +{ + do { + STBI_ASSERT(z->code_buffer < (1U << z->num_bits)); + z->code_buffer |= (unsigned int) stbi__zget8(z) << z->num_bits; + z->num_bits += 8; + } while (z->num_bits <= 24); +} + +stbi_inline static unsigned int stbi__zreceive(stbi__zbuf *z, int n) +{ + unsigned int k; + if (z->num_bits < n) stbi__fill_bits(z); + k = z->code_buffer & ((1 << n) - 1); + z->code_buffer >>= n; + z->num_bits -= n; + return k; +} + +static int stbi__zhuffman_decode_slowpath(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s,k; + // not resolved by fast table, so compute it the slow way + // use jpeg approach, which requires MSbits at top + k = stbi__bit_reverse(a->code_buffer, 16); + for (s=STBI__ZFAST_BITS+1; ; ++s) + if (k < z->maxcode[s]) + break; + if (s == 16) return -1; // invalid code! + // code size is s, so: + b = (k >> (16-s)) - z->firstcode[s] + z->firstsymbol[s]; + STBI_ASSERT(z->size[b] == s); + a->code_buffer >>= s; + a->num_bits -= s; + return z->value[b]; +} + +stbi_inline static int stbi__zhuffman_decode(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s; + if (a->num_bits < 16) stbi__fill_bits(a); + b = z->fast[a->code_buffer & STBI__ZFAST_MASK]; + if (b) { + s = b >> 9; + a->code_buffer >>= s; + a->num_bits -= s; + return b & 511; + } + return stbi__zhuffman_decode_slowpath(a, z); +} + +static int stbi__zexpand(stbi__zbuf *z, char *zout, int n) // need to make room for n bytes +{ + char *q; + int cur, limit, old_limit; + z->zout = zout; + if (!z->z_expandable) return stbi__err("output buffer limit","Corrupt PNG"); + cur = (int) (z->zout - z->zout_start); + limit = old_limit = (int) (z->zout_end - z->zout_start); + while (cur + n > limit) + limit *= 2; + q = (char *) STBI_REALLOC_SIZED(z->zout_start, old_limit, limit); + STBI_NOTUSED(old_limit); + if (q == NULL) return stbi__err("outofmem", "Out of memory"); + z->zout_start = q; + z->zout = q + cur; + z->zout_end = q + limit; + return 1; +} + +static const int stbi__zlength_base[31] = { + 3,4,5,6,7,8,9,10,11,13, + 15,17,19,23,27,31,35,43,51,59, + 67,83,99,115,131,163,195,227,258,0,0 }; + +static const int stbi__zlength_extra[31]= +{ 0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0,0,0 }; + +static const int stbi__zdist_base[32] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193, +257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577,0,0}; + +static const int stbi__zdist_extra[32] = +{ 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13}; + +static int stbi__parse_huffman_block(stbi__zbuf *a) +{ + char *zout = a->zout; + for(;;) { + int z = stbi__zhuffman_decode(a, &a->z_length); + if (z < 256) { + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); // error in huffman codes + if (zout >= a->zout_end) { + if (!stbi__zexpand(a, zout, 1)) return 0; + zout = a->zout; + } + *zout++ = (char) z; + } else { + stbi_uc *p; + int len,dist; + if (z == 256) { + a->zout = zout; + return 1; + } + z -= 257; + len = stbi__zlength_base[z]; + if (stbi__zlength_extra[z]) len += stbi__zreceive(a, stbi__zlength_extra[z]); + z = stbi__zhuffman_decode(a, &a->z_distance); + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); + dist = stbi__zdist_base[z]; + if (stbi__zdist_extra[z]) dist += stbi__zreceive(a, stbi__zdist_extra[z]); + if (zout - a->zout_start < dist) return stbi__err("bad dist","Corrupt PNG"); + if (zout + len > a->zout_end) { + if (!stbi__zexpand(a, zout, len)) return 0; + zout = a->zout; + } + p = (stbi_uc *) (zout - dist); + if (dist == 1) { // run of one byte; common in images. + stbi_uc v = *p; + if (len) { do *zout++ = v; while (--len); } + } else { + if (len) { do *zout++ = *p++; while (--len); } + } + } + } +} + +static int stbi__compute_huffman_codes(stbi__zbuf *a) +{ + static const stbi_uc length_dezigzag[19] = { 16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15 }; + stbi__zhuffman z_codelength; + stbi_uc lencodes[286+32+137];//padding for maximum single op + stbi_uc codelength_sizes[19]; + int i,n; + + int hlit = stbi__zreceive(a,5) + 257; + int hdist = stbi__zreceive(a,5) + 1; + int hclen = stbi__zreceive(a,4) + 4; + int ntot = hlit + hdist; + + memset(codelength_sizes, 0, sizeof(codelength_sizes)); + for (i=0; i < hclen; ++i) { + int s = stbi__zreceive(a,3); + codelength_sizes[length_dezigzag[i]] = (stbi_uc) s; + } + if (!stbi__zbuild_huffman(&z_codelength, codelength_sizes, 19)) return 0; + + n = 0; + while (n < ntot) { + int c = stbi__zhuffman_decode(a, &z_codelength); + if (c < 0 || c >= 19) return stbi__err("bad codelengths", "Corrupt PNG"); + if (c < 16) + lencodes[n++] = (stbi_uc) c; + else { + stbi_uc fill = 0; + if (c == 16) { + c = stbi__zreceive(a,2)+3; + if (n == 0) return stbi__err("bad codelengths", "Corrupt PNG"); + fill = lencodes[n-1]; + } else if (c == 17) + c = stbi__zreceive(a,3)+3; + else { + STBI_ASSERT(c == 18); + c = stbi__zreceive(a,7)+11; + } + if (ntot - n < c) return stbi__err("bad codelengths", "Corrupt PNG"); + memset(lencodes+n, fill, c); + n += c; + } + } + if (n != ntot) return stbi__err("bad codelengths","Corrupt PNG"); + if (!stbi__zbuild_huffman(&a->z_length, lencodes, hlit)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, lencodes+hlit, hdist)) return 0; + return 1; +} + +static int stbi__parse_uncompressed_block(stbi__zbuf *a) +{ + stbi_uc header[4]; + int len,nlen,k; + if (a->num_bits & 7) + stbi__zreceive(a, a->num_bits & 7); // discard + // drain the bit-packed data into header + k = 0; + while (a->num_bits > 0) { + header[k++] = (stbi_uc) (a->code_buffer & 255); // suppress MSVC run-time check + a->code_buffer >>= 8; + a->num_bits -= 8; + } + STBI_ASSERT(a->num_bits == 0); + // now fill header the normal way + while (k < 4) + header[k++] = stbi__zget8(a); + len = header[1] * 256 + header[0]; + nlen = header[3] * 256 + header[2]; + if (nlen != (len ^ 0xffff)) return stbi__err("zlib corrupt","Corrupt PNG"); + if (a->zbuffer + len > a->zbuffer_end) return stbi__err("read past buffer","Corrupt PNG"); + if (a->zout + len > a->zout_end) + if (!stbi__zexpand(a, a->zout, len)) return 0; + memcpy(a->zout, a->zbuffer, len); + a->zbuffer += len; + a->zout += len; + return 1; +} + +static int stbi__parse_zlib_header(stbi__zbuf *a) +{ + int cmf = stbi__zget8(a); + int cm = cmf & 15; + /* int cinfo = cmf >> 4; */ + int flg = stbi__zget8(a); + if ((cmf*256+flg) % 31 != 0) return stbi__err("bad zlib header","Corrupt PNG"); // zlib spec + if (flg & 32) return stbi__err("no preset dict","Corrupt PNG"); // preset dictionary not allowed in png + if (cm != 8) return stbi__err("bad compression","Corrupt PNG"); // DEFLATE required for png + // window = 1 << (8 + cinfo)... but who cares, we fully buffer output + return 1; +} + +static const stbi_uc stbi__zdefault_length[288] = +{ + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, 7,7,7,7,7,7,7,7,8,8,8,8,8,8,8,8 +}; +static const stbi_uc stbi__zdefault_distance[32] = +{ + 5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5 +}; +/* +Init algorithm: +{ + int i; // use <= to match clearly with spec + for (i=0; i <= 143; ++i) stbi__zdefault_length[i] = 8; + for ( ; i <= 255; ++i) stbi__zdefault_length[i] = 9; + for ( ; i <= 279; ++i) stbi__zdefault_length[i] = 7; + for ( ; i <= 287; ++i) stbi__zdefault_length[i] = 8; + for (i=0; i <= 31; ++i) stbi__zdefault_distance[i] = 5; +} +*/ + +static int stbi__parse_zlib(stbi__zbuf *a, int parse_header) +{ + int final, type; + if (parse_header) + if (!stbi__parse_zlib_header(a)) return 0; + a->num_bits = 0; + a->code_buffer = 0; + do { + final = stbi__zreceive(a,1); + type = stbi__zreceive(a,2); + if (type == 0) { + if (!stbi__parse_uncompressed_block(a)) return 0; + } else if (type == 3) { + return 0; + } else { + if (type == 1) { + // use fixed code lengths + if (!stbi__zbuild_huffman(&a->z_length , stbi__zdefault_length , 288)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, stbi__zdefault_distance, 32)) return 0; + } else { + if (!stbi__compute_huffman_codes(a)) return 0; + } + if (!stbi__parse_huffman_block(a)) return 0; + } + } while (!final); + return 1; +} + +static int stbi__do_zlib(stbi__zbuf *a, char *obuf, int olen, int exp, int parse_header) +{ + a->zout_start = obuf; + a->zout = obuf; + a->zout_end = obuf + olen; + a->z_expandable = exp; + + return stbi__parse_zlib(a, parse_header); +} + +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, 1)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF char *stbi_zlib_decode_malloc(char const *buffer, int len, int *outlen) +{ + return stbi_zlib_decode_malloc_guesssize(buffer, len, 16384, outlen); +} + +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, parse_header)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, char const *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 1)) + return (int) (a.zout - a.zout_start); + else + return -1; +} + +STBIDEF char *stbi_zlib_decode_noheader_malloc(char const *buffer, int len, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(16384); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer+len; + if (stbi__do_zlib(&a, p, 16384, 1, 0)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 0)) + return (int) (a.zout - a.zout_start); + else + return -1; +} +#endif + +// public domain "baseline" PNG decoder v0.10 Sean Barrett 2006-11-18 +// simple implementation +// - only 8-bit samples +// - no CRC checking +// - allocates lots of intermediate memory +// - avoids problem of streaming data between subsystems +// - avoids explicit window management +// performance +// - uses stb_zlib, a PD zlib implementation with fast huffman decoding + +#ifndef STBI_NO_PNG +typedef struct +{ + stbi__uint32 length; + stbi__uint32 type; +} stbi__pngchunk; + +static stbi__pngchunk stbi__get_chunk_header(stbi__context *s) +{ + stbi__pngchunk c; + c.length = stbi__get32be(s); + c.type = stbi__get32be(s); + return c; +} + +static int stbi__check_png_header(stbi__context *s) +{ + static const stbi_uc png_sig[8] = { 137,80,78,71,13,10,26,10 }; + int i; + for (i=0; i < 8; ++i) + if (stbi__get8(s) != png_sig[i]) return stbi__err("bad png sig","Not a PNG"); + return 1; +} + +typedef struct +{ + stbi__context *s; + stbi_uc *idata, *expanded, *out; + int depth; +} stbi__png; + + +enum { + STBI__F_none=0, + STBI__F_sub=1, + STBI__F_up=2, + STBI__F_avg=3, + STBI__F_paeth=4, + // synthetic filters used for first scanline to avoid needing a dummy row of 0s + STBI__F_avg_first, + STBI__F_paeth_first +}; + +static stbi_uc first_row_filter[5] = +{ + STBI__F_none, + STBI__F_sub, + STBI__F_none, + STBI__F_avg_first, + STBI__F_paeth_first +}; + +static int stbi__paeth(int a, int b, int c) +{ + int p = a + b - c; + int pa = abs(p-a); + int pb = abs(p-b); + int pc = abs(p-c); + if (pa <= pb && pa <= pc) return a; + if (pb <= pc) return b; + return c; +} + +static const stbi_uc stbi__depth_scale_table[9] = { 0, 0xff, 0x55, 0, 0x11, 0,0,0, 0x01 }; + +// create the png data from post-deflated data +static int stbi__create_png_image_raw(stbi__png *a, stbi_uc *raw, stbi__uint32 raw_len, int out_n, stbi__uint32 x, stbi__uint32 y, int depth, int color) +{ + int bytes = (depth == 16? 2 : 1); + stbi__context *s = a->s; + stbi__uint32 i,j,stride = x*out_n*bytes; + stbi__uint32 img_len, img_width_bytes; + int k; + int img_n = s->img_n; // copy it into a local for later + + int output_bytes = out_n*bytes; + int filter_bytes = img_n*bytes; + int width = x; + + STBI_ASSERT(out_n == s->img_n || out_n == s->img_n+1); + a->out = (stbi_uc *) stbi__malloc_mad3(x, y, output_bytes, 0); // extra bytes to write off the end into + if (!a->out) return stbi__err("outofmem", "Out of memory"); + + if (!stbi__mad3sizes_valid(img_n, x, depth, 7)) return stbi__err("too large", "Corrupt PNG"); + img_width_bytes = (((img_n * x * depth) + 7) >> 3); + img_len = (img_width_bytes + 1) * y; + + // we used to check for exact match between raw_len and img_len on non-interlaced PNGs, + // but issue #276 reported a PNG in the wild that had extra data at the end (all zeros), + // so just check for raw_len < img_len always. + if (raw_len < img_len) return stbi__err("not enough pixels","Corrupt PNG"); + + for (j=0; j < y; ++j) { + stbi_uc *cur = a->out + stride*j; + stbi_uc *prior; + int filter = *raw++; + + if (filter > 4) + return stbi__err("invalid filter","Corrupt PNG"); + + if (depth < 8) { + STBI_ASSERT(img_width_bytes <= x); + cur += x*out_n - img_width_bytes; // store output to the rightmost img_len bytes, so we can decode in place + filter_bytes = 1; + width = img_width_bytes; + } + prior = cur - stride; // bugfix: need to compute this after 'cur +=' computation above + + // if first row, use special filter that doesn't sample previous row + if (j == 0) filter = first_row_filter[filter]; + + // handle first byte explicitly + for (k=0; k < filter_bytes; ++k) { + switch (filter) { + case STBI__F_none : cur[k] = raw[k]; break; + case STBI__F_sub : cur[k] = raw[k]; break; + case STBI__F_up : cur[k] = STBI__BYTECAST(raw[k] + prior[k]); break; + case STBI__F_avg : cur[k] = STBI__BYTECAST(raw[k] + (prior[k]>>1)); break; + case STBI__F_paeth : cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(0,prior[k],0)); break; + case STBI__F_avg_first : cur[k] = raw[k]; break; + case STBI__F_paeth_first: cur[k] = raw[k]; break; + } + } + + if (depth == 8) { + if (img_n != out_n) + cur[img_n] = 255; // first pixel + raw += img_n; + cur += out_n; + prior += out_n; + } else if (depth == 16) { + if (img_n != out_n) { + cur[filter_bytes] = 255; // first pixel top byte + cur[filter_bytes+1] = 255; // first pixel bottom byte + } + raw += filter_bytes; + cur += output_bytes; + prior += output_bytes; + } else { + raw += 1; + cur += 1; + prior += 1; + } + + // this is a little gross, so that we don't switch per-pixel or per-component + if (depth < 8 || img_n == out_n) { + int nk = (width - 1)*filter_bytes; + #define STBI__CASE(f) \ + case f: \ + for (k=0; k < nk; ++k) + switch (filter) { + // "none" filter turns into a memcpy here; make that explicit. + case STBI__F_none: memcpy(cur, raw, nk); break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k-filter_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k-filter_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes],prior[k],prior[k-filter_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k-filter_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes],0,0)); } break; + } + #undef STBI__CASE + raw += nk; + } else { + STBI_ASSERT(img_n+1 == out_n); + #define STBI__CASE(f) \ + case f: \ + for (i=x-1; i >= 1; --i, cur[filter_bytes]=255,raw+=filter_bytes,cur+=output_bytes,prior+=output_bytes) \ + for (k=0; k < filter_bytes; ++k) + switch (filter) { + STBI__CASE(STBI__F_none) { cur[k] = raw[k]; } break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k- output_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k- output_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k- output_bytes],prior[k],prior[k- output_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k- output_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k- output_bytes],0,0)); } break; + } + #undef STBI__CASE + + // the loop above sets the high byte of the pixels' alpha, but for + // 16 bit png files we also need the low byte set. we'll do that here. + if (depth == 16) { + cur = a->out + stride*j; // start at the beginning of the row again + for (i=0; i < x; ++i,cur+=output_bytes) { + cur[filter_bytes+1] = 255; + } + } + } + } + + // we make a separate pass to expand bits to pixels; for performance, + // this could run two scanlines behind the above code, so it won't + // intefere with filtering but will still be in the cache. + if (depth < 8) { + for (j=0; j < y; ++j) { + stbi_uc *cur = a->out + stride*j; + stbi_uc *in = a->out + stride*j + x*out_n - img_width_bytes; + // unpack 1/2/4-bit into a 8-bit buffer. allows us to keep the common 8-bit path optimal at minimal cost for 1/2/4-bit + // png guarante byte alignment, if width is not multiple of 8/4/2 we'll decode dummy trailing data that will be skipped in the later loop + stbi_uc scale = (color == 0) ? stbi__depth_scale_table[depth] : 1; // scale grayscale values to 0..255 range + + // note that the final byte might overshoot and write more data than desired. + // we can allocate enough data that this never writes out of memory, but it + // could also overwrite the next scanline. can it overwrite non-empty data + // on the next scanline? yes, consider 1-pixel-wide scanlines with 1-bit-per-pixel. + // so we need to explicitly clamp the final ones + + if (depth == 4) { + for (k=x*img_n; k >= 2; k-=2, ++in) { + *cur++ = scale * ((*in >> 4) ); + *cur++ = scale * ((*in ) & 0x0f); + } + if (k > 0) *cur++ = scale * ((*in >> 4) ); + } else if (depth == 2) { + for (k=x*img_n; k >= 4; k-=4, ++in) { + *cur++ = scale * ((*in >> 6) ); + *cur++ = scale * ((*in >> 4) & 0x03); + *cur++ = scale * ((*in >> 2) & 0x03); + *cur++ = scale * ((*in ) & 0x03); + } + if (k > 0) *cur++ = scale * ((*in >> 6) ); + if (k > 1) *cur++ = scale * ((*in >> 4) & 0x03); + if (k > 2) *cur++ = scale * ((*in >> 2) & 0x03); + } else if (depth == 1) { + for (k=x*img_n; k >= 8; k-=8, ++in) { + *cur++ = scale * ((*in >> 7) ); + *cur++ = scale * ((*in >> 6) & 0x01); + *cur++ = scale * ((*in >> 5) & 0x01); + *cur++ = scale * ((*in >> 4) & 0x01); + *cur++ = scale * ((*in >> 3) & 0x01); + *cur++ = scale * ((*in >> 2) & 0x01); + *cur++ = scale * ((*in >> 1) & 0x01); + *cur++ = scale * ((*in ) & 0x01); + } + if (k > 0) *cur++ = scale * ((*in >> 7) ); + if (k > 1) *cur++ = scale * ((*in >> 6) & 0x01); + if (k > 2) *cur++ = scale * ((*in >> 5) & 0x01); + if (k > 3) *cur++ = scale * ((*in >> 4) & 0x01); + if (k > 4) *cur++ = scale * ((*in >> 3) & 0x01); + if (k > 5) *cur++ = scale * ((*in >> 2) & 0x01); + if (k > 6) *cur++ = scale * ((*in >> 1) & 0x01); + } + if (img_n != out_n) { + int q; + // insert alpha = 255 + cur = a->out + stride*j; + if (img_n == 1) { + for (q=x-1; q >= 0; --q) { + cur[q*2+1] = 255; + cur[q*2+0] = cur[q]; + } + } else { + STBI_ASSERT(img_n == 3); + for (q=x-1; q >= 0; --q) { + cur[q*4+3] = 255; + cur[q*4+2] = cur[q*3+2]; + cur[q*4+1] = cur[q*3+1]; + cur[q*4+0] = cur[q*3+0]; + } + } + } + } + } else if (depth == 16) { + // force the image data from big-endian to platform-native. + // this is done in a separate pass due to the decoding relying + // on the data being untouched, but could probably be done + // per-line during decode if care is taken. + stbi_uc *cur = a->out; + stbi__uint16 *cur16 = (stbi__uint16*)cur; + + for(i=0; i < x*y*out_n; ++i,cur16++,cur+=2) { + *cur16 = (cur[0] << 8) | cur[1]; + } + } + + return 1; +} + +static int stbi__create_png_image(stbi__png *a, stbi_uc *image_data, stbi__uint32 image_data_len, int out_n, int depth, int color, int interlaced) +{ + int bytes = (depth == 16 ? 2 : 1); + int out_bytes = out_n * bytes; + stbi_uc *final; + int p; + if (!interlaced) + return stbi__create_png_image_raw(a, image_data, image_data_len, out_n, a->s->img_x, a->s->img_y, depth, color); + + // de-interlacing + final = (stbi_uc *) stbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0); + for (p=0; p < 7; ++p) { + int xorig[] = { 0,4,0,2,0,1,0 }; + int yorig[] = { 0,0,4,0,2,0,1 }; + int xspc[] = { 8,8,4,4,2,2,1 }; + int yspc[] = { 8,8,8,4,4,2,2 }; + int i,j,x,y; + // pass1_x[4] = 0, pass1_x[5] = 1, pass1_x[12] = 1 + x = (a->s->img_x - xorig[p] + xspc[p]-1) / xspc[p]; + y = (a->s->img_y - yorig[p] + yspc[p]-1) / yspc[p]; + if (x && y) { + stbi__uint32 img_len = ((((a->s->img_n * x * depth) + 7) >> 3) + 1) * y; + if (!stbi__create_png_image_raw(a, image_data, image_data_len, out_n, x, y, depth, color)) { + STBI_FREE(final); + return 0; + } + for (j=0; j < y; ++j) { + for (i=0; i < x; ++i) { + int out_y = j*yspc[p]+yorig[p]; + int out_x = i*xspc[p]+xorig[p]; + memcpy(final + out_y*a->s->img_x*out_bytes + out_x*out_bytes, + a->out + (j*x+i)*out_bytes, out_bytes); + } + } + STBI_FREE(a->out); + image_data += img_len; + image_data_len -= img_len; + } + } + a->out = final; + + return 1; +} + +static int stbi__compute_transparency(stbi__png *z, stbi_uc tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + // compute color-based transparency, assuming we've + // already got 255 as the alpha value in the output + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i=0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 255); + p += 2; + } + } else { + for (i=0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int stbi__compute_transparency16(stbi__png *z, stbi__uint16 tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi__uint16 *p = (stbi__uint16*) z->out; + + // compute color-based transparency, assuming we've + // already got 65535 as the alpha value in the output + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i = 0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 65535); + p += 2; + } + } else { + for (i = 0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int stbi__expand_png_palette(stbi__png *a, stbi_uc *palette, int len, int pal_img_n) +{ + stbi__uint32 i, pixel_count = a->s->img_x * a->s->img_y; + stbi_uc *p, *temp_out, *orig = a->out; + + p = (stbi_uc *) stbi__malloc_mad2(pixel_count, pal_img_n, 0); + if (p == NULL) return stbi__err("outofmem", "Out of memory"); + + // between here and free(out) below, exitting would leak + temp_out = p; + + if (pal_img_n == 3) { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p += 3; + } + } else { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p[3] = palette[n+3]; + p += 4; + } + } + STBI_FREE(a->out); + a->out = temp_out; + + STBI_NOTUSED(len); + + return 1; +} + +static int stbi__unpremultiply_on_load = 0; +static int stbi__de_iphone_flag = 0; + +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply) +{ + stbi__unpremultiply_on_load = flag_true_if_should_unpremultiply; +} + +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert) +{ + stbi__de_iphone_flag = flag_true_if_should_convert; +} + +static void stbi__de_iphone(stbi__png *z) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + if (s->img_out_n == 3) { // convert bgr to rgb + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 3; + } + } else { + STBI_ASSERT(s->img_out_n == 4); + if (stbi__unpremultiply_on_load) { + // convert bgr to rgb and unpremultiply + for (i=0; i < pixel_count; ++i) { + stbi_uc a = p[3]; + stbi_uc t = p[0]; + if (a) { + stbi_uc half = a / 2; + p[0] = (p[2] * 255 + half) / a; + p[1] = (p[1] * 255 + half) / a; + p[2] = ( t * 255 + half) / a; + } else { + p[0] = p[2]; + p[2] = t; + } + p += 4; + } + } else { + // convert bgr to rgb + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 4; + } + } + } +} + +#define STBI__PNG_TYPE(a,b,c,d) (((unsigned) (a) << 24) + ((unsigned) (b) << 16) + ((unsigned) (c) << 8) + (unsigned) (d)) + +static int stbi__parse_png_file(stbi__png *z, int scan, int req_comp) +{ + stbi_uc palette[1024], pal_img_n=0; + stbi_uc has_trans=0, tc[3]; + stbi__uint16 tc16[3]; + stbi__uint32 ioff=0, idata_limit=0, i, pal_len=0; + int first=1,k,interlace=0, color=0, is_iphone=0; + stbi__context *s = z->s; + + z->expanded = NULL; + z->idata = NULL; + z->out = NULL; + + if (!stbi__check_png_header(s)) return 0; + + if (scan == STBI__SCAN_type) return 1; + + for (;;) { + stbi__pngchunk c = stbi__get_chunk_header(s); + switch (c.type) { + case STBI__PNG_TYPE('C','g','B','I'): + is_iphone = 1; + stbi__skip(s, c.length); + break; + case STBI__PNG_TYPE('I','H','D','R'): { + int comp,filter; + if (!first) return stbi__err("multiple IHDR","Corrupt PNG"); + first = 0; + if (c.length != 13) return stbi__err("bad IHDR len","Corrupt PNG"); + s->img_x = stbi__get32be(s); if (s->img_x > (1 << 24)) return stbi__err("too large","Very large image (corrupt?)"); + s->img_y = stbi__get32be(s); if (s->img_y > (1 << 24)) return stbi__err("too large","Very large image (corrupt?)"); + z->depth = stbi__get8(s); if (z->depth != 1 && z->depth != 2 && z->depth != 4 && z->depth != 8 && z->depth != 16) return stbi__err("1/2/4/8/16-bit only","PNG not supported: 1/2/4/8/16-bit only"); + color = stbi__get8(s); if (color > 6) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3 && z->depth == 16) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3) pal_img_n = 3; else if (color & 1) return stbi__err("bad ctype","Corrupt PNG"); + comp = stbi__get8(s); if (comp) return stbi__err("bad comp method","Corrupt PNG"); + filter= stbi__get8(s); if (filter) return stbi__err("bad filter method","Corrupt PNG"); + interlace = stbi__get8(s); if (interlace>1) return stbi__err("bad interlace method","Corrupt PNG"); + if (!s->img_x || !s->img_y) return stbi__err("0-pixel image","Corrupt PNG"); + if (!pal_img_n) { + s->img_n = (color & 2 ? 3 : 1) + (color & 4 ? 1 : 0); + if ((1 << 30) / s->img_x / s->img_n < s->img_y) return stbi__err("too large", "Image too large to decode"); + if (scan == STBI__SCAN_header) return 1; + } else { + // if paletted, then pal_n is our final components, and + // img_n is # components to decompress/filter. + s->img_n = 1; + if ((1 << 30) / s->img_x / 4 < s->img_y) return stbi__err("too large","Corrupt PNG"); + // if SCAN_header, have to scan to see if we have a tRNS + } + break; + } + + case STBI__PNG_TYPE('P','L','T','E'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (c.length > 256*3) return stbi__err("invalid PLTE","Corrupt PNG"); + pal_len = c.length / 3; + if (pal_len * 3 != c.length) return stbi__err("invalid PLTE","Corrupt PNG"); + for (i=0; i < pal_len; ++i) { + palette[i*4+0] = stbi__get8(s); + palette[i*4+1] = stbi__get8(s); + palette[i*4+2] = stbi__get8(s); + palette[i*4+3] = 255; + } + break; + } + + case STBI__PNG_TYPE('t','R','N','S'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (z->idata) return stbi__err("tRNS after IDAT","Corrupt PNG"); + if (pal_img_n) { + if (scan == STBI__SCAN_header) { s->img_n = 4; return 1; } + if (pal_len == 0) return stbi__err("tRNS before PLTE","Corrupt PNG"); + if (c.length > pal_len) return stbi__err("bad tRNS len","Corrupt PNG"); + pal_img_n = 4; + for (i=0; i < c.length; ++i) + palette[i*4+3] = stbi__get8(s); + } else { + if (!(s->img_n & 1)) return stbi__err("tRNS with alpha","Corrupt PNG"); + if (c.length != (stbi__uint32) s->img_n*2) return stbi__err("bad tRNS len","Corrupt PNG"); + has_trans = 1; + if (z->depth == 16) { + for (k = 0; k < s->img_n; ++k) tc16[k] = (stbi__uint16)stbi__get16be(s); // copy the values as-is + } else { + for (k = 0; k < s->img_n; ++k) tc[k] = (stbi_uc)(stbi__get16be(s) & 255) * stbi__depth_scale_table[z->depth]; // non 8-bit images will be larger + } + } + break; + } + + case STBI__PNG_TYPE('I','D','A','T'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (pal_img_n && !pal_len) return stbi__err("no PLTE","Corrupt PNG"); + if (scan == STBI__SCAN_header) { s->img_n = pal_img_n; return 1; } + if ((int)(ioff + c.length) < (int)ioff) return 0; + if (ioff + c.length > idata_limit) { + stbi__uint32 idata_limit_old = idata_limit; + stbi_uc *p; + if (idata_limit == 0) idata_limit = c.length > 4096 ? c.length : 4096; + while (ioff + c.length > idata_limit) + idata_limit *= 2; + STBI_NOTUSED(idata_limit_old); + p = (stbi_uc *) STBI_REALLOC_SIZED(z->idata, idata_limit_old, idata_limit); if (p == NULL) return stbi__err("outofmem", "Out of memory"); + z->idata = p; + } + if (!stbi__getn(s, z->idata+ioff,c.length)) return stbi__err("outofdata","Corrupt PNG"); + ioff += c.length; + break; + } + + case STBI__PNG_TYPE('I','E','N','D'): { + stbi__uint32 raw_len, bpl; + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (scan != STBI__SCAN_load) return 1; + if (z->idata == NULL) return stbi__err("no IDAT","Corrupt PNG"); + // initial guess for decoded data size to avoid unnecessary reallocs + bpl = (s->img_x * z->depth + 7) / 8; // bytes per line, per component + raw_len = bpl * s->img_y * s->img_n /* pixels */ + s->img_y /* filter mode per row */; + z->expanded = (stbi_uc *) stbi_zlib_decode_malloc_guesssize_headerflag((char *) z->idata, ioff, raw_len, (int *) &raw_len, !is_iphone); + if (z->expanded == NULL) return 0; // zlib should set error + STBI_FREE(z->idata); z->idata = NULL; + if ((req_comp == s->img_n+1 && req_comp != 3 && !pal_img_n) || has_trans) + s->img_out_n = s->img_n+1; + else + s->img_out_n = s->img_n; + if (!stbi__create_png_image(z, z->expanded, raw_len, s->img_out_n, z->depth, color, interlace)) return 0; + if (has_trans) { + if (z->depth == 16) { + if (!stbi__compute_transparency16(z, tc16, s->img_out_n)) return 0; + } else { + if (!stbi__compute_transparency(z, tc, s->img_out_n)) return 0; + } + } + if (is_iphone && stbi__de_iphone_flag && s->img_out_n > 2) + stbi__de_iphone(z); + if (pal_img_n) { + // pal_img_n == 3 or 4 + s->img_n = pal_img_n; // record the actual colors we had + s->img_out_n = pal_img_n; + if (req_comp >= 3) s->img_out_n = req_comp; + if (!stbi__expand_png_palette(z, palette, pal_len, s->img_out_n)) + return 0; + } else if (has_trans) { + // non-paletted image with tRNS -> source image has (constant) alpha + ++s->img_n; + } + STBI_FREE(z->expanded); z->expanded = NULL; + return 1; + } + + default: + // if critical, fail + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if ((c.type & (1 << 29)) == 0) { + #ifndef STBI_NO_FAILURE_STRINGS + // not threadsafe + static char invalid_chunk[] = "XXXX PNG chunk not known"; + invalid_chunk[0] = STBI__BYTECAST(c.type >> 24); + invalid_chunk[1] = STBI__BYTECAST(c.type >> 16); + invalid_chunk[2] = STBI__BYTECAST(c.type >> 8); + invalid_chunk[3] = STBI__BYTECAST(c.type >> 0); + #endif + return stbi__err(invalid_chunk, "PNG not supported: unknown PNG chunk type"); + } + stbi__skip(s, c.length); + break; + } + // end of PNG chunk, read and skip CRC + stbi__get32be(s); + } +} + +static void *stbi__do_png(stbi__png *p, int *x, int *y, int *n, int req_comp, stbi__result_info *ri) +{ + void *result=NULL; + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + if (stbi__parse_png_file(p, STBI__SCAN_load, req_comp)) { + if (p->depth < 8) + ri->bits_per_channel = 8; + else + ri->bits_per_channel = p->depth; + result = p->out; + p->out = NULL; + if (req_comp && req_comp != p->s->img_out_n) { + if (ri->bits_per_channel == 8) + result = stbi__convert_format((unsigned char *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + else + result = stbi__convert_format16((stbi__uint16 *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + p->s->img_out_n = req_comp; + if (result == NULL) return result; + } + *x = p->s->img_x; + *y = p->s->img_y; + if (n) *n = p->s->img_n; + } + STBI_FREE(p->out); p->out = NULL; + STBI_FREE(p->expanded); p->expanded = NULL; + STBI_FREE(p->idata); p->idata = NULL; + + return result; +} + +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi__png p; + p.s = s; + return stbi__do_png(&p, x,y,comp,req_comp, ri); +} + +static int stbi__png_test(stbi__context *s) +{ + int r; + r = stbi__check_png_header(s); + stbi__rewind(s); + return r; +} + +static int stbi__png_info_raw(stbi__png *p, int *x, int *y, int *comp) +{ + if (!stbi__parse_png_file(p, STBI__SCAN_header, 0)) { + stbi__rewind( p->s ); + return 0; + } + if (x) *x = p->s->img_x; + if (y) *y = p->s->img_y; + if (comp) *comp = p->s->img_n; + return 1; +} + +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__png p; + p.s = s; + return stbi__png_info_raw(&p, x, y, comp); +} + +static int stbi__png_is16(stbi__context *s) +{ + stbi__png p; + p.s = s; + if (!stbi__png_info_raw(&p, NULL, NULL, NULL)) + return 0; + if (p.depth != 16) { + stbi__rewind(p.s); + return 0; + } + return 1; +} +#endif + +// Microsoft/Windows BMP image + +#ifndef STBI_NO_BMP +static int stbi__bmp_test_raw(stbi__context *s) +{ + int r; + int sz; + if (stbi__get8(s) != 'B') return 0; + if (stbi__get8(s) != 'M') return 0; + stbi__get32le(s); // discard filesize + stbi__get16le(s); // discard reserved + stbi__get16le(s); // discard reserved + stbi__get32le(s); // discard data offset + sz = stbi__get32le(s); + r = (sz == 12 || sz == 40 || sz == 56 || sz == 108 || sz == 124); + return r; +} + +static int stbi__bmp_test(stbi__context *s) +{ + int r = stbi__bmp_test_raw(s); + stbi__rewind(s); + return r; +} + + +// returns 0..31 for the highest set bit +static int stbi__high_bit(unsigned int z) +{ + int n=0; + if (z == 0) return -1; + if (z >= 0x10000) n += 16, z >>= 16; + if (z >= 0x00100) n += 8, z >>= 8; + if (z >= 0x00010) n += 4, z >>= 4; + if (z >= 0x00004) n += 2, z >>= 2; + if (z >= 0x00002) n += 1, z >>= 1; + return n; +} + +static int stbi__bitcount(unsigned int a) +{ + a = (a & 0x55555555) + ((a >> 1) & 0x55555555); // max 2 + a = (a & 0x33333333) + ((a >> 2) & 0x33333333); // max 4 + a = (a + (a >> 4)) & 0x0f0f0f0f; // max 8 per 4, now 8 bits + a = (a + (a >> 8)); // max 16 per 8 bits + a = (a + (a >> 16)); // max 32 per 8 bits + return a & 0xff; +} + +// extract an arbitrarily-aligned N-bit value (N=bits) +// from v, and then make it 8-bits long and fractionally +// extend it to full full range. +static int stbi__shiftsigned(int v, int shift, int bits) +{ + static unsigned int mul_table[9] = { + 0, + 0xff/*0b11111111*/, 0x55/*0b01010101*/, 0x49/*0b01001001*/, 0x11/*0b00010001*/, + 0x21/*0b00100001*/, 0x41/*0b01000001*/, 0x81/*0b10000001*/, 0x01/*0b00000001*/, + }; + static unsigned int shift_table[9] = { + 0, 0,0,1,0,2,4,6,0, + }; + if (shift < 0) + v <<= -shift; + else + v >>= shift; + STBI_ASSERT(v >= 0 && v < 256); + v >>= (8-bits); + STBI_ASSERT(bits >= 0 && bits <= 8); + return (int) ((unsigned) v * mul_table[bits]) >> shift_table[bits]; +} + +typedef struct +{ + int bpp, offset, hsz; + unsigned int mr,mg,mb,ma, all_a; +} stbi__bmp_data; + +static void *stbi__bmp_parse_header(stbi__context *s, stbi__bmp_data *info) +{ + int hsz; + if (stbi__get8(s) != 'B' || stbi__get8(s) != 'M') return stbi__errpuc("not BMP", "Corrupt BMP"); + stbi__get32le(s); // discard filesize + stbi__get16le(s); // discard reserved + stbi__get16le(s); // discard reserved + info->offset = stbi__get32le(s); + info->hsz = hsz = stbi__get32le(s); + info->mr = info->mg = info->mb = info->ma = 0; + + if (hsz != 12 && hsz != 40 && hsz != 56 && hsz != 108 && hsz != 124) return stbi__errpuc("unknown BMP", "BMP type not supported: unknown"); + if (hsz == 12) { + s->img_x = stbi__get16le(s); + s->img_y = stbi__get16le(s); + } else { + s->img_x = stbi__get32le(s); + s->img_y = stbi__get32le(s); + } + if (stbi__get16le(s) != 1) return stbi__errpuc("bad BMP", "bad BMP"); + info->bpp = stbi__get16le(s); + if (hsz != 12) { + int compress = stbi__get32le(s); + if (compress == 1 || compress == 2) return stbi__errpuc("BMP RLE", "BMP type not supported: RLE"); + stbi__get32le(s); // discard sizeof + stbi__get32le(s); // discard hres + stbi__get32le(s); // discard vres + stbi__get32le(s); // discard colorsused + stbi__get32le(s); // discard max important + if (hsz == 40 || hsz == 56) { + if (hsz == 56) { + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + } + if (info->bpp == 16 || info->bpp == 32) { + if (compress == 0) { + if (info->bpp == 32) { + info->mr = 0xffu << 16; + info->mg = 0xffu << 8; + info->mb = 0xffu << 0; + info->ma = 0xffu << 24; + info->all_a = 0; // if all_a is 0 at end, then we loaded alpha channel but it was all 0 + } else { + info->mr = 31u << 10; + info->mg = 31u << 5; + info->mb = 31u << 0; + } + } else if (compress == 3) { + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + // not documented, but generated by photoshop and handled by mspaint + if (info->mr == info->mg && info->mg == info->mb) { + // ?!?!? + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else { + int i; + if (hsz != 108 && hsz != 124) + return stbi__errpuc("bad BMP", "bad BMP"); + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + info->ma = stbi__get32le(s); + stbi__get32le(s); // discard color space + for (i=0; i < 12; ++i) + stbi__get32le(s); // discard color space parameters + if (hsz == 124) { + stbi__get32le(s); // discard rendering intent + stbi__get32le(s); // discard offset of profile data + stbi__get32le(s); // discard size of profile data + stbi__get32le(s); // discard reserved + } + } + } + return (void *) 1; +} + + +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + unsigned int mr=0,mg=0,mb=0,ma=0, all_a; + stbi_uc pal[256][4]; + int psize=0,i,j,width; + int flip_vertically, pad, target; + stbi__bmp_data info; + STBI_NOTUSED(ri); + + info.all_a = 255; + if (stbi__bmp_parse_header(s, &info) == NULL) + return NULL; // error code already set + + flip_vertically = ((int) s->img_y) > 0; + s->img_y = abs((int) s->img_y); + + mr = info.mr; + mg = info.mg; + mb = info.mb; + ma = info.ma; + all_a = info.all_a; + + if (info.hsz == 12) { + if (info.bpp < 24) + psize = (info.offset - 14 - 24) / 3; + } else { + if (info.bpp < 16) + psize = (info.offset - 14 - info.hsz) >> 2; + } + + s->img_n = ma ? 4 : 3; + if (req_comp && req_comp >= 3) // we can directly decode 3 or 4 + target = req_comp; + else + target = s->img_n; // if they want monochrome, we'll post-convert + + // sanity-check size + if (!stbi__mad3sizes_valid(target, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "Corrupt BMP"); + + out = (stbi_uc *) stbi__malloc_mad3(target, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + if (info.bpp < 16) { + int z=0; + if (psize == 0 || psize > 256) { STBI_FREE(out); return stbi__errpuc("invalid", "Corrupt BMP"); } + for (i=0; i < psize; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + if (info.hsz != 12) stbi__get8(s); + pal[i][3] = 255; + } + stbi__skip(s, info.offset - 14 - info.hsz - psize * (info.hsz == 12 ? 3 : 4)); + if (info.bpp == 1) width = (s->img_x + 7) >> 3; + else if (info.bpp == 4) width = (s->img_x + 1) >> 1; + else if (info.bpp == 8) width = s->img_x; + else { STBI_FREE(out); return stbi__errpuc("bad bpp", "Corrupt BMP"); } + pad = (-width)&3; + if (info.bpp == 1) { + for (j=0; j < (int) s->img_y; ++j) { + int bit_offset = 7, v = stbi__get8(s); + for (i=0; i < (int) s->img_x; ++i) { + int color = (v>>bit_offset)&0x1; + out[z++] = pal[color][0]; + out[z++] = pal[color][1]; + out[z++] = pal[color][2]; + if((--bit_offset) < 0) { + bit_offset = 7; + v = stbi__get8(s); + } + } + stbi__skip(s, pad); + } + } else { + for (j=0; j < (int) s->img_y; ++j) { + for (i=0; i < (int) s->img_x; i += 2) { + int v=stbi__get8(s),v2=0; + if (info.bpp == 4) { + v2 = v & 15; + v >>= 4; + } + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + v = (info.bpp == 8) ? stbi__get8(s) : v2; + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + } + stbi__skip(s, pad); + } + } + } else { + int rshift=0,gshift=0,bshift=0,ashift=0,rcount=0,gcount=0,bcount=0,acount=0; + int z = 0; + int easy=0; + stbi__skip(s, info.offset - 14 - info.hsz); + if (info.bpp == 24) width = 3 * s->img_x; + else if (info.bpp == 16) width = 2*s->img_x; + else /* bpp = 32 and pad = 0 */ width=0; + pad = (-width) & 3; + if (info.bpp == 24) { + easy = 1; + } else if (info.bpp == 32) { + if (mb == 0xff && mg == 0xff00 && mr == 0x00ff0000 && ma == 0xff000000) + easy = 2; + } + if (!easy) { + if (!mr || !mg || !mb) { STBI_FREE(out); return stbi__errpuc("bad masks", "Corrupt BMP"); } + // right shift amt to put high bit in position #7 + rshift = stbi__high_bit(mr)-7; rcount = stbi__bitcount(mr); + gshift = stbi__high_bit(mg)-7; gcount = stbi__bitcount(mg); + bshift = stbi__high_bit(mb)-7; bcount = stbi__bitcount(mb); + ashift = stbi__high_bit(ma)-7; acount = stbi__bitcount(ma); + } + for (j=0; j < (int) s->img_y; ++j) { + if (easy) { + for (i=0; i < (int) s->img_x; ++i) { + unsigned char a; + out[z+2] = stbi__get8(s); + out[z+1] = stbi__get8(s); + out[z+0] = stbi__get8(s); + z += 3; + a = (easy == 2 ? stbi__get8(s) : 255); + all_a |= a; + if (target == 4) out[z++] = a; + } + } else { + int bpp = info.bpp; + for (i=0; i < (int) s->img_x; ++i) { + stbi__uint32 v = (bpp == 16 ? (stbi__uint32) stbi__get16le(s) : stbi__get32le(s)); + unsigned int a; + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mr, rshift, rcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mg, gshift, gcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mb, bshift, bcount)); + a = (ma ? stbi__shiftsigned(v & ma, ashift, acount) : 255); + all_a |= a; + if (target == 4) out[z++] = STBI__BYTECAST(a); + } + } + stbi__skip(s, pad); + } + } + + // if alpha channel is all 0s, replace with all 255s + if (target == 4 && all_a == 0) + for (i=4*s->img_x*s->img_y-1; i >= 0; i -= 4) + out[i] = 255; + + if (flip_vertically) { + stbi_uc t; + for (j=0; j < (int) s->img_y>>1; ++j) { + stbi_uc *p1 = out + j *s->img_x*target; + stbi_uc *p2 = out + (s->img_y-1-j)*s->img_x*target; + for (i=0; i < (int) s->img_x*target; ++i) { + t = p1[i], p1[i] = p2[i], p2[i] = t; + } + } + } + + if (req_comp && req_comp != target) { + out = stbi__convert_format(out, target, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + return out; +} +#endif + +// Targa Truevision - TGA +// by Jonathan Dummer +#ifndef STBI_NO_TGA +// returns STBI_rgb or whatever, 0 on error +static int stbi__tga_get_comp(int bits_per_pixel, int is_grey, int* is_rgb16) +{ + // only RGB or RGBA (incl. 16bit) or grey allowed + if (is_rgb16) *is_rgb16 = 0; + switch(bits_per_pixel) { + case 8: return STBI_grey; + case 16: if(is_grey) return STBI_grey_alpha; + // fallthrough + case 15: if(is_rgb16) *is_rgb16 = 1; + return STBI_rgb; + case 24: // fallthrough + case 32: return bits_per_pixel/8; + default: return 0; + } +} + +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp) +{ + int tga_w, tga_h, tga_comp, tga_image_type, tga_bits_per_pixel, tga_colormap_bpp; + int sz, tga_colormap_type; + stbi__get8(s); // discard Offset + tga_colormap_type = stbi__get8(s); // colormap type + if( tga_colormap_type > 1 ) { + stbi__rewind(s); + return 0; // only RGB or indexed allowed + } + tga_image_type = stbi__get8(s); // image type + if ( tga_colormap_type == 1 ) { // colormapped (paletted) image + if (tga_image_type != 1 && tga_image_type != 9) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); // skip index of first colormap entry and number of entries + sz = stbi__get8(s); // check bits per palette color entry + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); // skip image x and y origin + tga_colormap_bpp = sz; + } else { // "normal" image w/o colormap - only RGB or grey allowed, +/- RLE + if ( (tga_image_type != 2) && (tga_image_type != 3) && (tga_image_type != 10) && (tga_image_type != 11) ) { + stbi__rewind(s); + return 0; // only RGB or grey allowed, +/- RLE + } + stbi__skip(s,9); // skip colormap specification and image x/y origin + tga_colormap_bpp = 0; + } + tga_w = stbi__get16le(s); + if( tga_w < 1 ) { + stbi__rewind(s); + return 0; // test width + } + tga_h = stbi__get16le(s); + if( tga_h < 1 ) { + stbi__rewind(s); + return 0; // test height + } + tga_bits_per_pixel = stbi__get8(s); // bits per pixel + stbi__get8(s); // ignore alpha bits + if (tga_colormap_bpp != 0) { + if((tga_bits_per_pixel != 8) && (tga_bits_per_pixel != 16)) { + // when using a colormap, tga_bits_per_pixel is the size of the indexes + // I don't think anything but 8 or 16bit indexes makes sense + stbi__rewind(s); + return 0; + } + tga_comp = stbi__tga_get_comp(tga_colormap_bpp, 0, NULL); + } else { + tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3) || (tga_image_type == 11), NULL); + } + if(!tga_comp) { + stbi__rewind(s); + return 0; + } + if (x) *x = tga_w; + if (y) *y = tga_h; + if (comp) *comp = tga_comp; + return 1; // seems to have passed everything +} + +static int stbi__tga_test(stbi__context *s) +{ + int res = 0; + int sz, tga_color_type; + stbi__get8(s); // discard Offset + tga_color_type = stbi__get8(s); // color type + if ( tga_color_type > 1 ) goto errorEnd; // only RGB or indexed allowed + sz = stbi__get8(s); // image type + if ( tga_color_type == 1 ) { // colormapped (paletted) image + if (sz != 1 && sz != 9) goto errorEnd; // colortype 1 demands image type 1 or 9 + stbi__skip(s,4); // skip index of first colormap entry and number of entries + sz = stbi__get8(s); // check bits per palette color entry + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + stbi__skip(s,4); // skip image x and y origin + } else { // "normal" image w/o colormap + if ( (sz != 2) && (sz != 3) && (sz != 10) && (sz != 11) ) goto errorEnd; // only RGB or grey allowed, +/- RLE + stbi__skip(s,9); // skip colormap specification and image x/y origin + } + if ( stbi__get16le(s) < 1 ) goto errorEnd; // test width + if ( stbi__get16le(s) < 1 ) goto errorEnd; // test height + sz = stbi__get8(s); // bits per pixel + if ( (tga_color_type == 1) && (sz != 8) && (sz != 16) ) goto errorEnd; // for colormapped images, bpp is size of an index + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + + res = 1; // if we got this far, everything's good and we can return 1 instead of 0 + +errorEnd: + stbi__rewind(s); + return res; +} + +// read 16bit value and convert to 24bit RGB +static void stbi__tga_read_rgb16(stbi__context *s, stbi_uc* out) +{ + stbi__uint16 px = (stbi__uint16)stbi__get16le(s); + stbi__uint16 fiveBitMask = 31; + // we have 3 channels with 5bits each + int r = (px >> 10) & fiveBitMask; + int g = (px >> 5) & fiveBitMask; + int b = px & fiveBitMask; + // Note that this saves the data in RGB(A) order, so it doesn't need to be swapped later + out[0] = (stbi_uc)((r * 255)/31); + out[1] = (stbi_uc)((g * 255)/31); + out[2] = (stbi_uc)((b * 255)/31); + + // some people claim that the most significant bit might be used for alpha + // (possibly if an alpha-bit is set in the "image descriptor byte") + // but that only made 16bit test images completely translucent.. + // so let's treat all 15 and 16bit TGAs as RGB with no alpha. +} + +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + // read in the TGA header stuff + int tga_offset = stbi__get8(s); + int tga_indexed = stbi__get8(s); + int tga_image_type = stbi__get8(s); + int tga_is_RLE = 0; + int tga_palette_start = stbi__get16le(s); + int tga_palette_len = stbi__get16le(s); + int tga_palette_bits = stbi__get8(s); + int tga_x_origin = stbi__get16le(s); + int tga_y_origin = stbi__get16le(s); + int tga_width = stbi__get16le(s); + int tga_height = stbi__get16le(s); + int tga_bits_per_pixel = stbi__get8(s); + int tga_comp, tga_rgb16=0; + int tga_inverted = stbi__get8(s); + // int tga_alpha_bits = tga_inverted & 15; // the 4 lowest bits - unused (useless?) + // image data + unsigned char *tga_data; + unsigned char *tga_palette = NULL; + int i, j; + unsigned char raw_data[4] = {0}; + int RLE_count = 0; + int RLE_repeating = 0; + int read_next_pixel = 1; + STBI_NOTUSED(ri); + + // do a tiny bit of precessing + if ( tga_image_type >= 8 ) + { + tga_image_type -= 8; + tga_is_RLE = 1; + } + tga_inverted = 1 - ((tga_inverted >> 5) & 1); + + // If I'm paletted, then I'll use the number of bits from the palette + if ( tga_indexed ) tga_comp = stbi__tga_get_comp(tga_palette_bits, 0, &tga_rgb16); + else tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3), &tga_rgb16); + + if(!tga_comp) // shouldn't really happen, stbi__tga_test() should have ensured basic consistency + return stbi__errpuc("bad format", "Can't find out TGA pixelformat"); + + // tga info + *x = tga_width; + *y = tga_height; + if (comp) *comp = tga_comp; + + if (!stbi__mad3sizes_valid(tga_width, tga_height, tga_comp, 0)) + return stbi__errpuc("too large", "Corrupt TGA"); + + tga_data = (unsigned char*)stbi__malloc_mad3(tga_width, tga_height, tga_comp, 0); + if (!tga_data) return stbi__errpuc("outofmem", "Out of memory"); + + // skip to the data's starting position (offset usually = 0) + stbi__skip(s, tga_offset ); + + if ( !tga_indexed && !tga_is_RLE && !tga_rgb16 ) { + for (i=0; i < tga_height; ++i) { + int row = tga_inverted ? tga_height -i - 1 : i; + stbi_uc *tga_row = tga_data + row*tga_width*tga_comp; + stbi__getn(s, tga_row, tga_width * tga_comp); + } + } else { + // do I need to load a palette? + if ( tga_indexed) + { + // any data to skip? (offset usually = 0) + stbi__skip(s, tga_palette_start ); + // load the palette + tga_palette = (unsigned char*)stbi__malloc_mad2(tga_palette_len, tga_comp, 0); + if (!tga_palette) { + STBI_FREE(tga_data); + return stbi__errpuc("outofmem", "Out of memory"); + } + if (tga_rgb16) { + stbi_uc *pal_entry = tga_palette; + STBI_ASSERT(tga_comp == STBI_rgb); + for (i=0; i < tga_palette_len; ++i) { + stbi__tga_read_rgb16(s, pal_entry); + pal_entry += tga_comp; + } + } else if (!stbi__getn(s, tga_palette, tga_palette_len * tga_comp)) { + STBI_FREE(tga_data); + STBI_FREE(tga_palette); + return stbi__errpuc("bad palette", "Corrupt TGA"); + } + } + // load the data + for (i=0; i < tga_width * tga_height; ++i) + { + // if I'm in RLE mode, do I need to get a RLE stbi__pngchunk? + if ( tga_is_RLE ) + { + if ( RLE_count == 0 ) + { + // yep, get the next byte as a RLE command + int RLE_cmd = stbi__get8(s); + RLE_count = 1 + (RLE_cmd & 127); + RLE_repeating = RLE_cmd >> 7; + read_next_pixel = 1; + } else if ( !RLE_repeating ) + { + read_next_pixel = 1; + } + } else + { + read_next_pixel = 1; + } + // OK, if I need to read a pixel, do it now + if ( read_next_pixel ) + { + // load however much data we did have + if ( tga_indexed ) + { + // read in index, then perform the lookup + int pal_idx = (tga_bits_per_pixel == 8) ? stbi__get8(s) : stbi__get16le(s); + if ( pal_idx >= tga_palette_len ) { + // invalid index + pal_idx = 0; + } + pal_idx *= tga_comp; + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = tga_palette[pal_idx+j]; + } + } else if(tga_rgb16) { + STBI_ASSERT(tga_comp == STBI_rgb); + stbi__tga_read_rgb16(s, raw_data); + } else { + // read in the data raw + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = stbi__get8(s); + } + } + // clear the reading flag for the next pixel + read_next_pixel = 0; + } // end of reading a pixel + + // copy data + for (j = 0; j < tga_comp; ++j) + tga_data[i*tga_comp+j] = raw_data[j]; + + // in case we're in RLE mode, keep counting down + --RLE_count; + } + // do I need to invert the image? + if ( tga_inverted ) + { + for (j = 0; j*2 < tga_height; ++j) + { + int index1 = j * tga_width * tga_comp; + int index2 = (tga_height - 1 - j) * tga_width * tga_comp; + for (i = tga_width * tga_comp; i > 0; --i) + { + unsigned char temp = tga_data[index1]; + tga_data[index1] = tga_data[index2]; + tga_data[index2] = temp; + ++index1; + ++index2; + } + } + } + // clear my palette, if I had one + if ( tga_palette != NULL ) + { + STBI_FREE( tga_palette ); + } + } + + // swap RGB - if the source data was RGB16, it already is in the right order + if (tga_comp >= 3 && !tga_rgb16) + { + unsigned char* tga_pixel = tga_data; + for (i=0; i < tga_width * tga_height; ++i) + { + unsigned char temp = tga_pixel[0]; + tga_pixel[0] = tga_pixel[2]; + tga_pixel[2] = temp; + tga_pixel += tga_comp; + } + } + + // convert to target component count + if (req_comp && req_comp != tga_comp) + tga_data = stbi__convert_format(tga_data, tga_comp, req_comp, tga_width, tga_height); + + // the things I do to get rid of an error message, and yet keep + // Microsoft's C compilers happy... [8^( + tga_palette_start = tga_palette_len = tga_palette_bits = + tga_x_origin = tga_y_origin = 0; + // OK, done + return tga_data; +} +#endif + +// ************************************************************************************************* +// Photoshop PSD loader -- PD by Thatcher Ulrich, integration by Nicolas Schulz, tweaked by STB + +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s) +{ + int r = (stbi__get32be(s) == 0x38425053); + stbi__rewind(s); + return r; +} + +static int stbi__psd_decode_rle(stbi__context *s, stbi_uc *p, int pixelCount) +{ + int count, nleft, len; + + count = 0; + while ((nleft = pixelCount - count) > 0) { + len = stbi__get8(s); + if (len == 128) { + // No-op. + } else if (len < 128) { + // Copy next len+1 bytes literally. + len++; + if (len > nleft) return 0; // corrupt data + count += len; + while (len) { + *p = stbi__get8(s); + p += 4; + len--; + } + } else if (len > 128) { + stbi_uc val; + // Next -len+1 bytes in the dest are replicated from next source byte. + // (Interpret len as a negative 8-bit int.) + len = 257 - len; + if (len > nleft) return 0; // corrupt data + val = stbi__get8(s); + count += len; + while (len) { + *p = val; + p += 4; + len--; + } + } + } + + return 1; +} + +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + int pixelCount; + int channelCount, compression; + int channel, i; + int bitdepth; + int w,h; + stbi_uc *out; + STBI_NOTUSED(ri); + + // Check identifier + if (stbi__get32be(s) != 0x38425053) // "8BPS" + return stbi__errpuc("not PSD", "Corrupt PSD image"); + + // Check file type version. + if (stbi__get16be(s) != 1) + return stbi__errpuc("wrong version", "Unsupported version of PSD image"); + + // Skip 6 reserved bytes. + stbi__skip(s, 6 ); + + // Read the number of channels (R, G, B, A, etc). + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) + return stbi__errpuc("wrong channel count", "Unsupported number of channels in PSD image"); + + // Read the rows and columns of the image. + h = stbi__get32be(s); + w = stbi__get32be(s); + + // Make sure the depth is 8 bits. + bitdepth = stbi__get16be(s); + if (bitdepth != 8 && bitdepth != 16) + return stbi__errpuc("unsupported bit depth", "PSD bit depth is not 8 or 16 bit"); + + // Make sure the color mode is RGB. + // Valid options are: + // 0: Bitmap + // 1: Grayscale + // 2: Indexed color + // 3: RGB color + // 4: CMYK color + // 7: Multichannel + // 8: Duotone + // 9: Lab color + if (stbi__get16be(s) != 3) + return stbi__errpuc("wrong color format", "PSD is not in RGB color format"); + + // Skip the Mode Data. (It's the palette for indexed color; other info for other modes.) + stbi__skip(s,stbi__get32be(s) ); + + // Skip the image resources. (resolution, pen tool paths, etc) + stbi__skip(s, stbi__get32be(s) ); + + // Skip the reserved data. + stbi__skip(s, stbi__get32be(s) ); + + // Find out if the data is compressed. + // Known values: + // 0: no compression + // 1: RLE compressed + compression = stbi__get16be(s); + if (compression > 1) + return stbi__errpuc("bad compression", "PSD has an unknown compression format"); + + // Check size + if (!stbi__mad3sizes_valid(4, w, h, 0)) + return stbi__errpuc("too large", "Corrupt PSD"); + + // Create the destination image. + + if (!compression && bitdepth == 16 && bpc == 16) { + out = (stbi_uc *) stbi__malloc_mad3(8, w, h, 0); + ri->bits_per_channel = 16; + } else + out = (stbi_uc *) stbi__malloc(4 * w*h); + + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + pixelCount = w*h; + + // Initialize the data to zero. + //memset( out, 0, pixelCount * 4 ); + + // Finally, the image data. + if (compression) { + // RLE as used by .PSD and .TIFF + // Loop until you get the number of unpacked bytes you are expecting: + // Read the next source byte into n. + // If n is between 0 and 127 inclusive, copy the next n+1 bytes literally. + // Else if n is between -127 and -1 inclusive, copy the next byte -n+1 times. + // Else if n is 128, noop. + // Endloop + + // The RLE-compressed data is preceeded by a 2-byte data count for each row in the data, + // which we're going to just skip. + stbi__skip(s, h * channelCount * 2 ); + + // Read the RLE data by channel. + for (channel = 0; channel < 4; channel++) { + stbi_uc *p; + + p = out+channel; + if (channel >= channelCount) { + // Fill this channel with default data. + for (i = 0; i < pixelCount; i++, p += 4) + *p = (channel == 3 ? 255 : 0); + } else { + // Read the RLE data. + if (!stbi__psd_decode_rle(s, p, pixelCount)) { + STBI_FREE(out); + return stbi__errpuc("corrupt", "bad RLE data"); + } + } + } + + } else { + // We're at the raw image data. It's each channel in order (Red, Green, Blue, Alpha, ...) + // where each channel consists of an 8-bit (or 16-bit) value for each pixel in the image. + + // Read the data by channel. + for (channel = 0; channel < 4; channel++) { + if (channel >= channelCount) { + // Fill this channel with default data. + if (bitdepth == 16 && bpc == 16) { + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + stbi__uint16 val = channel == 3 ? 65535 : 0; + for (i = 0; i < pixelCount; i++, q += 4) + *q = val; + } else { + stbi_uc *p = out+channel; + stbi_uc val = channel == 3 ? 255 : 0; + for (i = 0; i < pixelCount; i++, p += 4) + *p = val; + } + } else { + if (ri->bits_per_channel == 16) { // output bpc + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + for (i = 0; i < pixelCount; i++, q += 4) + *q = (stbi__uint16) stbi__get16be(s); + } else { + stbi_uc *p = out+channel; + if (bitdepth == 16) { // input bpc + for (i = 0; i < pixelCount; i++, p += 4) + *p = (stbi_uc) (stbi__get16be(s) >> 8); + } else { + for (i = 0; i < pixelCount; i++, p += 4) + *p = stbi__get8(s); + } + } + } + } + } + + // remove weird white matte from PSD + if (channelCount >= 4) { + if (ri->bits_per_channel == 16) { + for (i=0; i < w*h; ++i) { + stbi__uint16 *pixel = (stbi__uint16 *) out + 4*i; + if (pixel[3] != 0 && pixel[3] != 65535) { + float a = pixel[3] / 65535.0f; + float ra = 1.0f / a; + float inv_a = 65535.0f * (1 - ra); + pixel[0] = (stbi__uint16) (pixel[0]*ra + inv_a); + pixel[1] = (stbi__uint16) (pixel[1]*ra + inv_a); + pixel[2] = (stbi__uint16) (pixel[2]*ra + inv_a); + } + } + } else { + for (i=0; i < w*h; ++i) { + unsigned char *pixel = out + 4*i; + if (pixel[3] != 0 && pixel[3] != 255) { + float a = pixel[3] / 255.0f; + float ra = 1.0f / a; + float inv_a = 255.0f * (1 - ra); + pixel[0] = (unsigned char) (pixel[0]*ra + inv_a); + pixel[1] = (unsigned char) (pixel[1]*ra + inv_a); + pixel[2] = (unsigned char) (pixel[2]*ra + inv_a); + } + } + } + } + + // convert to desired output format + if (req_comp && req_comp != 4) { + if (ri->bits_per_channel == 16) + out = (stbi_uc *) stbi__convert_format16((stbi__uint16 *) out, 4, req_comp, w, h); + else + out = stbi__convert_format(out, 4, req_comp, w, h); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + + if (comp) *comp = 4; + *y = h; + *x = w; + + return out; +} +#endif + +// ************************************************************************************************* +// Softimage PIC loader +// by Tom Seddon +// +// See http://softimage.wiki.softimage.com/index.php/INFO:_PIC_file_format +// See http://ozviz.wasp.uwa.edu.au/~pbourke/dataformats/softimagepic/ + +#ifndef STBI_NO_PIC +static int stbi__pic_is4(stbi__context *s,const char *str) +{ + int i; + for (i=0; i<4; ++i) + if (stbi__get8(s) != (stbi_uc)str[i]) + return 0; + + return 1; +} + +static int stbi__pic_test_core(stbi__context *s) +{ + int i; + + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) + return 0; + + for(i=0;i<84;++i) + stbi__get8(s); + + if (!stbi__pic_is4(s,"PICT")) + return 0; + + return 1; +} + +typedef struct +{ + stbi_uc size,type,channel; +} stbi__pic_packet; + +static stbi_uc *stbi__readval(stbi__context *s, int channel, stbi_uc *dest) +{ + int mask=0x80, i; + + for (i=0; i<4; ++i, mask>>=1) { + if (channel & mask) { + if (stbi__at_eof(s)) return stbi__errpuc("bad file","PIC file too short"); + dest[i]=stbi__get8(s); + } + } + + return dest; +} + +static void stbi__copyval(int channel,stbi_uc *dest,const stbi_uc *src) +{ + int mask=0x80,i; + + for (i=0;i<4; ++i, mask>>=1) + if (channel&mask) + dest[i]=src[i]; +} + +static stbi_uc *stbi__pic_load_core(stbi__context *s,int width,int height,int *comp, stbi_uc *result) +{ + int act_comp=0,num_packets=0,y,chained; + stbi__pic_packet packets[10]; + + // this will (should...) cater for even some bizarre stuff like having data + // for the same channel in multiple packets. + do { + stbi__pic_packet *packet; + + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return stbi__errpuc("bad format","too many packets"); + + packet = &packets[num_packets++]; + + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + + act_comp |= packet->channel; + + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (reading packets)"); + if (packet->size != 8) return stbi__errpuc("bad format","packet isn't 8bpp"); + } while (chained); + + *comp = (act_comp & 0x10 ? 4 : 3); // has alpha channel? + + for(y=0; ytype) { + default: + return stbi__errpuc("bad format","packet has bad compression type"); + + case 0: {//uncompressed + int x; + + for(x=0;xchannel,dest)) + return 0; + break; + } + + case 1://Pure RLE + { + int left=width, i; + + while (left>0) { + stbi_uc count,value[4]; + + count=stbi__get8(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pure read count)"); + + if (count > left) + count = (stbi_uc) left; + + if (!stbi__readval(s,packet->channel,value)) return 0; + + for(i=0; ichannel,dest,value); + left -= count; + } + } + break; + + case 2: {//Mixed RLE + int left=width; + while (left>0) { + int count = stbi__get8(s), i; + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (mixed read count)"); + + if (count >= 128) { // Repeated + stbi_uc value[4]; + + if (count==128) + count = stbi__get16be(s); + else + count -= 127; + if (count > left) + return stbi__errpuc("bad file","scanline overrun"); + + if (!stbi__readval(s,packet->channel,value)) + return 0; + + for(i=0;ichannel,dest,value); + } else { // Raw + ++count; + if (count>left) return stbi__errpuc("bad file","scanline overrun"); + + for(i=0;ichannel,dest)) + return 0; + } + left-=count; + } + break; + } + } + } + } + + return result; +} + +static void *stbi__pic_load(stbi__context *s,int *px,int *py,int *comp,int req_comp, stbi__result_info *ri) +{ + stbi_uc *result; + int i, x,y, internal_comp; + STBI_NOTUSED(ri); + + if (!comp) comp = &internal_comp; + + for (i=0; i<92; ++i) + stbi__get8(s); + + x = stbi__get16be(s); + y = stbi__get16be(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pic header)"); + if (!stbi__mad3sizes_valid(x, y, 4, 0)) return stbi__errpuc("too large", "PIC image too large to decode"); + + stbi__get32be(s); //skip `ratio' + stbi__get16be(s); //skip `fields' + stbi__get16be(s); //skip `pad' + + // intermediate buffer is RGBA + result = (stbi_uc *) stbi__malloc_mad3(x, y, 4, 0); + memset(result, 0xff, x*y*4); + + if (!stbi__pic_load_core(s,x,y,comp, result)) { + STBI_FREE(result); + result=0; + } + *px = x; + *py = y; + if (req_comp == 0) req_comp = *comp; + result=stbi__convert_format(result,4,req_comp,x,y); + + return result; +} + +static int stbi__pic_test(stbi__context *s) +{ + int r = stbi__pic_test_core(s); + stbi__rewind(s); + return r; +} +#endif + +// ************************************************************************************************* +// GIF loader -- public domain by Jean-Marc Lienher -- simplified/shrunk by stb + +#ifndef STBI_NO_GIF +typedef struct +{ + stbi__int16 prefix; + stbi_uc first; + stbi_uc suffix; +} stbi__gif_lzw; + +typedef struct +{ + int w,h; + stbi_uc *out; // output buffer (always 4 components) + stbi_uc *background; // The current "background" as far as a gif is concerned + stbi_uc *history; + int flags, bgindex, ratio, transparent, eflags; + stbi_uc pal[256][4]; + stbi_uc lpal[256][4]; + stbi__gif_lzw codes[8192]; + stbi_uc *color_table; + int parse, step; + int lflags; + int start_x, start_y; + int max_x, max_y; + int cur_x, cur_y; + int line_size; + int delay; +} stbi__gif; + +static int stbi__gif_test_raw(stbi__context *s) +{ + int sz; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') return 0; + sz = stbi__get8(s); + if (sz != '9' && sz != '7') return 0; + if (stbi__get8(s) != 'a') return 0; + return 1; +} + +static int stbi__gif_test(stbi__context *s) +{ + int r = stbi__gif_test_raw(s); + stbi__rewind(s); + return r; +} + +static void stbi__gif_parse_colortable(stbi__context *s, stbi_uc pal[256][4], int num_entries, int transp) +{ + int i; + for (i=0; i < num_entries; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + pal[i][3] = transp == i ? 0 : 255; + } +} + +static int stbi__gif_header(stbi__context *s, stbi__gif *g, int *comp, int is_info) +{ + stbi_uc version; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') + return stbi__err("not GIF", "Corrupt GIF"); + + version = stbi__get8(s); + if (version != '7' && version != '9') return stbi__err("not GIF", "Corrupt GIF"); + if (stbi__get8(s) != 'a') return stbi__err("not GIF", "Corrupt GIF"); + + stbi__g_failure_reason = ""; + g->w = stbi__get16le(s); + g->h = stbi__get16le(s); + g->flags = stbi__get8(s); + g->bgindex = stbi__get8(s); + g->ratio = stbi__get8(s); + g->transparent = -1; + + if (comp != 0) *comp = 4; // can't actually tell whether it's 3 or 4 until we parse the comments + + if (is_info) return 1; + + if (g->flags & 0x80) + stbi__gif_parse_colortable(s,g->pal, 2 << (g->flags & 7), -1); + + return 1; +} + +static int stbi__gif_info_raw(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__gif* g = (stbi__gif*) stbi__malloc(sizeof(stbi__gif)); + if (!stbi__gif_header(s, g, comp, 1)) { + STBI_FREE(g); + stbi__rewind( s ); + return 0; + } + if (x) *x = g->w; + if (y) *y = g->h; + STBI_FREE(g); + return 1; +} + +static void stbi__out_gif_code(stbi__gif *g, stbi__uint16 code) +{ + stbi_uc *p, *c; + int idx; + + // recurse to decode the prefixes, since the linked-list is backwards, + // and working backwards through an interleaved image would be nasty + if (g->codes[code].prefix >= 0) + stbi__out_gif_code(g, g->codes[code].prefix); + + if (g->cur_y >= g->max_y) return; + + idx = g->cur_x + g->cur_y; + p = &g->out[idx]; + g->history[idx / 4] = 1; + + c = &g->color_table[g->codes[code].suffix * 4]; + if (c[3] > 128) { // don't render transparent pixels; + p[0] = c[2]; + p[1] = c[1]; + p[2] = c[0]; + p[3] = c[3]; + } + g->cur_x += 4; + + if (g->cur_x >= g->max_x) { + g->cur_x = g->start_x; + g->cur_y += g->step; + + while (g->cur_y >= g->max_y && g->parse > 0) { + g->step = (1 << g->parse) * g->line_size; + g->cur_y = g->start_y + (g->step >> 1); + --g->parse; + } + } +} + +static stbi_uc *stbi__process_gif_raster(stbi__context *s, stbi__gif *g) +{ + stbi_uc lzw_cs; + stbi__int32 len, init_code; + stbi__uint32 first; + stbi__int32 codesize, codemask, avail, oldcode, bits, valid_bits, clear; + stbi__gif_lzw *p; + + lzw_cs = stbi__get8(s); + if (lzw_cs > 12) return NULL; + clear = 1 << lzw_cs; + first = 1; + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + bits = 0; + valid_bits = 0; + for (init_code = 0; init_code < clear; init_code++) { + g->codes[init_code].prefix = -1; + g->codes[init_code].first = (stbi_uc) init_code; + g->codes[init_code].suffix = (stbi_uc) init_code; + } + + // support no starting clear code + avail = clear+2; + oldcode = -1; + + len = 0; + for(;;) { + if (valid_bits < codesize) { + if (len == 0) { + len = stbi__get8(s); // start new block + if (len == 0) + return g->out; + } + --len; + bits |= (stbi__int32) stbi__get8(s) << valid_bits; + valid_bits += 8; + } else { + stbi__int32 code = bits & codemask; + bits >>= codesize; + valid_bits -= codesize; + // @OPTIMIZE: is there some way we can accelerate the non-clear path? + if (code == clear) { // clear code + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + avail = clear + 2; + oldcode = -1; + first = 0; + } else if (code == clear + 1) { // end of stream code + stbi__skip(s, len); + while ((len = stbi__get8(s)) > 0) + stbi__skip(s,len); + return g->out; + } else if (code <= avail) { + if (first) { + return stbi__errpuc("no clear code", "Corrupt GIF"); + } + + if (oldcode >= 0) { + p = &g->codes[avail++]; + if (avail > 8192) { + return stbi__errpuc("too many codes", "Corrupt GIF"); + } + + p->prefix = (stbi__int16) oldcode; + p->first = g->codes[oldcode].first; + p->suffix = (code == avail) ? p->first : g->codes[code].first; + } else if (code == avail) + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + + stbi__out_gif_code(g, (stbi__uint16) code); + + if ((avail & codemask) == 0 && avail <= 0x0FFF) { + codesize++; + codemask = (1 << codesize) - 1; + } + + oldcode = code; + } else { + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + } + } + } +} + +// this function is designed to support animated gifs, although stb_image doesn't support it +// two back is the image from two frames ago, used for a very specific disposal format +static stbi_uc *stbi__gif_load_next(stbi__context *s, stbi__gif *g, int *comp, int req_comp, stbi_uc *two_back) +{ + int dispose; + int first_frame; + int pi; + int pcount; + + // on first frame, any non-written pixels get the background colour (non-transparent) + first_frame = 0; + if (g->out == 0) { + if (!stbi__gif_header(s, g, comp,0)) return 0; // stbi__g_failure_reason set by stbi__gif_header + g->out = (stbi_uc *) stbi__malloc(4 * g->w * g->h); + g->background = (stbi_uc *) stbi__malloc(4 * g->w * g->h); + g->history = (stbi_uc *) stbi__malloc(g->w * g->h); + if (g->out == 0) return stbi__errpuc("outofmem", "Out of memory"); + + // image is treated as "tranparent" at the start - ie, nothing overwrites the current background; + // background colour is only used for pixels that are not rendered first frame, after that "background" + // color refers to teh color that was there the previous frame. + memset( g->out, 0x00, 4 * g->w * g->h ); + memset( g->background, 0x00, 4 * g->w * g->h ); // state of the background (starts transparent) + memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame + first_frame = 1; + } else { + // second frame - how do we dispoase of the previous one? + dispose = (g->eflags & 0x1C) >> 2; + pcount = g->w * g->h; + + if ((dispose == 3) && (two_back == 0)) { + dispose = 2; // if I don't have an image to revert back to, default to the old background + } + + if (dispose == 3) { // use previous graphic + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &two_back[pi * 4], 4 ); + } + } + } else if (dispose == 2) { + // restore what was changed last frame to background before that frame; + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &g->background[pi * 4], 4 ); + } + } + } else { + // This is a non-disposal case eithe way, so just + // leave the pixels as is, and they will become the new background + // 1: do not dispose + // 0: not specified. + } + + // background is what out is after the undoing of the previou frame; + memcpy( g->background, g->out, 4 * g->w * g->h ); + } + + // clear my history; + memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame + + for (;;) { + int tag = stbi__get8(s); + switch (tag) { + case 0x2C: /* Image Descriptor */ + { + stbi__int32 x, y, w, h; + stbi_uc *o; + + x = stbi__get16le(s); + y = stbi__get16le(s); + w = stbi__get16le(s); + h = stbi__get16le(s); + if (((x + w) > (g->w)) || ((y + h) > (g->h))) + return stbi__errpuc("bad Image Descriptor", "Corrupt GIF"); + + g->line_size = g->w * 4; + g->start_x = x * 4; + g->start_y = y * g->line_size; + g->max_x = g->start_x + w * 4; + g->max_y = g->start_y + h * g->line_size; + g->cur_x = g->start_x; + g->cur_y = g->start_y; + + g->lflags = stbi__get8(s); + + if (g->lflags & 0x40) { + g->step = 8 * g->line_size; // first interlaced spacing + g->parse = 3; + } else { + g->step = g->line_size; + g->parse = 0; + } + + if (g->lflags & 0x80) { + stbi__gif_parse_colortable(s,g->lpal, 2 << (g->lflags & 7), g->eflags & 0x01 ? g->transparent : -1); + g->color_table = (stbi_uc *) g->lpal; + } else if (g->flags & 0x80) { + g->color_table = (stbi_uc *) g->pal; + } else + return stbi__errpuc("missing color table", "Corrupt GIF"); + + o = stbi__process_gif_raster(s, g); + if (o == NULL) return NULL; + + // if this was the first frame, + pcount = g->w * g->h; + if (first_frame && (g->bgindex > 0)) { + // if first frame, any pixel not drawn to gets the background color + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi] == 0) { + g->pal[g->bgindex][3] = 255; // just in case it was made transparent, undo that; It will be reset next frame if need be; + memcpy( &g->out[pi * 4], &g->pal[g->bgindex], 4 ); + } + } + } + + return o; + } + + case 0x21: // Comment Extension. + { + int len; + int ext = stbi__get8(s); + if (ext == 0xF9) { // Graphic Control Extension. + len = stbi__get8(s); + if (len == 4) { + g->eflags = stbi__get8(s); + g->delay = 10 * stbi__get16le(s); // delay - 1/100th of a second, saving as 1/1000ths. + + // unset old transparent + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 255; + } + if (g->eflags & 0x01) { + g->transparent = stbi__get8(s); + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 0; + } + } else { + // don't need transparent + stbi__skip(s, 1); + g->transparent = -1; + } + } else { + stbi__skip(s, len); + break; + } + } + while ((len = stbi__get8(s)) != 0) { + stbi__skip(s, len); + } + break; + } + + case 0x3B: // gif stream termination code + return (stbi_uc *) s; // using '1' causes warning on some compilers + + default: + return stbi__errpuc("unknown code", "Corrupt GIF"); + } + } +} + +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + if (stbi__gif_test(s)) { + int layers = 0; + stbi_uc *u = 0; + stbi_uc *out = 0; + stbi_uc *two_back = 0; + stbi__gif g; + int stride; + memset(&g, 0, sizeof(g)); + if (delays) { + *delays = 0; + } + + do { + u = stbi__gif_load_next(s, &g, comp, req_comp, two_back); + if (u == (stbi_uc *) s) u = 0; // end of animated gif marker + + if (u) { + *x = g.w; + *y = g.h; + ++layers; + stride = g.w * g.h * 4; + + if (out) { + out = (stbi_uc*) STBI_REALLOC( out, layers * stride ); + if (delays) { + *delays = (int*) STBI_REALLOC( *delays, sizeof(int) * layers ); + } + } else { + out = (stbi_uc*)stbi__malloc( layers * stride ); + if (delays) { + *delays = (int*) stbi__malloc( layers * sizeof(int) ); + } + } + memcpy( out + ((layers - 1) * stride), u, stride ); + if (layers >= 2) { + two_back = out - 2 * stride; + } + + if (delays) { + (*delays)[layers - 1U] = g.delay; + } + } + } while (u != 0); + + // free temp buffer; + STBI_FREE(g.out); + STBI_FREE(g.history); + STBI_FREE(g.background); + + // do the final conversion after loading everything; + if (req_comp && req_comp != 4) + out = stbi__convert_format(out, 4, req_comp, layers * g.w, g.h); + + *z = layers; + return out; + } else { + return stbi__errpuc("not GIF", "Image was not as a gif type."); + } +} + +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *u = 0; + stbi__gif g; + memset(&g, 0, sizeof(g)); + + u = stbi__gif_load_next(s, &g, comp, req_comp, 0); + if (u == (stbi_uc *) s) u = 0; // end of animated gif marker + if (u) { + *x = g.w; + *y = g.h; + + // moved conversion to after successful load so that the same + // can be done for multiple frames. + if (req_comp && req_comp != 4) + u = stbi__convert_format(u, 4, req_comp, g.w, g.h); + } + + // free buffers needed for multiple frame loading; + STBI_FREE(g.history); + STBI_FREE(g.background); + + return u; +} + +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp) +{ + return stbi__gif_info_raw(s,x,y,comp); +} +#endif + +// ************************************************************************************************* +// Radiance RGBE HDR loader +// originally by Nicolas Schulz +#ifndef STBI_NO_HDR +static int stbi__hdr_test_core(stbi__context *s, const char *signature) +{ + int i; + for (i=0; signature[i]; ++i) + if (stbi__get8(s) != signature[i]) + return 0; + stbi__rewind(s); + return 1; +} + +static int stbi__hdr_test(stbi__context* s) +{ + int r = stbi__hdr_test_core(s, "#?RADIANCE\n"); + stbi__rewind(s); + if(!r) { + r = stbi__hdr_test_core(s, "#?RGBE\n"); + stbi__rewind(s); + } + return r; +} + +#define STBI__HDR_BUFLEN 1024 +static char *stbi__hdr_gettoken(stbi__context *z, char *buffer) +{ + int len=0; + char c = '\0'; + + c = (char) stbi__get8(z); + + while (!stbi__at_eof(z) && c != '\n') { + buffer[len++] = c; + if (len == STBI__HDR_BUFLEN-1) { + // flush to end of line + while (!stbi__at_eof(z) && stbi__get8(z) != '\n') + ; + break; + } + c = (char) stbi__get8(z); + } + + buffer[len] = 0; + return buffer; +} + +static void stbi__hdr_convert(float *output, stbi_uc *input, int req_comp) +{ + if ( input[3] != 0 ) { + float f1; + // Exponent + f1 = (float) ldexp(1.0f, input[3] - (int)(128 + 8)); + if (req_comp <= 2) + output[0] = (input[0] + input[1] + input[2]) * f1 / 3; + else { + output[0] = input[0] * f1; + output[1] = input[1] * f1; + output[2] = input[2] * f1; + } + if (req_comp == 2) output[1] = 1; + if (req_comp == 4) output[3] = 1; + } else { + switch (req_comp) { + case 4: output[3] = 1; /* fallthrough */ + case 3: output[0] = output[1] = output[2] = 0; + break; + case 2: output[1] = 1; /* fallthrough */ + case 1: output[0] = 0; + break; + } + } +} + +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int width, height; + stbi_uc *scanline; + float *hdr_data; + int len; + unsigned char count, value; + int i, j, k, c1,c2, z; + const char *headerToken; + STBI_NOTUSED(ri); + + // Check identifier + headerToken = stbi__hdr_gettoken(s,buffer); + if (strcmp(headerToken, "#?RADIANCE") != 0 && strcmp(headerToken, "#?RGBE") != 0) + return stbi__errpf("not HDR", "Corrupt HDR image"); + + // Parse header + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + + if (!valid) return stbi__errpf("unsupported format", "Unsupported HDR format"); + + // Parse width and height + // can't use sscanf() if we're not using stdio! + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + height = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + width = (int) strtol(token, NULL, 10); + + *x = width; + *y = height; + + if (comp) *comp = 3; + if (req_comp == 0) req_comp = 3; + + if (!stbi__mad4sizes_valid(width, height, req_comp, sizeof(float), 0)) + return stbi__errpf("too large", "HDR image is too large"); + + // Read data + hdr_data = (float *) stbi__malloc_mad4(width, height, req_comp, sizeof(float), 0); + if (!hdr_data) + return stbi__errpf("outofmem", "Out of memory"); + + // Load image data + // image data is stored as some number of sca + if ( width < 8 || width >= 32768) { + // Read flat data + for (j=0; j < height; ++j) { + for (i=0; i < width; ++i) { + stbi_uc rgbe[4]; + main_decode_loop: + stbi__getn(s, rgbe, 4); + stbi__hdr_convert(hdr_data + j * width * req_comp + i * req_comp, rgbe, req_comp); + } + } + } else { + // Read RLE-encoded data + scanline = NULL; + + for (j = 0; j < height; ++j) { + c1 = stbi__get8(s); + c2 = stbi__get8(s); + len = stbi__get8(s); + if (c1 != 2 || c2 != 2 || (len & 0x80)) { + // not run-length encoded, so we have to actually use THIS data as a decoded + // pixel (note this can't be a valid pixel--one of RGB must be >= 128) + stbi_uc rgbe[4]; + rgbe[0] = (stbi_uc) c1; + rgbe[1] = (stbi_uc) c2; + rgbe[2] = (stbi_uc) len; + rgbe[3] = (stbi_uc) stbi__get8(s); + stbi__hdr_convert(hdr_data, rgbe, req_comp); + i = 1; + j = 0; + STBI_FREE(scanline); + goto main_decode_loop; // yes, this makes no sense + } + len <<= 8; + len |= stbi__get8(s); + if (len != width) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("invalid decoded scanline length", "corrupt HDR"); } + if (scanline == NULL) { + scanline = (stbi_uc *) stbi__malloc_mad2(width, 4, 0); + if (!scanline) { + STBI_FREE(hdr_data); + return stbi__errpf("outofmem", "Out of memory"); + } + } + + for (k = 0; k < 4; ++k) { + int nleft; + i = 0; + while ((nleft = width - i) > 0) { + count = stbi__get8(s); + if (count > 128) { + // Run + value = stbi__get8(s); + count -= 128; + if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = value; + } else { + // Dump + if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = stbi__get8(s); + } + } + } + for (i=0; i < width; ++i) + stbi__hdr_convert(hdr_data+(j*width + i)*req_comp, scanline + i*4, req_comp); + } + if (scanline) + STBI_FREE(scanline); + } + + return hdr_data; +} + +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int dummy; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + if (stbi__hdr_test(s) == 0) { + stbi__rewind( s ); + return 0; + } + + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + + if (!valid) { + stbi__rewind( s ); + return 0; + } + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *y = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *x = (int) strtol(token, NULL, 10); + *comp = 3; + return 1; +} +#endif // STBI_NO_HDR + +#ifndef STBI_NO_BMP +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp) +{ + void *p; + stbi__bmp_data info; + + info.all_a = 255; + p = stbi__bmp_parse_header(s, &info); + stbi__rewind( s ); + if (p == NULL) + return 0; + if (x) *x = s->img_x; + if (y) *y = s->img_y; + if (comp) *comp = info.ma ? 4 : 3; + return 1; +} +#endif + +#ifndef STBI_NO_PSD +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp) +{ + int channelCount, dummy, depth; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + *y = stbi__get32be(s); + *x = stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 8 && depth != 16) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 3) { + stbi__rewind( s ); + return 0; + } + *comp = 4; + return 1; +} + +static int stbi__psd_is16(stbi__context *s) +{ + int channelCount, depth; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + (void) stbi__get32be(s); + (void) stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 16) { + stbi__rewind( s ); + return 0; + } + return 1; +} +#endif + +#ifndef STBI_NO_PIC +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp) +{ + int act_comp=0,num_packets=0,chained,dummy; + stbi__pic_packet packets[10]; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) { + stbi__rewind(s); + return 0; + } + + stbi__skip(s, 88); + + *x = stbi__get16be(s); + *y = stbi__get16be(s); + if (stbi__at_eof(s)) { + stbi__rewind( s); + return 0; + } + if ( (*x) != 0 && (1 << 28) / (*x) < (*y)) { + stbi__rewind( s ); + return 0; + } + + stbi__skip(s, 8); + + do { + stbi__pic_packet *packet; + + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return 0; + + packet = &packets[num_packets++]; + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + act_comp |= packet->channel; + + if (stbi__at_eof(s)) { + stbi__rewind( s ); + return 0; + } + if (packet->size != 8) { + stbi__rewind( s ); + return 0; + } + } while (chained); + + *comp = (act_comp & 0x10 ? 4 : 3); + + return 1; +} +#endif + +// ************************************************************************************************* +// Portable Gray Map and Portable Pixel Map loader +// by Ken Miller +// +// PGM: http://netpbm.sourceforge.net/doc/pgm.html +// PPM: http://netpbm.sourceforge.net/doc/ppm.html +// +// Known limitations: +// Does not support comments in the header section +// Does not support ASCII image data (formats P2 and P3) +// Does not support 16-bit-per-channel + +#ifndef STBI_NO_PNM + +static int stbi__pnm_test(stbi__context *s) +{ + char p, t; + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind( s ); + return 0; + } + return 1; +} + +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + STBI_NOTUSED(ri); + + if (!stbi__pnm_info(s, (int *)&s->img_x, (int *)&s->img_y, (int *)&s->img_n)) + return 0; + + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + + if (!stbi__mad3sizes_valid(s->img_n, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "PNM too large"); + + out = (stbi_uc *) stbi__malloc_mad3(s->img_n, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + stbi__getn(s, out, s->img_n * s->img_x * s->img_y); + + if (req_comp && req_comp != s->img_n) { + out = stbi__convert_format(out, s->img_n, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + return out; +} + +static int stbi__pnm_isspace(char c) +{ + return c == ' ' || c == '\t' || c == '\n' || c == '\v' || c == '\f' || c == '\r'; +} + +static void stbi__pnm_skip_whitespace(stbi__context *s, char *c) +{ + for (;;) { + while (!stbi__at_eof(s) && stbi__pnm_isspace(*c)) + *c = (char) stbi__get8(s); + + if (stbi__at_eof(s) || *c != '#') + break; + + while (!stbi__at_eof(s) && *c != '\n' && *c != '\r' ) + *c = (char) stbi__get8(s); + } +} + +static int stbi__pnm_isdigit(char c) +{ + return c >= '0' && c <= '9'; +} + +static int stbi__pnm_getinteger(stbi__context *s, char *c) +{ + int value = 0; + + while (!stbi__at_eof(s) && stbi__pnm_isdigit(*c)) { + value = value*10 + (*c - '0'); + *c = (char) stbi__get8(s); + } + + return value; +} + +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp) +{ + int maxv, dummy; + char c, p, t; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + stbi__rewind(s); + + // Get identifier + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind(s); + return 0; + } + + *comp = (t == '6') ? 3 : 1; // '5' is 1-component .pgm; '6' is 3-component .ppm + + c = (char) stbi__get8(s); + stbi__pnm_skip_whitespace(s, &c); + + *x = stbi__pnm_getinteger(s, &c); // read width + stbi__pnm_skip_whitespace(s, &c); + + *y = stbi__pnm_getinteger(s, &c); // read height + stbi__pnm_skip_whitespace(s, &c); + + maxv = stbi__pnm_getinteger(s, &c); // read max value + + if (maxv > 255) + return stbi__err("max value > 255", "PPM image not 8-bit"); + else + return 1; +} +#endif + +static int stbi__info_main(stbi__context *s, int *x, int *y, int *comp) +{ + #ifndef STBI_NO_JPEG + if (stbi__jpeg_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PNG + if (stbi__png_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_GIF + if (stbi__gif_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_BMP + if (stbi__bmp_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PSD + if (stbi__psd_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PIC + if (stbi__pic_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PNM + if (stbi__pnm_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_HDR + if (stbi__hdr_info(s, x, y, comp)) return 1; + #endif + + // test tga last because it's a crappy test! + #ifndef STBI_NO_TGA + if (stbi__tga_info(s, x, y, comp)) + return 1; + #endif + return stbi__err("unknown image type", "Image not of any known type, or corrupt"); +} + +static int stbi__is_16_main(stbi__context *s) +{ + #ifndef STBI_NO_PNG + if (stbi__png_is16(s)) return 1; + #endif + + #ifndef STBI_NO_PSD + if (stbi__psd_is16(s)) return 1; + #endif + + return 0; +} + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info(char const *filename, int *x, int *y, int *comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_info_from_file(f, x, y, comp); + fclose(f); + return result; +} + +STBIDEF int stbi_info_from_file(FILE *f, int *x, int *y, int *comp) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__info_main(&s,x,y,comp); + fseek(f,pos,SEEK_SET); + return r; +} + +STBIDEF int stbi_is_16_bit(char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_is_16_bit_from_file(f); + fclose(f); + return result; +} + +STBIDEF int stbi_is_16_bit_from_file(FILE *f) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__is_16_main(&s); + fseek(f,pos,SEEK_SET); + return r; +} +#endif // !STBI_NO_STDIO + +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__info_main(&s,x,y,comp); +} + +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *c, void *user, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__info_main(&s,x,y,comp); +} + +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__is_16_main(&s); +} + +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *c, void *user) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__is_16_main(&s); +} + +#endif // STB_IMAGE_IMPLEMENTATION + +/* + revision history: + 2.19 (2018-02-11) fix warning + 2.18 (2018-01-30) fix warnings + 2.17 (2018-01-29) change sbti__shiftsigned to avoid clang -O2 bug + 1-bit BMP + *_is_16_bit api + avoid warnings + 2.16 (2017-07-23) all functions have 16-bit variants; + STBI_NO_STDIO works again; + compilation fixes; + fix rounding in unpremultiply; + optimize vertical flip; + disable raw_len validation; + documentation fixes + 2.15 (2017-03-18) fix png-1,2,4 bug; now all Imagenet JPGs decode; + warning fixes; disable run-time SSE detection on gcc; + uniform handling of optional "return" values; + thread-safe initialization of zlib tables + 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs + 2.13 (2016-11-29) add 16-bit API, only supported for PNG right now + 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes + 2.11 (2016-04-02) allocate large structures on the stack + remove white matting for transparent PSD + fix reported channel count for PNG & BMP + re-enable SSE2 in non-gcc 64-bit + support RGB-formatted JPEG + read 16-bit PNGs (only as 8-bit) + 2.10 (2016-01-22) avoid warning introduced in 2.09 by STBI_REALLOC_SIZED + 2.09 (2016-01-16) allow comments in PNM files + 16-bit-per-pixel TGA (not bit-per-component) + info() for TGA could break due to .hdr handling + info() for BMP to shares code instead of sloppy parse + can use STBI_REALLOC_SIZED if allocator doesn't support realloc + code cleanup + 2.08 (2015-09-13) fix to 2.07 cleanup, reading RGB PSD as RGBA + 2.07 (2015-09-13) fix compiler warnings + partial animated GIF support + limited 16-bpc PSD support + #ifdef unused functions + bug with < 92 byte PIC,PNM,HDR,TGA + 2.06 (2015-04-19) fix bug where PSD returns wrong '*comp' value + 2.05 (2015-04-19) fix bug in progressive JPEG handling, fix warning + 2.04 (2015-04-15) try to re-enable SIMD on MinGW 64-bit + 2.03 (2015-04-12) extra corruption checking (mmozeiko) + stbi_set_flip_vertically_on_load (nguillemot) + fix NEON support; fix mingw support + 2.02 (2015-01-19) fix incorrect assert, fix warning + 2.01 (2015-01-17) fix various warnings; suppress SIMD on gcc 32-bit without -msse2 + 2.00b (2014-12-25) fix STBI_MALLOC in progressive JPEG + 2.00 (2014-12-25) optimize JPG, including x86 SSE2 & NEON SIMD (ryg) + progressive JPEG (stb) + PGM/PPM support (Ken Miller) + STBI_MALLOC,STBI_REALLOC,STBI_FREE + GIF bugfix -- seemingly never worked + STBI_NO_*, STBI_ONLY_* + 1.48 (2014-12-14) fix incorrectly-named assert() + 1.47 (2014-12-14) 1/2/4-bit PNG support, both direct and paletted (Omar Cornut & stb) + optimize PNG (ryg) + fix bug in interlaced PNG with user-specified channel count (stb) + 1.46 (2014-08-26) + fix broken tRNS chunk (colorkey-style transparency) in non-paletted PNG + 1.45 (2014-08-16) + fix MSVC-ARM internal compiler error by wrapping malloc + 1.44 (2014-08-07) + various warning fixes from Ronny Chevalier + 1.43 (2014-07-15) + fix MSVC-only compiler problem in code changed in 1.42 + 1.42 (2014-07-09) + don't define _CRT_SECURE_NO_WARNINGS (affects user code) + fixes to stbi__cleanup_jpeg path + added STBI_ASSERT to avoid requiring assert.h + 1.41 (2014-06-25) + fix search&replace from 1.36 that messed up comments/error messages + 1.40 (2014-06-22) + fix gcc struct-initialization warning + 1.39 (2014-06-15) + fix to TGA optimization when req_comp != number of components in TGA; + fix to GIF loading because BMP wasn't rewinding (whoops, no GIFs in my test suite) + add support for BMP version 5 (more ignored fields) + 1.38 (2014-06-06) + suppress MSVC warnings on integer casts truncating values + fix accidental rename of 'skip' field of I/O + 1.37 (2014-06-04) + remove duplicate typedef + 1.36 (2014-06-03) + convert to header file single-file library + if de-iphone isn't set, load iphone images color-swapped instead of returning NULL + 1.35 (2014-05-27) + various warnings + fix broken STBI_SIMD path + fix bug where stbi_load_from_file no longer left file pointer in correct place + fix broken non-easy path for 32-bit BMP (possibly never used) + TGA optimization by Arseny Kapoulkine + 1.34 (unknown) + use STBI_NOTUSED in stbi__resample_row_generic(), fix one more leak in tga failure case + 1.33 (2011-07-14) + make stbi_is_hdr work in STBI_NO_HDR (as specified), minor compiler-friendly improvements + 1.32 (2011-07-13) + support for "info" function for all supported filetypes (SpartanJ) + 1.31 (2011-06-20) + a few more leak fixes, bug in PNG handling (SpartanJ) + 1.30 (2011-06-11) + added ability to load files via callbacks to accomidate custom input streams (Ben Wenger) + removed deprecated format-specific test/load functions + removed support for installable file formats (stbi_loader) -- would have been broken for IO callbacks anyway + error cases in bmp and tga give messages and don't leak (Raymond Barbiero, grisha) + fix inefficiency in decoding 32-bit BMP (David Woo) + 1.29 (2010-08-16) + various warning fixes from Aurelien Pocheville + 1.28 (2010-08-01) + fix bug in GIF palette transparency (SpartanJ) + 1.27 (2010-08-01) + cast-to-stbi_uc to fix warnings + 1.26 (2010-07-24) + fix bug in file buffering for PNG reported by SpartanJ + 1.25 (2010-07-17) + refix trans_data warning (Won Chun) + 1.24 (2010-07-12) + perf improvements reading from files on platforms with lock-heavy fgetc() + minor perf improvements for jpeg + deprecated type-specific functions so we'll get feedback if they're needed + attempt to fix trans_data warning (Won Chun) + 1.23 fixed bug in iPhone support + 1.22 (2010-07-10) + removed image *writing* support + stbi_info support from Jetro Lauha + GIF support from Jean-Marc Lienher + iPhone PNG-extensions from James Brown + warning-fixes from Nicolas Schulz and Janez Zemva (i.stbi__err. Janez (U+017D)emva) + 1.21 fix use of 'stbi_uc' in header (reported by jon blow) + 1.20 added support for Softimage PIC, by Tom Seddon + 1.19 bug in interlaced PNG corruption check (found by ryg) + 1.18 (2008-08-02) + fix a threading bug (local mutable static) + 1.17 support interlaced PNG + 1.16 major bugfix - stbi__convert_format converted one too many pixels + 1.15 initialize some fields for thread safety + 1.14 fix threadsafe conversion bug + header-file-only version (#define STBI_HEADER_FILE_ONLY before including) + 1.13 threadsafe + 1.12 const qualifiers in the API + 1.11 Support installable IDCT, colorspace conversion routines + 1.10 Fixes for 64-bit (don't use "unsigned long") + optimized upsampling by Fabian "ryg" Giesen + 1.09 Fix format-conversion for PSD code (bad global variables!) + 1.08 Thatcher Ulrich's PSD code integrated by Nicolas Schulz + 1.07 attempt to fix C++ warning/errors again + 1.06 attempt to fix C++ warning/errors again + 1.05 fix TGA loading to return correct *comp and use good luminance calc + 1.04 default float alpha is 1, not 255; use 'void *' for stbi_image_free + 1.03 bugfixes to STBI_NO_STDIO, STBI_NO_HDR + 1.02 support for (subset of) HDR files, float interface for preferred access to them + 1.01 fix bug: possible bug in handling right-side up bmps... not sure + fix bug: the stbi__bmp_load() and stbi__tga_load() functions didn't work at all + 1.00 interface to zlib that skips zlib header + 0.99 correct handling of alpha in palette + 0.98 TGA loader by lonesock; dynamically add loaders (untested) + 0.97 jpeg errors on too large a file; also catch another malloc failure + 0.96 fix detection of invalid v value - particleman@mollyrocket forum + 0.95 during header scan, seek to markers in case of padding + 0.94 STBI_NO_STDIO to disable stdio usage; rename all #defines the same + 0.93 handle jpegtran output; verbose errors + 0.92 read 4,8,16,24,32-bit BMP files of several formats + 0.91 output 24-bit Windows 3.0 BMP files + 0.90 fix a few more warnings; bump version number to approach 1.0 + 0.61 bugfixes due to Marc LeBlanc, Christopher Lloyd + 0.60 fix compiling as c++ + 0.59 fix warnings: merge Dave Moore's -Wall fixes + 0.58 fix bug: zlib uncompressed mode len/nlen was wrong endian + 0.57 fix bug: jpg last huffman symbol before marker was >9 bits but less than 16 available + 0.56 fix bug: zlib uncompressed mode len vs. nlen + 0.55 fix bug: restart_interval not initialized to 0 + 0.54 allow NULL for 'int *comp' + 0.53 fix bug in png 3->4; speedup png decoding + 0.52 png handles req_comp=3,4 directly; minor cleanup; jpeg comments + 0.51 obey req_comp requests, 1-component jpegs return as 1-component, + on 'test' only check type, not whether we support this variant + 0.50 (2006-11-19) + first released version +*/ + + +/* +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person obtaining a copy of +this software and associated documentation files (the "Software"), to deal in +the Software without restriction, including without limitation the rights to +use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies +of the Software, and to permit persons to whom the Software is furnished to do +so, subject to the following conditions: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ diff --git a/src/stb_truetype.h b/src/stb_truetype.h new file mode 100644 index 0000000..62595a1 --- /dev/null +++ b/src/stb_truetype.h @@ -0,0 +1,5011 @@ +// stb_truetype.h - v1.24 - public domain +// authored from 2009-2020 by Sean Barrett / RAD Game Tools +// +// ======================================================================= +// +// NO SECURITY GUARANTEE -- DO NOT USE THIS ON UNTRUSTED FONT FILES +// +// This library does no range checking of the offsets found in the file, +// meaning an attacker can use it to read arbitrary memory. +// +// ======================================================================= +// +// This library processes TrueType files: +// parse files +// extract glyph metrics +// extract glyph shapes +// render glyphs to one-channel bitmaps with antialiasing (box filter) +// render glyphs to one-channel SDF bitmaps (signed-distance field/function) +// +// Todo: +// non-MS cmaps +// crashproof on bad data +// hinting? (no longer patented) +// cleartype-style AA? +// optimize: use simple memory allocator for intermediates +// optimize: build edge-list directly from curves +// optimize: rasterize directly from curves? +// +// ADDITIONAL CONTRIBUTORS +// +// Mikko Mononen: compound shape support, more cmap formats +// Tor Andersson: kerning, subpixel rendering +// Dougall Johnson: OpenType / Type 2 font handling +// Daniel Ribeiro Maciel: basic GPOS-based kerning +// +// Misc other: +// Ryan Gordon +// Simon Glass +// github:IntellectualKitty +// Imanol Celaya +// Daniel Ribeiro Maciel +// +// Bug/warning reports/fixes: +// "Zer" on mollyrocket Fabian "ryg" Giesen github:NiLuJe +// Cass Everitt Martins Mozeiko github:aloucks +// stoiko (Haemimont Games) Cap Petschulat github:oyvindjam +// Brian Hook Omar Cornut github:vassvik +// Walter van Niftrik Ryan Griege +// David Gow Peter LaValle +// David Given Sergey Popov +// Ivan-Assen Ivanov Giumo X. Clanjor +// Anthony Pesch Higor Euripedes +// Johan Duparc Thomas Fields +// Hou Qiming Derek Vinyard +// Rob Loach Cort Stratton +// Kenney Phillis Jr. Brian Costabile +// Ken Voskuil (kaesve) +// +// VERSION HISTORY +// +// 1.24 (2020-02-05) fix warning +// 1.23 (2020-02-02) query SVG data for glyphs; query whole kerning table (but only kern not GPOS) +// 1.22 (2019-08-11) minimize missing-glyph duplication; fix kerning if both 'GPOS' and 'kern' are defined +// 1.21 (2019-02-25) fix warning +// 1.20 (2019-02-07) PackFontRange skips missing codepoints; GetScaleFontVMetrics() +// 1.19 (2018-02-11) GPOS kerning, STBTT_fmod +// 1.18 (2018-01-29) add missing function +// 1.17 (2017-07-23) make more arguments const; doc fix +// 1.16 (2017-07-12) SDF support +// 1.15 (2017-03-03) make more arguments const +// 1.14 (2017-01-16) num-fonts-in-TTC function +// 1.13 (2017-01-02) support OpenType fonts, certain Apple fonts +// 1.12 (2016-10-25) suppress warnings about casting away const with -Wcast-qual +// 1.11 (2016-04-02) fix unused-variable warning +// 1.10 (2016-04-02) user-defined fabs(); rare memory leak; remove duplicate typedef +// 1.09 (2016-01-16) warning fix; avoid crash on outofmem; use allocation userdata properly +// 1.08 (2015-09-13) document stbtt_Rasterize(); fixes for vertical & horizontal edges +// 1.07 (2015-08-01) allow PackFontRanges to accept arrays of sparse codepoints; +// variant PackFontRanges to pack and render in separate phases; +// fix stbtt_GetFontOFfsetForIndex (never worked for non-0 input?); +// fixed an assert() bug in the new rasterizer +// replace assert() with STBTT_assert() in new rasterizer +// +// Full history can be found at the end of this file. +// +// LICENSE +// +// See end of file for license information. +// +// USAGE +// +// Include this file in whatever places need to refer to it. In ONE C/C++ +// file, write: +// #define STB_TRUETYPE_IMPLEMENTATION +// before the #include of this file. This expands out the actual +// implementation into that C/C++ file. +// +// To make the implementation private to the file that generates the implementation, +// #define STBTT_STATIC +// +// Simple 3D API (don't ship this, but it's fine for tools and quick start) +// stbtt_BakeFontBitmap() -- bake a font to a bitmap for use as texture +// stbtt_GetBakedQuad() -- compute quad to draw for a given char +// +// Improved 3D API (more shippable): +// #include "stb_rect_pack.h" -- optional, but you really want it +// stbtt_PackBegin() +// stbtt_PackSetOversampling() -- for improved quality on small fonts +// stbtt_PackFontRanges() -- pack and renders +// stbtt_PackEnd() +// stbtt_GetPackedQuad() +// +// "Load" a font file from a memory buffer (you have to keep the buffer loaded) +// stbtt_InitFont() +// stbtt_GetFontOffsetForIndex() -- indexing for TTC font collections +// stbtt_GetNumberOfFonts() -- number of fonts for TTC font collections +// +// Render a unicode codepoint to a bitmap +// stbtt_GetCodepointBitmap() -- allocates and returns a bitmap +// stbtt_MakeCodepointBitmap() -- renders into bitmap you provide +// stbtt_GetCodepointBitmapBox() -- how big the bitmap must be +// +// Character advance/positioning +// stbtt_GetCodepointHMetrics() +// stbtt_GetFontVMetrics() +// stbtt_GetFontVMetricsOS2() +// stbtt_GetCodepointKernAdvance() +// +// Starting with version 1.06, the rasterizer was replaced with a new, +// faster and generally-more-precise rasterizer. The new rasterizer more +// accurately measures pixel coverage for anti-aliasing, except in the case +// where multiple shapes overlap, in which case it overestimates the AA pixel +// coverage. Thus, anti-aliasing of intersecting shapes may look wrong. If +// this turns out to be a problem, you can re-enable the old rasterizer with +// #define STBTT_RASTERIZER_VERSION 1 +// which will incur about a 15% speed hit. +// +// ADDITIONAL DOCUMENTATION +// +// Immediately after this block comment are a series of sample programs. +// +// After the sample programs is the "header file" section. This section +// includes documentation for each API function. +// +// Some important concepts to understand to use this library: +// +// Codepoint +// Characters are defined by unicode codepoints, e.g. 65 is +// uppercase A, 231 is lowercase c with a cedilla, 0x7e30 is +// the hiragana for "ma". +// +// Glyph +// A visual character shape (every codepoint is rendered as +// some glyph) +// +// Glyph index +// A font-specific integer ID representing a glyph +// +// Baseline +// Glyph shapes are defined relative to a baseline, which is the +// bottom of uppercase characters. Characters extend both above +// and below the baseline. +// +// Current Point +// As you draw text to the screen, you keep track of a "current point" +// which is the origin of each character. The current point's vertical +// position is the baseline. Even "baked fonts" use this model. +// +// Vertical Font Metrics +// The vertical qualities of the font, used to vertically position +// and space the characters. See docs for stbtt_GetFontVMetrics. +// +// Font Size in Pixels or Points +// The preferred interface for specifying font sizes in stb_truetype +// is to specify how tall the font's vertical extent should be in pixels. +// If that sounds good enough, skip the next paragraph. +// +// Most font APIs instead use "points", which are a common typographic +// measurement for describing font size, defined as 72 points per inch. +// stb_truetype provides a point API for compatibility. However, true +// "per inch" conventions don't make much sense on computer displays +// since different monitors have different number of pixels per +// inch. For example, Windows traditionally uses a convention that +// there are 96 pixels per inch, thus making 'inch' measurements have +// nothing to do with inches, and thus effectively defining a point to +// be 1.333 pixels. Additionally, the TrueType font data provides +// an explicit scale factor to scale a given font's glyphs to points, +// but the author has observed that this scale factor is often wrong +// for non-commercial fonts, thus making fonts scaled in points +// according to the TrueType spec incoherently sized in practice. +// +// DETAILED USAGE: +// +// Scale: +// Select how high you want the font to be, in points or pixels. +// Call ScaleForPixelHeight or ScaleForMappingEmToPixels to compute +// a scale factor SF that will be used by all other functions. +// +// Baseline: +// You need to select a y-coordinate that is the baseline of where +// your text will appear. Call GetFontBoundingBox to get the baseline-relative +// bounding box for all characters. SF*-y0 will be the distance in pixels +// that the worst-case character could extend above the baseline, so if +// you want the top edge of characters to appear at the top of the +// screen where y=0, then you would set the baseline to SF*-y0. +// +// Current point: +// Set the current point where the first character will appear. The +// first character could extend left of the current point; this is font +// dependent. You can either choose a current point that is the leftmost +// point and hope, or add some padding, or check the bounding box or +// left-side-bearing of the first character to be displayed and set +// the current point based on that. +// +// Displaying a character: +// Compute the bounding box of the character. It will contain signed values +// relative to . I.e. if it returns x0,y0,x1,y1, +// then the character should be displayed in the rectangle from +// to = 32 && *text < 128) { + stbtt_aligned_quad q; + stbtt_GetBakedQuad(cdata, 512,512, *text-32, &x,&y,&q,1);//1=opengl & d3d10+,0=d3d9 + glTexCoord2f(q.s0,q.t1); glVertex2f(q.x0,q.y0); + glTexCoord2f(q.s1,q.t1); glVertex2f(q.x1,q.y0); + glTexCoord2f(q.s1,q.t0); glVertex2f(q.x1,q.y1); + glTexCoord2f(q.s0,q.t0); glVertex2f(q.x0,q.y1); + } + ++text; + } + glEnd(); +} +#endif +// +// +////////////////////////////////////////////////////////////////////////////// +// +// Complete program (this compiles): get a single bitmap, print as ASCII art +// +#if 0 +#include +#define STB_TRUETYPE_IMPLEMENTATION // force following include to generate implementation +#include "stb_truetype.h" + +char ttf_buffer[1<<25]; + +int main(int argc, char **argv) +{ + stbtt_fontinfo font; + unsigned char *bitmap; + int w,h,i,j,c = (argc > 1 ? atoi(argv[1]) : 'a'), s = (argc > 2 ? atoi(argv[2]) : 20); + + fread(ttf_buffer, 1, 1<<25, fopen(argc > 3 ? argv[3] : "c:/windows/fonts/arialbd.ttf", "rb")); + + stbtt_InitFont(&font, ttf_buffer, stbtt_GetFontOffsetForIndex(ttf_buffer,0)); + bitmap = stbtt_GetCodepointBitmap(&font, 0,stbtt_ScaleForPixelHeight(&font, s), c, &w, &h, 0,0); + + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) + putchar(" .:ioVM@"[bitmap[j*w+i]>>5]); + putchar('\n'); + } + return 0; +} +#endif +// +// Output: +// +// .ii. +// @@@@@@. +// V@Mio@@o +// :i. V@V +// :oM@@M +// :@@@MM@M +// @@o o@M +// :@@. M@M +// @@@o@@@@ +// :M@@V:@@. +// +////////////////////////////////////////////////////////////////////////////// +// +// Complete program: print "Hello World!" banner, with bugs +// +#if 0 +char buffer[24<<20]; +unsigned char screen[20][79]; + +int main(int arg, char **argv) +{ + stbtt_fontinfo font; + int i,j,ascent,baseline,ch=0; + float scale, xpos=2; // leave a little padding in case the character extends left + char *text = "Heljo World!"; // intentionally misspelled to show 'lj' brokenness + + fread(buffer, 1, 1000000, fopen("c:/windows/fonts/arialbd.ttf", "rb")); + stbtt_InitFont(&font, buffer, 0); + + scale = stbtt_ScaleForPixelHeight(&font, 15); + stbtt_GetFontVMetrics(&font, &ascent,0,0); + baseline = (int) (ascent*scale); + + while (text[ch]) { + int advance,lsb,x0,y0,x1,y1; + float x_shift = xpos - (float) floor(xpos); + stbtt_GetCodepointHMetrics(&font, text[ch], &advance, &lsb); + stbtt_GetCodepointBitmapBoxSubpixel(&font, text[ch], scale,scale,x_shift,0, &x0,&y0,&x1,&y1); + stbtt_MakeCodepointBitmapSubpixel(&font, &screen[baseline + y0][(int) xpos + x0], x1-x0,y1-y0, 79, scale,scale,x_shift,0, text[ch]); + // note that this stomps the old data, so where character boxes overlap (e.g. 'lj') it's wrong + // because this API is really for baking character bitmaps into textures. if you want to render + // a sequence of characters, you really need to render each bitmap to a temp buffer, then + // "alpha blend" that into the working buffer + xpos += (advance * scale); + if (text[ch+1]) + xpos += scale*stbtt_GetCodepointKernAdvance(&font, text[ch],text[ch+1]); + ++ch; + } + + for (j=0; j < 20; ++j) { + for (i=0; i < 78; ++i) + putchar(" .:ioVM@"[screen[j][i]>>5]); + putchar('\n'); + } + + return 0; +} +#endif + + +////////////////////////////////////////////////////////////////////////////// +////////////////////////////////////////////////////////////////////////////// +//// +//// INTEGRATION WITH YOUR CODEBASE +//// +//// The following sections allow you to supply alternate definitions +//// of C library functions used by stb_truetype, e.g. if you don't +//// link with the C runtime library. + +#ifdef STB_TRUETYPE_IMPLEMENTATION + // #define your own (u)stbtt_int8/16/32 before including to override this + #ifndef stbtt_uint8 + typedef unsigned char stbtt_uint8; + typedef signed char stbtt_int8; + typedef unsigned short stbtt_uint16; + typedef signed short stbtt_int16; + typedef unsigned int stbtt_uint32; + typedef signed int stbtt_int32; + #endif + + typedef char stbtt__check_size32[sizeof(stbtt_int32)==4 ? 1 : -1]; + typedef char stbtt__check_size16[sizeof(stbtt_int16)==2 ? 1 : -1]; + + // e.g. #define your own STBTT_ifloor/STBTT_iceil() to avoid math.h + #ifndef STBTT_ifloor + #include + #define STBTT_ifloor(x) ((int) floor(x)) + #define STBTT_iceil(x) ((int) ceil(x)) + #endif + + #ifndef STBTT_sqrt + #include + #define STBTT_sqrt(x) sqrt(x) + #define STBTT_pow(x,y) pow(x,y) + #endif + + #ifndef STBTT_fmod + #include + #define STBTT_fmod(x,y) fmod(x,y) + #endif + + #ifndef STBTT_cos + #include + #define STBTT_cos(x) cos(x) + #define STBTT_acos(x) acos(x) + #endif + + #ifndef STBTT_fabs + #include + #define STBTT_fabs(x) fabs(x) + #endif + + // #define your own functions "STBTT_malloc" / "STBTT_free" to avoid malloc.h + #ifndef STBTT_malloc + #include + #define STBTT_malloc(x,u) ((void)(u),malloc(x)) + #define STBTT_free(x,u) ((void)(u),free(x)) + #endif + + #ifndef STBTT_assert + #include + #define STBTT_assert(x) assert(x) + #endif + + #ifndef STBTT_strlen + #include + #define STBTT_strlen(x) strlen(x) + #endif + + #ifndef STBTT_memcpy + #include + #define STBTT_memcpy memcpy + #define STBTT_memset memset + #endif +#endif + +/////////////////////////////////////////////////////////////////////////////// +/////////////////////////////////////////////////////////////////////////////// +//// +//// INTERFACE +//// +//// + +#ifndef __STB_INCLUDE_STB_TRUETYPE_H__ +#define __STB_INCLUDE_STB_TRUETYPE_H__ + +#ifdef STBTT_STATIC +#define STBTT_DEF static +#else +#define STBTT_DEF extern +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +// private structure +typedef struct +{ + unsigned char *data; + int cursor; + int size; +} stbtt__buf; + +////////////////////////////////////////////////////////////////////////////// +// +// TEXTURE BAKING API +// +// If you use this API, you only have to call two functions ever. +// + +typedef struct +{ + unsigned short x0,y0,x1,y1; // coordinates of bbox in bitmap + float xoff,yoff,xadvance; +} stbtt_bakedchar; + +STBTT_DEF int stbtt_BakeFontBitmap(const unsigned char *data, int offset, // font location (use offset=0 for plain .ttf) + float pixel_height, // height of font in pixels + unsigned char *pixels, int pw, int ph, // bitmap to be filled in + int first_char, int num_chars, // characters to bake + stbtt_bakedchar *chardata); // you allocate this, it's num_chars long +// if return is positive, the first unused row of the bitmap +// if return is negative, returns the negative of the number of characters that fit +// if return is 0, no characters fit and no rows were used +// This uses a very crappy packing. + +typedef struct +{ + float x0,y0,s0,t0; // top-left + float x1,y1,s1,t1; // bottom-right +} stbtt_aligned_quad; + +STBTT_DEF void stbtt_GetBakedQuad(const stbtt_bakedchar *chardata, int pw, int ph, // same data as above + int char_index, // character to display + float *xpos, float *ypos, // pointers to current position in screen pixel space + stbtt_aligned_quad *q, // output: quad to draw + int opengl_fillrule); // true if opengl fill rule; false if DX9 or earlier +// Call GetBakedQuad with char_index = 'character - first_char', and it +// creates the quad you need to draw and advances the current position. +// +// The coordinate system used assumes y increases downwards. +// +// Characters will extend both above and below the current position; +// see discussion of "BASELINE" above. +// +// It's inefficient; you might want to c&p it and optimize it. + +STBTT_DEF void stbtt_GetScaledFontVMetrics(const unsigned char *fontdata, int index, float size, float *ascent, float *descent, float *lineGap); +// Query the font vertical metrics without having to create a font first. + + +////////////////////////////////////////////////////////////////////////////// +// +// NEW TEXTURE BAKING API +// +// This provides options for packing multiple fonts into one atlas, not +// perfectly but better than nothing. + +typedef struct +{ + unsigned short x0,y0,x1,y1; // coordinates of bbox in bitmap + float xoff,yoff,xadvance; + float xoff2,yoff2; +} stbtt_packedchar; + +typedef struct stbtt_pack_context stbtt_pack_context; +typedef struct stbtt_fontinfo stbtt_fontinfo; +#ifndef STB_RECT_PACK_VERSION +typedef struct stbrp_rect stbrp_rect; +#endif + +STBTT_DEF int stbtt_PackBegin(stbtt_pack_context *spc, unsigned char *pixels, int width, int height, int stride_in_bytes, int padding, void *alloc_context); +// Initializes a packing context stored in the passed-in stbtt_pack_context. +// Future calls using this context will pack characters into the bitmap passed +// in here: a 1-channel bitmap that is width * height. stride_in_bytes is +// the distance from one row to the next (or 0 to mean they are packed tightly +// together). "padding" is the amount of padding to leave between each +// character (normally you want '1' for bitmaps you'll use as textures with +// bilinear filtering). +// +// Returns 0 on failure, 1 on success. + +STBTT_DEF void stbtt_PackEnd (stbtt_pack_context *spc); +// Cleans up the packing context and frees all memory. + +#define STBTT_POINT_SIZE(x) (-(x)) + +STBTT_DEF int stbtt_PackFontRange(stbtt_pack_context *spc, const unsigned char *fontdata, int font_index, float font_size, + int first_unicode_char_in_range, int num_chars_in_range, stbtt_packedchar *chardata_for_range); +// Creates character bitmaps from the font_index'th font found in fontdata (use +// font_index=0 if you don't know what that is). It creates num_chars_in_range +// bitmaps for characters with unicode values starting at first_unicode_char_in_range +// and increasing. Data for how to render them is stored in chardata_for_range; +// pass these to stbtt_GetPackedQuad to get back renderable quads. +// +// font_size is the full height of the character from ascender to descender, +// as computed by stbtt_ScaleForPixelHeight. To use a point size as computed +// by stbtt_ScaleForMappingEmToPixels, wrap the point size in STBTT_POINT_SIZE() +// and pass that result as 'font_size': +// ..., 20 , ... // font max minus min y is 20 pixels tall +// ..., STBTT_POINT_SIZE(20), ... // 'M' is 20 pixels tall + +typedef struct +{ + float font_size; + int first_unicode_codepoint_in_range; // if non-zero, then the chars are continuous, and this is the first codepoint + int *array_of_unicode_codepoints; // if non-zero, then this is an array of unicode codepoints + int num_chars; + stbtt_packedchar *chardata_for_range; // output + unsigned char h_oversample, v_oversample; // don't set these, they're used internally +} stbtt_pack_range; + +STBTT_DEF int stbtt_PackFontRanges(stbtt_pack_context *spc, const unsigned char *fontdata, int font_index, stbtt_pack_range *ranges, int num_ranges); +// Creates character bitmaps from multiple ranges of characters stored in +// ranges. This will usually create a better-packed bitmap than multiple +// calls to stbtt_PackFontRange. Note that you can call this multiple +// times within a single PackBegin/PackEnd. + +STBTT_DEF void stbtt_PackSetOversampling(stbtt_pack_context *spc, unsigned int h_oversample, unsigned int v_oversample); +// Oversampling a font increases the quality by allowing higher-quality subpixel +// positioning, and is especially valuable at smaller text sizes. +// +// This function sets the amount of oversampling for all following calls to +// stbtt_PackFontRange(s) or stbtt_PackFontRangesGatherRects for a given +// pack context. The default (no oversampling) is achieved by h_oversample=1 +// and v_oversample=1. The total number of pixels required is +// h_oversample*v_oversample larger than the default; for example, 2x2 +// oversampling requires 4x the storage of 1x1. For best results, render +// oversampled textures with bilinear filtering. Look at the readme in +// stb/tests/oversample for information about oversampled fonts +// +// To use with PackFontRangesGather etc., you must set it before calls +// call to PackFontRangesGatherRects. + +STBTT_DEF void stbtt_PackSetSkipMissingCodepoints(stbtt_pack_context *spc, int skip); +// If skip != 0, this tells stb_truetype to skip any codepoints for which +// there is no corresponding glyph. If skip=0, which is the default, then +// codepoints without a glyph recived the font's "missing character" glyph, +// typically an empty box by convention. + +STBTT_DEF void stbtt_GetPackedQuad(const stbtt_packedchar *chardata, int pw, int ph, // same data as above + int char_index, // character to display + float *xpos, float *ypos, // pointers to current position in screen pixel space + stbtt_aligned_quad *q, // output: quad to draw + int align_to_integer); + +STBTT_DEF int stbtt_PackFontRangesGatherRects(stbtt_pack_context *spc, const stbtt_fontinfo *info, stbtt_pack_range *ranges, int num_ranges, stbrp_rect *rects); +STBTT_DEF void stbtt_PackFontRangesPackRects(stbtt_pack_context *spc, stbrp_rect *rects, int num_rects); +STBTT_DEF int stbtt_PackFontRangesRenderIntoRects(stbtt_pack_context *spc, const stbtt_fontinfo *info, stbtt_pack_range *ranges, int num_ranges, stbrp_rect *rects); +// Calling these functions in sequence is roughly equivalent to calling +// stbtt_PackFontRanges(). If you more control over the packing of multiple +// fonts, or if you want to pack custom data into a font texture, take a look +// at the source to of stbtt_PackFontRanges() and create a custom version +// using these functions, e.g. call GatherRects multiple times, +// building up a single array of rects, then call PackRects once, +// then call RenderIntoRects repeatedly. This may result in a +// better packing than calling PackFontRanges multiple times +// (or it may not). + +// this is an opaque structure that you shouldn't mess with which holds +// all the context needed from PackBegin to PackEnd. +struct stbtt_pack_context { + void *user_allocator_context; + void *pack_info; + int width; + int height; + int stride_in_bytes; + int padding; + int skip_missing; + unsigned int h_oversample, v_oversample; + unsigned char *pixels; + void *nodes; +}; + +////////////////////////////////////////////////////////////////////////////// +// +// FONT LOADING +// +// + +STBTT_DEF int stbtt_GetNumberOfFonts(const unsigned char *data); +// This function will determine the number of fonts in a font file. TrueType +// collection (.ttc) files may contain multiple fonts, while TrueType font +// (.ttf) files only contain one font. The number of fonts can be used for +// indexing with the previous function where the index is between zero and one +// less than the total fonts. If an error occurs, -1 is returned. + +STBTT_DEF int stbtt_GetFontOffsetForIndex(const unsigned char *data, int index); +// Each .ttf/.ttc file may have more than one font. Each font has a sequential +// index number starting from 0. Call this function to get the font offset for +// a given index; it returns -1 if the index is out of range. A regular .ttf +// file will only define one font and it always be at offset 0, so it will +// return '0' for index 0, and -1 for all other indices. + +// The following structure is defined publicly so you can declare one on +// the stack or as a global or etc, but you should treat it as opaque. +struct stbtt_fontinfo +{ + void * userdata; + unsigned char * data; // pointer to .ttf file + int fontstart; // offset of start of font + + int numGlyphs; // number of glyphs, needed for range checking + + int loca,head,glyf,hhea,hmtx,kern,gpos,svg; // table locations as offset from start of .ttf + int index_map; // a cmap mapping for our chosen character encoding + int indexToLocFormat; // format needed to map from glyph index to glyph + + stbtt__buf cff; // cff font data + stbtt__buf charstrings; // the charstring index + stbtt__buf gsubrs; // global charstring subroutines index + stbtt__buf subrs; // private charstring subroutines index + stbtt__buf fontdicts; // array of font dicts + stbtt__buf fdselect; // map from glyph to fontdict +}; + +STBTT_DEF int stbtt_InitFont(stbtt_fontinfo *info, const unsigned char *data, int offset); +// Given an offset into the file that defines a font, this function builds +// the necessary cached info for the rest of the system. You must allocate +// the stbtt_fontinfo yourself, and stbtt_InitFont will fill it out. You don't +// need to do anything special to free it, because the contents are pure +// value data with no additional data structures. Returns 0 on failure. + + +////////////////////////////////////////////////////////////////////////////// +// +// CHARACTER TO GLYPH-INDEX CONVERSIOn + +STBTT_DEF int stbtt_FindGlyphIndex(const stbtt_fontinfo *info, int unicode_codepoint); +// If you're going to perform multiple operations on the same character +// and you want a speed-up, call this function with the character you're +// going to process, then use glyph-based functions instead of the +// codepoint-based functions. +// Returns 0 if the character codepoint is not defined in the font. + + +////////////////////////////////////////////////////////////////////////////// +// +// CHARACTER PROPERTIES +// + +STBTT_DEF float stbtt_ScaleForPixelHeight(const stbtt_fontinfo *info, float pixels); +// computes a scale factor to produce a font whose "height" is 'pixels' tall. +// Height is measured as the distance from the highest ascender to the lowest +// descender; in other words, it's equivalent to calling stbtt_GetFontVMetrics +// and computing: +// scale = pixels / (ascent - descent) +// so if you prefer to measure height by the ascent only, use a similar calculation. + +STBTT_DEF float stbtt_ScaleForMappingEmToPixels(const stbtt_fontinfo *info, float pixels); +// computes a scale factor to produce a font whose EM size is mapped to +// 'pixels' tall. This is probably what traditional APIs compute, but +// I'm not positive. + +STBTT_DEF void stbtt_GetFontVMetrics(const stbtt_fontinfo *info, int *ascent, int *descent, int *lineGap); +// ascent is the coordinate above the baseline the font extends; descent +// is the coordinate below the baseline the font extends (i.e. it is typically negative) +// lineGap is the spacing between one row's descent and the next row's ascent... +// so you should advance the vertical position by "*ascent - *descent + *lineGap" +// these are expressed in unscaled coordinates, so you must multiply by +// the scale factor for a given size + +STBTT_DEF int stbtt_GetFontVMetricsOS2(const stbtt_fontinfo *info, int *typoAscent, int *typoDescent, int *typoLineGap); +// analogous to GetFontVMetrics, but returns the "typographic" values from the OS/2 +// table (specific to MS/Windows TTF files). +// +// Returns 1 on success (table present), 0 on failure. + +STBTT_DEF void stbtt_GetFontBoundingBox(const stbtt_fontinfo *info, int *x0, int *y0, int *x1, int *y1); +// the bounding box around all possible characters + +STBTT_DEF void stbtt_GetCodepointHMetrics(const stbtt_fontinfo *info, int codepoint, int *advanceWidth, int *leftSideBearing); +// leftSideBearing is the offset from the current horizontal position to the left edge of the character +// advanceWidth is the offset from the current horizontal position to the next horizontal position +// these are expressed in unscaled coordinates + +STBTT_DEF int stbtt_GetCodepointKernAdvance(const stbtt_fontinfo *info, int ch1, int ch2); +// an additional amount to add to the 'advance' value between ch1 and ch2 + +STBTT_DEF int stbtt_GetCodepointBox(const stbtt_fontinfo *info, int codepoint, int *x0, int *y0, int *x1, int *y1); +// Gets the bounding box of the visible part of the glyph, in unscaled coordinates + +STBTT_DEF void stbtt_GetGlyphHMetrics(const stbtt_fontinfo *info, int glyph_index, int *advanceWidth, int *leftSideBearing); +STBTT_DEF int stbtt_GetGlyphKernAdvance(const stbtt_fontinfo *info, int glyph1, int glyph2); +STBTT_DEF int stbtt_GetGlyphBox(const stbtt_fontinfo *info, int glyph_index, int *x0, int *y0, int *x1, int *y1); +// as above, but takes one or more glyph indices for greater efficiency + +typedef struct stbtt_kerningentry +{ + int glyph1; // use stbtt_FindGlyphIndex + int glyph2; + int advance; +} stbtt_kerningentry; + +STBTT_DEF int stbtt_GetKerningTableLength(const stbtt_fontinfo *info); +STBTT_DEF int stbtt_GetKerningTable(const stbtt_fontinfo *info, stbtt_kerningentry* table, int table_length); +// Retrieves a complete list of all of the kerning pairs provided by the font +// stbtt_GetKerningTable never writes more than table_length entries and returns how many entries it did write. +// The table will be sorted by (a.glyph1 == b.glyph1)?(a.glyph2 < b.glyph2):(a.glyph1 < b.glyph1) + +////////////////////////////////////////////////////////////////////////////// +// +// GLYPH SHAPES (you probably don't need these, but they have to go before +// the bitmaps for C declaration-order reasons) +// + +#ifndef STBTT_vmove // you can predefine these to use different values (but why?) + enum { + STBTT_vmove=1, + STBTT_vline, + STBTT_vcurve, + STBTT_vcubic + }; +#endif + +#ifndef stbtt_vertex // you can predefine this to use different values + // (we share this with other code at RAD) + #define stbtt_vertex_type short // can't use stbtt_int16 because that's not visible in the header file + typedef struct + { + stbtt_vertex_type x,y,cx,cy,cx1,cy1; + unsigned char type,padding; + } stbtt_vertex; +#endif + +STBTT_DEF int stbtt_IsGlyphEmpty(const stbtt_fontinfo *info, int glyph_index); +// returns non-zero if nothing is drawn for this glyph + +STBTT_DEF int stbtt_GetCodepointShape(const stbtt_fontinfo *info, int unicode_codepoint, stbtt_vertex **vertices); +STBTT_DEF int stbtt_GetGlyphShape(const stbtt_fontinfo *info, int glyph_index, stbtt_vertex **vertices); +// returns # of vertices and fills *vertices with the pointer to them +// these are expressed in "unscaled" coordinates +// +// The shape is a series of contours. Each one starts with +// a STBTT_moveto, then consists of a series of mixed +// STBTT_lineto and STBTT_curveto segments. A lineto +// draws a line from previous endpoint to its x,y; a curveto +// draws a quadratic bezier from previous endpoint to +// its x,y, using cx,cy as the bezier control point. + +STBTT_DEF void stbtt_FreeShape(const stbtt_fontinfo *info, stbtt_vertex *vertices); +// frees the data allocated above + +STBTT_DEF int stbtt_GetCodepointSVG(const stbtt_fontinfo *info, int unicode_codepoint, const char **svg); +STBTT_DEF int stbtt_GetGlyphSVG(const stbtt_fontinfo *info, int gl, const char **svg); +// fills svg with the character's SVG data. +// returns data size or 0 if SVG not found. + +////////////////////////////////////////////////////////////////////////////// +// +// BITMAP RENDERING +// + +STBTT_DEF void stbtt_FreeBitmap(unsigned char *bitmap, void *userdata); +// frees the bitmap allocated below + +STBTT_DEF unsigned char *stbtt_GetCodepointBitmap(const stbtt_fontinfo *info, float scale_x, float scale_y, int codepoint, int *width, int *height, int *xoff, int *yoff); +// allocates a large-enough single-channel 8bpp bitmap and renders the +// specified character/glyph at the specified scale into it, with +// antialiasing. 0 is no coverage (transparent), 255 is fully covered (opaque). +// *width & *height are filled out with the width & height of the bitmap, +// which is stored left-to-right, top-to-bottom. +// +// xoff/yoff are the offset it pixel space from the glyph origin to the top-left of the bitmap + +STBTT_DEF unsigned char *stbtt_GetCodepointBitmapSubpixel(const stbtt_fontinfo *info, float scale_x, float scale_y, float shift_x, float shift_y, int codepoint, int *width, int *height, int *xoff, int *yoff); +// the same as stbtt_GetCodepoitnBitmap, but you can specify a subpixel +// shift for the character + +STBTT_DEF void stbtt_MakeCodepointBitmap(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, int codepoint); +// the same as stbtt_GetCodepointBitmap, but you pass in storage for the bitmap +// in the form of 'output', with row spacing of 'out_stride' bytes. the bitmap +// is clipped to out_w/out_h bytes. Call stbtt_GetCodepointBitmapBox to get the +// width and height and positioning info for it first. + +STBTT_DEF void stbtt_MakeCodepointBitmapSubpixel(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int codepoint); +// same as stbtt_MakeCodepointBitmap, but you can specify a subpixel +// shift for the character + +STBTT_DEF void stbtt_MakeCodepointBitmapSubpixelPrefilter(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int oversample_x, int oversample_y, float *sub_x, float *sub_y, int codepoint); +// same as stbtt_MakeCodepointBitmapSubpixel, but prefiltering +// is performed (see stbtt_PackSetOversampling) + +STBTT_DEF void stbtt_GetCodepointBitmapBox(const stbtt_fontinfo *font, int codepoint, float scale_x, float scale_y, int *ix0, int *iy0, int *ix1, int *iy1); +// get the bbox of the bitmap centered around the glyph origin; so the +// bitmap width is ix1-ix0, height is iy1-iy0, and location to place +// the bitmap top left is (leftSideBearing*scale,iy0). +// (Note that the bitmap uses y-increases-down, but the shape uses +// y-increases-up, so CodepointBitmapBox and CodepointBox are inverted.) + +STBTT_DEF void stbtt_GetCodepointBitmapBoxSubpixel(const stbtt_fontinfo *font, int codepoint, float scale_x, float scale_y, float shift_x, float shift_y, int *ix0, int *iy0, int *ix1, int *iy1); +// same as stbtt_GetCodepointBitmapBox, but you can specify a subpixel +// shift for the character + +// the following functions are equivalent to the above functions, but operate +// on glyph indices instead of Unicode codepoints (for efficiency) +STBTT_DEF unsigned char *stbtt_GetGlyphBitmap(const stbtt_fontinfo *info, float scale_x, float scale_y, int glyph, int *width, int *height, int *xoff, int *yoff); +STBTT_DEF unsigned char *stbtt_GetGlyphBitmapSubpixel(const stbtt_fontinfo *info, float scale_x, float scale_y, float shift_x, float shift_y, int glyph, int *width, int *height, int *xoff, int *yoff); +STBTT_DEF void stbtt_MakeGlyphBitmap(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, int glyph); +STBTT_DEF void stbtt_MakeGlyphBitmapSubpixel(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int glyph); +STBTT_DEF void stbtt_MakeGlyphBitmapSubpixelPrefilter(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int oversample_x, int oversample_y, float *sub_x, float *sub_y, int glyph); +STBTT_DEF void stbtt_GetGlyphBitmapBox(const stbtt_fontinfo *font, int glyph, float scale_x, float scale_y, int *ix0, int *iy0, int *ix1, int *iy1); +STBTT_DEF void stbtt_GetGlyphBitmapBoxSubpixel(const stbtt_fontinfo *font, int glyph, float scale_x, float scale_y,float shift_x, float shift_y, int *ix0, int *iy0, int *ix1, int *iy1); + + +// @TODO: don't expose this structure +typedef struct +{ + int w,h,stride; + unsigned char *pixels; +} stbtt__bitmap; + +// rasterize a shape with quadratic beziers into a bitmap +STBTT_DEF void stbtt_Rasterize(stbtt__bitmap *result, // 1-channel bitmap to draw into + float flatness_in_pixels, // allowable error of curve in pixels + stbtt_vertex *vertices, // array of vertices defining shape + int num_verts, // number of vertices in above array + float scale_x, float scale_y, // scale applied to input vertices + float shift_x, float shift_y, // translation applied to input vertices + int x_off, int y_off, // another translation applied to input + int invert, // if non-zero, vertically flip shape + void *userdata); // context for to STBTT_MALLOC + +////////////////////////////////////////////////////////////////////////////// +// +// Signed Distance Function (or Field) rendering + +STBTT_DEF void stbtt_FreeSDF(unsigned char *bitmap, void *userdata); +// frees the SDF bitmap allocated below + +STBTT_DEF unsigned char * stbtt_GetGlyphSDF(const stbtt_fontinfo *info, float scale, int glyph, int padding, unsigned char onedge_value, float pixel_dist_scale, int *width, int *height, int *xoff, int *yoff); +STBTT_DEF unsigned char * stbtt_GetCodepointSDF(const stbtt_fontinfo *info, float scale, int codepoint, int padding, unsigned char onedge_value, float pixel_dist_scale, int *width, int *height, int *xoff, int *yoff); +// These functions compute a discretized SDF field for a single character, suitable for storing +// in a single-channel texture, sampling with bilinear filtering, and testing against +// larger than some threshold to produce scalable fonts. +// info -- the font +// scale -- controls the size of the resulting SDF bitmap, same as it would be creating a regular bitmap +// glyph/codepoint -- the character to generate the SDF for +// padding -- extra "pixels" around the character which are filled with the distance to the character (not 0), +// which allows effects like bit outlines +// onedge_value -- value 0-255 to test the SDF against to reconstruct the character (i.e. the isocontour of the character) +// pixel_dist_scale -- what value the SDF should increase by when moving one SDF "pixel" away from the edge (on the 0..255 scale) +// if positive, > onedge_value is inside; if negative, < onedge_value is inside +// width,height -- output height & width of the SDF bitmap (including padding) +// xoff,yoff -- output origin of the character +// return value -- a 2D array of bytes 0..255, width*height in size +// +// pixel_dist_scale & onedge_value are a scale & bias that allows you to make +// optimal use of the limited 0..255 for your application, trading off precision +// and special effects. SDF values outside the range 0..255 are clamped to 0..255. +// +// Example: +// scale = stbtt_ScaleForPixelHeight(22) +// padding = 5 +// onedge_value = 180 +// pixel_dist_scale = 180/5.0 = 36.0 +// +// This will create an SDF bitmap in which the character is about 22 pixels +// high but the whole bitmap is about 22+5+5=32 pixels high. To produce a filled +// shape, sample the SDF at each pixel and fill the pixel if the SDF value +// is greater than or equal to 180/255. (You'll actually want to antialias, +// which is beyond the scope of this example.) Additionally, you can compute +// offset outlines (e.g. to stroke the character border inside & outside, +// or only outside). For example, to fill outside the character up to 3 SDF +// pixels, you would compare against (180-36.0*3)/255 = 72/255. The above +// choice of variables maps a range from 5 pixels outside the shape to +// 2 pixels inside the shape to 0..255; this is intended primarily for apply +// outside effects only (the interior range is needed to allow proper +// antialiasing of the font at *smaller* sizes) +// +// The function computes the SDF analytically at each SDF pixel, not by e.g. +// building a higher-res bitmap and approximating it. In theory the quality +// should be as high as possible for an SDF of this size & representation, but +// unclear if this is true in practice (perhaps building a higher-res bitmap +// and computing from that can allow drop-out prevention). +// +// The algorithm has not been optimized at all, so expect it to be slow +// if computing lots of characters or very large sizes. + + + +////////////////////////////////////////////////////////////////////////////// +// +// Finding the right font... +// +// You should really just solve this offline, keep your own tables +// of what font is what, and don't try to get it out of the .ttf file. +// That's because getting it out of the .ttf file is really hard, because +// the names in the file can appear in many possible encodings, in many +// possible languages, and e.g. if you need a case-insensitive comparison, +// the details of that depend on the encoding & language in a complex way +// (actually underspecified in truetype, but also gigantic). +// +// But you can use the provided functions in two possible ways: +// stbtt_FindMatchingFont() will use *case-sensitive* comparisons on +// unicode-encoded names to try to find the font you want; +// you can run this before calling stbtt_InitFont() +// +// stbtt_GetFontNameString() lets you get any of the various strings +// from the file yourself and do your own comparisons on them. +// You have to have called stbtt_InitFont() first. + + +STBTT_DEF int stbtt_FindMatchingFont(const unsigned char *fontdata, const char *name, int flags); +// returns the offset (not index) of the font that matches, or -1 if none +// if you use STBTT_MACSTYLE_DONTCARE, use a font name like "Arial Bold". +// if you use any other flag, use a font name like "Arial"; this checks +// the 'macStyle' header field; i don't know if fonts set this consistently +#define STBTT_MACSTYLE_DONTCARE 0 +#define STBTT_MACSTYLE_BOLD 1 +#define STBTT_MACSTYLE_ITALIC 2 +#define STBTT_MACSTYLE_UNDERSCORE 4 +#define STBTT_MACSTYLE_NONE 8 // <= not same as 0, this makes us check the bitfield is 0 + +STBTT_DEF int stbtt_CompareUTF8toUTF16_bigendian(const char *s1, int len1, const char *s2, int len2); +// returns 1/0 whether the first string interpreted as utf8 is identical to +// the second string interpreted as big-endian utf16... useful for strings from next func + +STBTT_DEF const char *stbtt_GetFontNameString(const stbtt_fontinfo *font, int *length, int platformID, int encodingID, int languageID, int nameID); +// returns the string (which may be big-endian double byte, e.g. for unicode) +// and puts the length in bytes in *length. +// +// some of the values for the IDs are below; for more see the truetype spec: +// http://developer.apple.com/textfonts/TTRefMan/RM06/Chap6name.html +// http://www.microsoft.com/typography/otspec/name.htm + +enum { // platformID + STBTT_PLATFORM_ID_UNICODE =0, + STBTT_PLATFORM_ID_MAC =1, + STBTT_PLATFORM_ID_ISO =2, + STBTT_PLATFORM_ID_MICROSOFT =3 +}; + +enum { // encodingID for STBTT_PLATFORM_ID_UNICODE + STBTT_UNICODE_EID_UNICODE_1_0 =0, + STBTT_UNICODE_EID_UNICODE_1_1 =1, + STBTT_UNICODE_EID_ISO_10646 =2, + STBTT_UNICODE_EID_UNICODE_2_0_BMP=3, + STBTT_UNICODE_EID_UNICODE_2_0_FULL=4 +}; + +enum { // encodingID for STBTT_PLATFORM_ID_MICROSOFT + STBTT_MS_EID_SYMBOL =0, + STBTT_MS_EID_UNICODE_BMP =1, + STBTT_MS_EID_SHIFTJIS =2, + STBTT_MS_EID_UNICODE_FULL =10 +}; + +enum { // encodingID for STBTT_PLATFORM_ID_MAC; same as Script Manager codes + STBTT_MAC_EID_ROMAN =0, STBTT_MAC_EID_ARABIC =4, + STBTT_MAC_EID_JAPANESE =1, STBTT_MAC_EID_HEBREW =5, + STBTT_MAC_EID_CHINESE_TRAD =2, STBTT_MAC_EID_GREEK =6, + STBTT_MAC_EID_KOREAN =3, STBTT_MAC_EID_RUSSIAN =7 +}; + +enum { // languageID for STBTT_PLATFORM_ID_MICROSOFT; same as LCID... + // problematic because there are e.g. 16 english LCIDs and 16 arabic LCIDs + STBTT_MS_LANG_ENGLISH =0x0409, STBTT_MS_LANG_ITALIAN =0x0410, + STBTT_MS_LANG_CHINESE =0x0804, STBTT_MS_LANG_JAPANESE =0x0411, + STBTT_MS_LANG_DUTCH =0x0413, STBTT_MS_LANG_KOREAN =0x0412, + STBTT_MS_LANG_FRENCH =0x040c, STBTT_MS_LANG_RUSSIAN =0x0419, + STBTT_MS_LANG_GERMAN =0x0407, STBTT_MS_LANG_SPANISH =0x0409, + STBTT_MS_LANG_HEBREW =0x040d, STBTT_MS_LANG_SWEDISH =0x041D +}; + +enum { // languageID for STBTT_PLATFORM_ID_MAC + STBTT_MAC_LANG_ENGLISH =0 , STBTT_MAC_LANG_JAPANESE =11, + STBTT_MAC_LANG_ARABIC =12, STBTT_MAC_LANG_KOREAN =23, + STBTT_MAC_LANG_DUTCH =4 , STBTT_MAC_LANG_RUSSIAN =32, + STBTT_MAC_LANG_FRENCH =1 , STBTT_MAC_LANG_SPANISH =6 , + STBTT_MAC_LANG_GERMAN =2 , STBTT_MAC_LANG_SWEDISH =5 , + STBTT_MAC_LANG_HEBREW =10, STBTT_MAC_LANG_CHINESE_SIMPLIFIED =33, + STBTT_MAC_LANG_ITALIAN =3 , STBTT_MAC_LANG_CHINESE_TRAD =19 +}; + +#ifdef __cplusplus +} +#endif + +#endif // __STB_INCLUDE_STB_TRUETYPE_H__ + +/////////////////////////////////////////////////////////////////////////////// +/////////////////////////////////////////////////////////////////////////////// +//// +//// IMPLEMENTATION +//// +//// + +#ifdef STB_TRUETYPE_IMPLEMENTATION + +#ifndef STBTT_MAX_OVERSAMPLE +#define STBTT_MAX_OVERSAMPLE 8 +#endif + +#if STBTT_MAX_OVERSAMPLE > 255 +#error "STBTT_MAX_OVERSAMPLE cannot be > 255" +#endif + +typedef int stbtt__test_oversample_pow2[(STBTT_MAX_OVERSAMPLE & (STBTT_MAX_OVERSAMPLE-1)) == 0 ? 1 : -1]; + +#ifndef STBTT_RASTERIZER_VERSION +#define STBTT_RASTERIZER_VERSION 2 +#endif + +#ifdef _MSC_VER +#define STBTT__NOTUSED(v) (void)(v) +#else +#define STBTT__NOTUSED(v) (void)sizeof(v) +#endif + +////////////////////////////////////////////////////////////////////////// +// +// stbtt__buf helpers to parse data from file +// + +static stbtt_uint8 stbtt__buf_get8(stbtt__buf *b) +{ + if (b->cursor >= b->size) + return 0; + return b->data[b->cursor++]; +} + +static stbtt_uint8 stbtt__buf_peek8(stbtt__buf *b) +{ + if (b->cursor >= b->size) + return 0; + return b->data[b->cursor]; +} + +static void stbtt__buf_seek(stbtt__buf *b, int o) +{ + STBTT_assert(!(o > b->size || o < 0)); + b->cursor = (o > b->size || o < 0) ? b->size : o; +} + +static void stbtt__buf_skip(stbtt__buf *b, int o) +{ + stbtt__buf_seek(b, b->cursor + o); +} + +static stbtt_uint32 stbtt__buf_get(stbtt__buf *b, int n) +{ + stbtt_uint32 v = 0; + int i; + STBTT_assert(n >= 1 && n <= 4); + for (i = 0; i < n; i++) + v = (v << 8) | stbtt__buf_get8(b); + return v; +} + +static stbtt__buf stbtt__new_buf(const void *p, size_t size) +{ + stbtt__buf r; + STBTT_assert(size < 0x40000000); + r.data = (stbtt_uint8*) p; + r.size = (int) size; + r.cursor = 0; + return r; +} + +#define stbtt__buf_get16(b) stbtt__buf_get((b), 2) +#define stbtt__buf_get32(b) stbtt__buf_get((b), 4) + +static stbtt__buf stbtt__buf_range(const stbtt__buf *b, int o, int s) +{ + stbtt__buf r = stbtt__new_buf(NULL, 0); + if (o < 0 || s < 0 || o > b->size || s > b->size - o) return r; + r.data = b->data + o; + r.size = s; + return r; +} + +static stbtt__buf stbtt__cff_get_index(stbtt__buf *b) +{ + int count, start, offsize; + start = b->cursor; + count = stbtt__buf_get16(b); + if (count) { + offsize = stbtt__buf_get8(b); + STBTT_assert(offsize >= 1 && offsize <= 4); + stbtt__buf_skip(b, offsize * count); + stbtt__buf_skip(b, stbtt__buf_get(b, offsize) - 1); + } + return stbtt__buf_range(b, start, b->cursor - start); +} + +static stbtt_uint32 stbtt__cff_int(stbtt__buf *b) +{ + int b0 = stbtt__buf_get8(b); + if (b0 >= 32 && b0 <= 246) return b0 - 139; + else if (b0 >= 247 && b0 <= 250) return (b0 - 247)*256 + stbtt__buf_get8(b) + 108; + else if (b0 >= 251 && b0 <= 254) return -(b0 - 251)*256 - stbtt__buf_get8(b) - 108; + else if (b0 == 28) return stbtt__buf_get16(b); + else if (b0 == 29) return stbtt__buf_get32(b); + STBTT_assert(0); + return 0; +} + +static void stbtt__cff_skip_operand(stbtt__buf *b) { + int v, b0 = stbtt__buf_peek8(b); + STBTT_assert(b0 >= 28); + if (b0 == 30) { + stbtt__buf_skip(b, 1); + while (b->cursor < b->size) { + v = stbtt__buf_get8(b); + if ((v & 0xF) == 0xF || (v >> 4) == 0xF) + break; + } + } else { + stbtt__cff_int(b); + } +} + +static stbtt__buf stbtt__dict_get(stbtt__buf *b, int key) +{ + stbtt__buf_seek(b, 0); + while (b->cursor < b->size) { + int start = b->cursor, end, op; + while (stbtt__buf_peek8(b) >= 28) + stbtt__cff_skip_operand(b); + end = b->cursor; + op = stbtt__buf_get8(b); + if (op == 12) op = stbtt__buf_get8(b) | 0x100; + if (op == key) return stbtt__buf_range(b, start, end-start); + } + return stbtt__buf_range(b, 0, 0); +} + +static void stbtt__dict_get_ints(stbtt__buf *b, int key, int outcount, stbtt_uint32 *out) +{ + int i; + stbtt__buf operands = stbtt__dict_get(b, key); + for (i = 0; i < outcount && operands.cursor < operands.size; i++) + out[i] = stbtt__cff_int(&operands); +} + +static int stbtt__cff_index_count(stbtt__buf *b) +{ + stbtt__buf_seek(b, 0); + return stbtt__buf_get16(b); +} + +static stbtt__buf stbtt__cff_index_get(stbtt__buf b, int i) +{ + int count, offsize, start, end; + stbtt__buf_seek(&b, 0); + count = stbtt__buf_get16(&b); + offsize = stbtt__buf_get8(&b); + STBTT_assert(i >= 0 && i < count); + STBTT_assert(offsize >= 1 && offsize <= 4); + stbtt__buf_skip(&b, i*offsize); + start = stbtt__buf_get(&b, offsize); + end = stbtt__buf_get(&b, offsize); + return stbtt__buf_range(&b, 2+(count+1)*offsize+start, end - start); +} + +////////////////////////////////////////////////////////////////////////// +// +// accessors to parse data from file +// + +// on platforms that don't allow misaligned reads, if we want to allow +// truetype fonts that aren't padded to alignment, define ALLOW_UNALIGNED_TRUETYPE + +#define ttBYTE(p) (* (stbtt_uint8 *) (p)) +#define ttCHAR(p) (* (stbtt_int8 *) (p)) +#define ttFixed(p) ttLONG(p) + +static stbtt_uint16 ttUSHORT(stbtt_uint8 *p) { return p[0]*256 + p[1]; } +static stbtt_int16 ttSHORT(stbtt_uint8 *p) { return p[0]*256 + p[1]; } +static stbtt_uint32 ttULONG(stbtt_uint8 *p) { return (p[0]<<24) + (p[1]<<16) + (p[2]<<8) + p[3]; } +static stbtt_int32 ttLONG(stbtt_uint8 *p) { return (p[0]<<24) + (p[1]<<16) + (p[2]<<8) + p[3]; } + +#define stbtt_tag4(p,c0,c1,c2,c3) ((p)[0] == (c0) && (p)[1] == (c1) && (p)[2] == (c2) && (p)[3] == (c3)) +#define stbtt_tag(p,str) stbtt_tag4(p,str[0],str[1],str[2],str[3]) + +static int stbtt__isfont(stbtt_uint8 *font) +{ + // check the version number + if (stbtt_tag4(font, '1',0,0,0)) return 1; // TrueType 1 + if (stbtt_tag(font, "typ1")) return 1; // TrueType with type 1 font -- we don't support this! + if (stbtt_tag(font, "OTTO")) return 1; // OpenType with CFF + if (stbtt_tag4(font, 0,1,0,0)) return 1; // OpenType 1.0 + if (stbtt_tag(font, "true")) return 1; // Apple specification for TrueType fonts + return 0; +} + +// @OPTIMIZE: binary search +static stbtt_uint32 stbtt__find_table(stbtt_uint8 *data, stbtt_uint32 fontstart, const char *tag) +{ + stbtt_int32 num_tables = ttUSHORT(data+fontstart+4); + stbtt_uint32 tabledir = fontstart + 12; + stbtt_int32 i; + for (i=0; i < num_tables; ++i) { + stbtt_uint32 loc = tabledir + 16*i; + if (stbtt_tag(data+loc+0, tag)) + return ttULONG(data+loc+8); + } + return 0; +} + +static int stbtt_GetFontOffsetForIndex_internal(unsigned char *font_collection, int index) +{ + // if it's just a font, there's only one valid index + if (stbtt__isfont(font_collection)) + return index == 0 ? 0 : -1; + + // check if it's a TTC + if (stbtt_tag(font_collection, "ttcf")) { + // version 1? + if (ttULONG(font_collection+4) == 0x00010000 || ttULONG(font_collection+4) == 0x00020000) { + stbtt_int32 n = ttLONG(font_collection+8); + if (index >= n) + return -1; + return ttULONG(font_collection+12+index*4); + } + } + return -1; +} + +static int stbtt_GetNumberOfFonts_internal(unsigned char *font_collection) +{ + // if it's just a font, there's only one valid font + if (stbtt__isfont(font_collection)) + return 1; + + // check if it's a TTC + if (stbtt_tag(font_collection, "ttcf")) { + // version 1? + if (ttULONG(font_collection+4) == 0x00010000 || ttULONG(font_collection+4) == 0x00020000) { + return ttLONG(font_collection+8); + } + } + return 0; +} + +static stbtt__buf stbtt__get_subrs(stbtt__buf cff, stbtt__buf fontdict) +{ + stbtt_uint32 subrsoff = 0, private_loc[2] = { 0, 0 }; + stbtt__buf pdict; + stbtt__dict_get_ints(&fontdict, 18, 2, private_loc); + if (!private_loc[1] || !private_loc[0]) return stbtt__new_buf(NULL, 0); + pdict = stbtt__buf_range(&cff, private_loc[1], private_loc[0]); + stbtt__dict_get_ints(&pdict, 19, 1, &subrsoff); + if (!subrsoff) return stbtt__new_buf(NULL, 0); + stbtt__buf_seek(&cff, private_loc[1]+subrsoff); + return stbtt__cff_get_index(&cff); +} + +// since most people won't use this, find this table the first time it's needed +static int stbtt__get_svg(stbtt_fontinfo *info) +{ + stbtt_uint32 t; + if (info->svg < 0) { + t = stbtt__find_table(info->data, info->fontstart, "SVG "); + if (t) { + stbtt_uint32 offset = ttULONG(info->data + t + 2); + info->svg = t + offset; + } else { + info->svg = 0; + } + } + return info->svg; +} + +static int stbtt_InitFont_internal(stbtt_fontinfo *info, unsigned char *data, int fontstart) +{ + stbtt_uint32 cmap, t; + stbtt_int32 i,numTables; + + info->data = data; + info->fontstart = fontstart; + info->cff = stbtt__new_buf(NULL, 0); + + cmap = stbtt__find_table(data, fontstart, "cmap"); // required + info->loca = stbtt__find_table(data, fontstart, "loca"); // required + info->head = stbtt__find_table(data, fontstart, "head"); // required + info->glyf = stbtt__find_table(data, fontstart, "glyf"); // required + info->hhea = stbtt__find_table(data, fontstart, "hhea"); // required + info->hmtx = stbtt__find_table(data, fontstart, "hmtx"); // required + info->kern = stbtt__find_table(data, fontstart, "kern"); // not required + info->gpos = stbtt__find_table(data, fontstart, "GPOS"); // not required + + if (!cmap || !info->head || !info->hhea || !info->hmtx) + return 0; + if (info->glyf) { + // required for truetype + if (!info->loca) return 0; + } else { + // initialization for CFF / Type2 fonts (OTF) + stbtt__buf b, topdict, topdictidx; + stbtt_uint32 cstype = 2, charstrings = 0, fdarrayoff = 0, fdselectoff = 0; + stbtt_uint32 cff; + + cff = stbtt__find_table(data, fontstart, "CFF "); + if (!cff) return 0; + + info->fontdicts = stbtt__new_buf(NULL, 0); + info->fdselect = stbtt__new_buf(NULL, 0); + + // @TODO this should use size from table (not 512MB) + info->cff = stbtt__new_buf(data+cff, 512*1024*1024); + b = info->cff; + + // read the header + stbtt__buf_skip(&b, 2); + stbtt__buf_seek(&b, stbtt__buf_get8(&b)); // hdrsize + + // @TODO the name INDEX could list multiple fonts, + // but we just use the first one. + stbtt__cff_get_index(&b); // name INDEX + topdictidx = stbtt__cff_get_index(&b); + topdict = stbtt__cff_index_get(topdictidx, 0); + stbtt__cff_get_index(&b); // string INDEX + info->gsubrs = stbtt__cff_get_index(&b); + + stbtt__dict_get_ints(&topdict, 17, 1, &charstrings); + stbtt__dict_get_ints(&topdict, 0x100 | 6, 1, &cstype); + stbtt__dict_get_ints(&topdict, 0x100 | 36, 1, &fdarrayoff); + stbtt__dict_get_ints(&topdict, 0x100 | 37, 1, &fdselectoff); + info->subrs = stbtt__get_subrs(b, topdict); + + // we only support Type 2 charstrings + if (cstype != 2) return 0; + if (charstrings == 0) return 0; + + if (fdarrayoff) { + // looks like a CID font + if (!fdselectoff) return 0; + stbtt__buf_seek(&b, fdarrayoff); + info->fontdicts = stbtt__cff_get_index(&b); + info->fdselect = stbtt__buf_range(&b, fdselectoff, b.size-fdselectoff); + } + + stbtt__buf_seek(&b, charstrings); + info->charstrings = stbtt__cff_get_index(&b); + } + + t = stbtt__find_table(data, fontstart, "maxp"); + if (t) + info->numGlyphs = ttUSHORT(data+t+4); + else + info->numGlyphs = 0xffff; + + info->svg = -1; + + // find a cmap encoding table we understand *now* to avoid searching + // later. (todo: could make this installable) + // the same regardless of glyph. + numTables = ttUSHORT(data + cmap + 2); + info->index_map = 0; + for (i=0; i < numTables; ++i) { + stbtt_uint32 encoding_record = cmap + 4 + 8 * i; + // find an encoding we understand: + switch(ttUSHORT(data+encoding_record)) { + case STBTT_PLATFORM_ID_MICROSOFT: + switch (ttUSHORT(data+encoding_record+2)) { + case STBTT_MS_EID_UNICODE_BMP: + case STBTT_MS_EID_UNICODE_FULL: + // MS/Unicode + info->index_map = cmap + ttULONG(data+encoding_record+4); + break; + } + break; + case STBTT_PLATFORM_ID_UNICODE: + // Mac/iOS has these + // all the encodingIDs are unicode, so we don't bother to check it + info->index_map = cmap + ttULONG(data+encoding_record+4); + break; + } + } + if (info->index_map == 0) + return 0; + + info->indexToLocFormat = ttUSHORT(data+info->head + 50); + return 1; +} + +STBTT_DEF int stbtt_FindGlyphIndex(const stbtt_fontinfo *info, int unicode_codepoint) +{ + stbtt_uint8 *data = info->data; + stbtt_uint32 index_map = info->index_map; + + stbtt_uint16 format = ttUSHORT(data + index_map + 0); + if (format == 0) { // apple byte encoding + stbtt_int32 bytes = ttUSHORT(data + index_map + 2); + if (unicode_codepoint < bytes-6) + return ttBYTE(data + index_map + 6 + unicode_codepoint); + return 0; + } else if (format == 6) { + stbtt_uint32 first = ttUSHORT(data + index_map + 6); + stbtt_uint32 count = ttUSHORT(data + index_map + 8); + if ((stbtt_uint32) unicode_codepoint >= first && (stbtt_uint32) unicode_codepoint < first+count) + return ttUSHORT(data + index_map + 10 + (unicode_codepoint - first)*2); + return 0; + } else if (format == 2) { + STBTT_assert(0); // @TODO: high-byte mapping for japanese/chinese/korean + return 0; + } else if (format == 4) { // standard mapping for windows fonts: binary search collection of ranges + stbtt_uint16 segcount = ttUSHORT(data+index_map+6) >> 1; + stbtt_uint16 searchRange = ttUSHORT(data+index_map+8) >> 1; + stbtt_uint16 entrySelector = ttUSHORT(data+index_map+10); + stbtt_uint16 rangeShift = ttUSHORT(data+index_map+12) >> 1; + + // do a binary search of the segments + stbtt_uint32 endCount = index_map + 14; + stbtt_uint32 search = endCount; + + if (unicode_codepoint > 0xffff) + return 0; + + // they lie from endCount .. endCount + segCount + // but searchRange is the nearest power of two, so... + if (unicode_codepoint >= ttUSHORT(data + search + rangeShift*2)) + search += rangeShift*2; + + // now decrement to bias correctly to find smallest + search -= 2; + while (entrySelector) { + stbtt_uint16 end; + searchRange >>= 1; + end = ttUSHORT(data + search + searchRange*2); + if (unicode_codepoint > end) + search += searchRange*2; + --entrySelector; + } + search += 2; + + { + stbtt_uint16 offset, start; + stbtt_uint16 item = (stbtt_uint16) ((search - endCount) >> 1); + + STBTT_assert(unicode_codepoint <= ttUSHORT(data + endCount + 2*item)); + start = ttUSHORT(data + index_map + 14 + segcount*2 + 2 + 2*item); + if (unicode_codepoint < start) + return 0; + + offset = ttUSHORT(data + index_map + 14 + segcount*6 + 2 + 2*item); + if (offset == 0) + return (stbtt_uint16) (unicode_codepoint + ttSHORT(data + index_map + 14 + segcount*4 + 2 + 2*item)); + + return ttUSHORT(data + offset + (unicode_codepoint-start)*2 + index_map + 14 + segcount*6 + 2 + 2*item); + } + } else if (format == 12 || format == 13) { + stbtt_uint32 ngroups = ttULONG(data+index_map+12); + stbtt_int32 low,high; + low = 0; high = (stbtt_int32)ngroups; + // Binary search the right group. + while (low < high) { + stbtt_int32 mid = low + ((high-low) >> 1); // rounds down, so low <= mid < high + stbtt_uint32 start_char = ttULONG(data+index_map+16+mid*12); + stbtt_uint32 end_char = ttULONG(data+index_map+16+mid*12+4); + if ((stbtt_uint32) unicode_codepoint < start_char) + high = mid; + else if ((stbtt_uint32) unicode_codepoint > end_char) + low = mid+1; + else { + stbtt_uint32 start_glyph = ttULONG(data+index_map+16+mid*12+8); + if (format == 12) + return start_glyph + unicode_codepoint-start_char; + else // format == 13 + return start_glyph; + } + } + return 0; // not found + } + // @TODO + STBTT_assert(0); + return 0; +} + +STBTT_DEF int stbtt_GetCodepointShape(const stbtt_fontinfo *info, int unicode_codepoint, stbtt_vertex **vertices) +{ + return stbtt_GetGlyphShape(info, stbtt_FindGlyphIndex(info, unicode_codepoint), vertices); +} + +static void stbtt_setvertex(stbtt_vertex *v, stbtt_uint8 type, stbtt_int32 x, stbtt_int32 y, stbtt_int32 cx, stbtt_int32 cy) +{ + v->type = type; + v->x = (stbtt_int16) x; + v->y = (stbtt_int16) y; + v->cx = (stbtt_int16) cx; + v->cy = (stbtt_int16) cy; +} + +static int stbtt__GetGlyfOffset(const stbtt_fontinfo *info, int glyph_index) +{ + int g1,g2; + + STBTT_assert(!info->cff.size); + + if (glyph_index >= info->numGlyphs) return -1; // glyph index out of range + if (info->indexToLocFormat >= 2) return -1; // unknown index->glyph map format + + if (info->indexToLocFormat == 0) { + g1 = info->glyf + ttUSHORT(info->data + info->loca + glyph_index * 2) * 2; + g2 = info->glyf + ttUSHORT(info->data + info->loca + glyph_index * 2 + 2) * 2; + } else { + g1 = info->glyf + ttULONG (info->data + info->loca + glyph_index * 4); + g2 = info->glyf + ttULONG (info->data + info->loca + glyph_index * 4 + 4); + } + + return g1==g2 ? -1 : g1; // if length is 0, return -1 +} + +static int stbtt__GetGlyphInfoT2(const stbtt_fontinfo *info, int glyph_index, int *x0, int *y0, int *x1, int *y1); + +STBTT_DEF int stbtt_GetGlyphBox(const stbtt_fontinfo *info, int glyph_index, int *x0, int *y0, int *x1, int *y1) +{ + if (info->cff.size) { + stbtt__GetGlyphInfoT2(info, glyph_index, x0, y0, x1, y1); + } else { + int g = stbtt__GetGlyfOffset(info, glyph_index); + if (g < 0) return 0; + + if (x0) *x0 = ttSHORT(info->data + g + 2); + if (y0) *y0 = ttSHORT(info->data + g + 4); + if (x1) *x1 = ttSHORT(info->data + g + 6); + if (y1) *y1 = ttSHORT(info->data + g + 8); + } + return 1; +} + +STBTT_DEF int stbtt_GetCodepointBox(const stbtt_fontinfo *info, int codepoint, int *x0, int *y0, int *x1, int *y1) +{ + return stbtt_GetGlyphBox(info, stbtt_FindGlyphIndex(info,codepoint), x0,y0,x1,y1); +} + +STBTT_DEF int stbtt_IsGlyphEmpty(const stbtt_fontinfo *info, int glyph_index) +{ + stbtt_int16 numberOfContours; + int g; + if (info->cff.size) + return stbtt__GetGlyphInfoT2(info, glyph_index, NULL, NULL, NULL, NULL) == 0; + g = stbtt__GetGlyfOffset(info, glyph_index); + if (g < 0) return 1; + numberOfContours = ttSHORT(info->data + g); + return numberOfContours == 0; +} + +static int stbtt__close_shape(stbtt_vertex *vertices, int num_vertices, int was_off, int start_off, + stbtt_int32 sx, stbtt_int32 sy, stbtt_int32 scx, stbtt_int32 scy, stbtt_int32 cx, stbtt_int32 cy) +{ + if (start_off) { + if (was_off) + stbtt_setvertex(&vertices[num_vertices++], STBTT_vcurve, (cx+scx)>>1, (cy+scy)>>1, cx,cy); + stbtt_setvertex(&vertices[num_vertices++], STBTT_vcurve, sx,sy,scx,scy); + } else { + if (was_off) + stbtt_setvertex(&vertices[num_vertices++], STBTT_vcurve,sx,sy,cx,cy); + else + stbtt_setvertex(&vertices[num_vertices++], STBTT_vline,sx,sy,0,0); + } + return num_vertices; +} + +static int stbtt__GetGlyphShapeTT(const stbtt_fontinfo *info, int glyph_index, stbtt_vertex **pvertices) +{ + stbtt_int16 numberOfContours; + stbtt_uint8 *endPtsOfContours; + stbtt_uint8 *data = info->data; + stbtt_vertex *vertices=0; + int num_vertices=0; + int g = stbtt__GetGlyfOffset(info, glyph_index); + + *pvertices = NULL; + + if (g < 0) return 0; + + numberOfContours = ttSHORT(data + g); + + if (numberOfContours > 0) { + stbtt_uint8 flags=0,flagcount; + stbtt_int32 ins, i,j=0,m,n, next_move, was_off=0, off, start_off=0; + stbtt_int32 x,y,cx,cy,sx,sy, scx,scy; + stbtt_uint8 *points; + endPtsOfContours = (data + g + 10); + ins = ttUSHORT(data + g + 10 + numberOfContours * 2); + points = data + g + 10 + numberOfContours * 2 + 2 + ins; + + n = 1+ttUSHORT(endPtsOfContours + numberOfContours*2-2); + + m = n + 2*numberOfContours; // a loose bound on how many vertices we might need + vertices = (stbtt_vertex *) STBTT_malloc(m * sizeof(vertices[0]), info->userdata); + if (vertices == 0) + return 0; + + next_move = 0; + flagcount=0; + + // in first pass, we load uninterpreted data into the allocated array + // above, shifted to the end of the array so we won't overwrite it when + // we create our final data starting from the front + + off = m - n; // starting offset for uninterpreted data, regardless of how m ends up being calculated + + // first load flags + + for (i=0; i < n; ++i) { + if (flagcount == 0) { + flags = *points++; + if (flags & 8) + flagcount = *points++; + } else + --flagcount; + vertices[off+i].type = flags; + } + + // now load x coordinates + x=0; + for (i=0; i < n; ++i) { + flags = vertices[off+i].type; + if (flags & 2) { + stbtt_int16 dx = *points++; + x += (flags & 16) ? dx : -dx; // ??? + } else { + if (!(flags & 16)) { + x = x + (stbtt_int16) (points[0]*256 + points[1]); + points += 2; + } + } + vertices[off+i].x = (stbtt_int16) x; + } + + // now load y coordinates + y=0; + for (i=0; i < n; ++i) { + flags = vertices[off+i].type; + if (flags & 4) { + stbtt_int16 dy = *points++; + y += (flags & 32) ? dy : -dy; // ??? + } else { + if (!(flags & 32)) { + y = y + (stbtt_int16) (points[0]*256 + points[1]); + points += 2; + } + } + vertices[off+i].y = (stbtt_int16) y; + } + + // now convert them to our format + num_vertices=0; + sx = sy = cx = cy = scx = scy = 0; + for (i=0; i < n; ++i) { + flags = vertices[off+i].type; + x = (stbtt_int16) vertices[off+i].x; + y = (stbtt_int16) vertices[off+i].y; + + if (next_move == i) { + if (i != 0) + num_vertices = stbtt__close_shape(vertices, num_vertices, was_off, start_off, sx,sy,scx,scy,cx,cy); + + // now start the new one + start_off = !(flags & 1); + if (start_off) { + // if we start off with an off-curve point, then when we need to find a point on the curve + // where we can start, and we need to save some state for when we wraparound. + scx = x; + scy = y; + if (!(vertices[off+i+1].type & 1)) { + // next point is also a curve point, so interpolate an on-point curve + sx = (x + (stbtt_int32) vertices[off+i+1].x) >> 1; + sy = (y + (stbtt_int32) vertices[off+i+1].y) >> 1; + } else { + // otherwise just use the next point as our start point + sx = (stbtt_int32) vertices[off+i+1].x; + sy = (stbtt_int32) vertices[off+i+1].y; + ++i; // we're using point i+1 as the starting point, so skip it + } + } else { + sx = x; + sy = y; + } + stbtt_setvertex(&vertices[num_vertices++], STBTT_vmove,sx,sy,0,0); + was_off = 0; + next_move = 1 + ttUSHORT(endPtsOfContours+j*2); + ++j; + } else { + if (!(flags & 1)) { // if it's a curve + if (was_off) // two off-curve control points in a row means interpolate an on-curve midpoint + stbtt_setvertex(&vertices[num_vertices++], STBTT_vcurve, (cx+x)>>1, (cy+y)>>1, cx, cy); + cx = x; + cy = y; + was_off = 1; + } else { + if (was_off) + stbtt_setvertex(&vertices[num_vertices++], STBTT_vcurve, x,y, cx, cy); + else + stbtt_setvertex(&vertices[num_vertices++], STBTT_vline, x,y,0,0); + was_off = 0; + } + } + } + num_vertices = stbtt__close_shape(vertices, num_vertices, was_off, start_off, sx,sy,scx,scy,cx,cy); + } else if (numberOfContours < 0) { + // Compound shapes. + int more = 1; + stbtt_uint8 *comp = data + g + 10; + num_vertices = 0; + vertices = 0; + while (more) { + stbtt_uint16 flags, gidx; + int comp_num_verts = 0, i; + stbtt_vertex *comp_verts = 0, *tmp = 0; + float mtx[6] = {1,0,0,1,0,0}, m, n; + + flags = ttSHORT(comp); comp+=2; + gidx = ttSHORT(comp); comp+=2; + + if (flags & 2) { // XY values + if (flags & 1) { // shorts + mtx[4] = ttSHORT(comp); comp+=2; + mtx[5] = ttSHORT(comp); comp+=2; + } else { + mtx[4] = ttCHAR(comp); comp+=1; + mtx[5] = ttCHAR(comp); comp+=1; + } + } + else { + // @TODO handle matching point + STBTT_assert(0); + } + if (flags & (1<<3)) { // WE_HAVE_A_SCALE + mtx[0] = mtx[3] = ttSHORT(comp)/16384.0f; comp+=2; + mtx[1] = mtx[2] = 0; + } else if (flags & (1<<6)) { // WE_HAVE_AN_X_AND_YSCALE + mtx[0] = ttSHORT(comp)/16384.0f; comp+=2; + mtx[1] = mtx[2] = 0; + mtx[3] = ttSHORT(comp)/16384.0f; comp+=2; + } else if (flags & (1<<7)) { // WE_HAVE_A_TWO_BY_TWO + mtx[0] = ttSHORT(comp)/16384.0f; comp+=2; + mtx[1] = ttSHORT(comp)/16384.0f; comp+=2; + mtx[2] = ttSHORT(comp)/16384.0f; comp+=2; + mtx[3] = ttSHORT(comp)/16384.0f; comp+=2; + } + + // Find transformation scales. + m = (float) STBTT_sqrt(mtx[0]*mtx[0] + mtx[1]*mtx[1]); + n = (float) STBTT_sqrt(mtx[2]*mtx[2] + mtx[3]*mtx[3]); + + // Get indexed glyph. + comp_num_verts = stbtt_GetGlyphShape(info, gidx, &comp_verts); + if (comp_num_verts > 0) { + // Transform vertices. + for (i = 0; i < comp_num_verts; ++i) { + stbtt_vertex* v = &comp_verts[i]; + stbtt_vertex_type x,y; + x=v->x; y=v->y; + v->x = (stbtt_vertex_type)(m * (mtx[0]*x + mtx[2]*y + mtx[4])); + v->y = (stbtt_vertex_type)(n * (mtx[1]*x + mtx[3]*y + mtx[5])); + x=v->cx; y=v->cy; + v->cx = (stbtt_vertex_type)(m * (mtx[0]*x + mtx[2]*y + mtx[4])); + v->cy = (stbtt_vertex_type)(n * (mtx[1]*x + mtx[3]*y + mtx[5])); + } + // Append vertices. + tmp = (stbtt_vertex*)STBTT_malloc((num_vertices+comp_num_verts)*sizeof(stbtt_vertex), info->userdata); + if (!tmp) { + if (vertices) STBTT_free(vertices, info->userdata); + if (comp_verts) STBTT_free(comp_verts, info->userdata); + return 0; + } + if (num_vertices > 0) STBTT_memcpy(tmp, vertices, num_vertices*sizeof(stbtt_vertex)); + STBTT_memcpy(tmp+num_vertices, comp_verts, comp_num_verts*sizeof(stbtt_vertex)); + if (vertices) STBTT_free(vertices, info->userdata); + vertices = tmp; + STBTT_free(comp_verts, info->userdata); + num_vertices += comp_num_verts; + } + // More components ? + more = flags & (1<<5); + } + } else { + // numberOfCounters == 0, do nothing + } + + *pvertices = vertices; + return num_vertices; +} + +typedef struct +{ + int bounds; + int started; + float first_x, first_y; + float x, y; + stbtt_int32 min_x, max_x, min_y, max_y; + + stbtt_vertex *pvertices; + int num_vertices; +} stbtt__csctx; + +#define STBTT__CSCTX_INIT(bounds) {bounds,0, 0,0, 0,0, 0,0,0,0, NULL, 0} + +static void stbtt__track_vertex(stbtt__csctx *c, stbtt_int32 x, stbtt_int32 y) +{ + if (x > c->max_x || !c->started) c->max_x = x; + if (y > c->max_y || !c->started) c->max_y = y; + if (x < c->min_x || !c->started) c->min_x = x; + if (y < c->min_y || !c->started) c->min_y = y; + c->started = 1; +} + +static void stbtt__csctx_v(stbtt__csctx *c, stbtt_uint8 type, stbtt_int32 x, stbtt_int32 y, stbtt_int32 cx, stbtt_int32 cy, stbtt_int32 cx1, stbtt_int32 cy1) +{ + if (c->bounds) { + stbtt__track_vertex(c, x, y); + if (type == STBTT_vcubic) { + stbtt__track_vertex(c, cx, cy); + stbtt__track_vertex(c, cx1, cy1); + } + } else { + stbtt_setvertex(&c->pvertices[c->num_vertices], type, x, y, cx, cy); + c->pvertices[c->num_vertices].cx1 = (stbtt_int16) cx1; + c->pvertices[c->num_vertices].cy1 = (stbtt_int16) cy1; + } + c->num_vertices++; +} + +static void stbtt__csctx_close_shape(stbtt__csctx *ctx) +{ + if (ctx->first_x != ctx->x || ctx->first_y != ctx->y) + stbtt__csctx_v(ctx, STBTT_vline, (int)ctx->first_x, (int)ctx->first_y, 0, 0, 0, 0); +} + +static void stbtt__csctx_rmove_to(stbtt__csctx *ctx, float dx, float dy) +{ + stbtt__csctx_close_shape(ctx); + ctx->first_x = ctx->x = ctx->x + dx; + ctx->first_y = ctx->y = ctx->y + dy; + stbtt__csctx_v(ctx, STBTT_vmove, (int)ctx->x, (int)ctx->y, 0, 0, 0, 0); +} + +static void stbtt__csctx_rline_to(stbtt__csctx *ctx, float dx, float dy) +{ + ctx->x += dx; + ctx->y += dy; + stbtt__csctx_v(ctx, STBTT_vline, (int)ctx->x, (int)ctx->y, 0, 0, 0, 0); +} + +static void stbtt__csctx_rccurve_to(stbtt__csctx *ctx, float dx1, float dy1, float dx2, float dy2, float dx3, float dy3) +{ + float cx1 = ctx->x + dx1; + float cy1 = ctx->y + dy1; + float cx2 = cx1 + dx2; + float cy2 = cy1 + dy2; + ctx->x = cx2 + dx3; + ctx->y = cy2 + dy3; + stbtt__csctx_v(ctx, STBTT_vcubic, (int)ctx->x, (int)ctx->y, (int)cx1, (int)cy1, (int)cx2, (int)cy2); +} + +static stbtt__buf stbtt__get_subr(stbtt__buf idx, int n) +{ + int count = stbtt__cff_index_count(&idx); + int bias = 107; + if (count >= 33900) + bias = 32768; + else if (count >= 1240) + bias = 1131; + n += bias; + if (n < 0 || n >= count) + return stbtt__new_buf(NULL, 0); + return stbtt__cff_index_get(idx, n); +} + +static stbtt__buf stbtt__cid_get_glyph_subrs(const stbtt_fontinfo *info, int glyph_index) +{ + stbtt__buf fdselect = info->fdselect; + int nranges, start, end, v, fmt, fdselector = -1, i; + + stbtt__buf_seek(&fdselect, 0); + fmt = stbtt__buf_get8(&fdselect); + if (fmt == 0) { + // untested + stbtt__buf_skip(&fdselect, glyph_index); + fdselector = stbtt__buf_get8(&fdselect); + } else if (fmt == 3) { + nranges = stbtt__buf_get16(&fdselect); + start = stbtt__buf_get16(&fdselect); + for (i = 0; i < nranges; i++) { + v = stbtt__buf_get8(&fdselect); + end = stbtt__buf_get16(&fdselect); + if (glyph_index >= start && glyph_index < end) { + fdselector = v; + break; + } + start = end; + } + } + if (fdselector == -1) stbtt__new_buf(NULL, 0); + return stbtt__get_subrs(info->cff, stbtt__cff_index_get(info->fontdicts, fdselector)); +} + +static int stbtt__run_charstring(const stbtt_fontinfo *info, int glyph_index, stbtt__csctx *c) +{ + int in_header = 1, maskbits = 0, subr_stack_height = 0, sp = 0, v, i, b0; + int has_subrs = 0, clear_stack; + float s[48]; + stbtt__buf subr_stack[10], subrs = info->subrs, b; + float f; + +#define STBTT__CSERR(s) (0) + + // this currently ignores the initial width value, which isn't needed if we have hmtx + b = stbtt__cff_index_get(info->charstrings, glyph_index); + while (b.cursor < b.size) { + i = 0; + clear_stack = 1; + b0 = stbtt__buf_get8(&b); + switch (b0) { + // @TODO implement hinting + case 0x13: // hintmask + case 0x14: // cntrmask + if (in_header) + maskbits += (sp / 2); // implicit "vstem" + in_header = 0; + stbtt__buf_skip(&b, (maskbits + 7) / 8); + break; + + case 0x01: // hstem + case 0x03: // vstem + case 0x12: // hstemhm + case 0x17: // vstemhm + maskbits += (sp / 2); + break; + + case 0x15: // rmoveto + in_header = 0; + if (sp < 2) return STBTT__CSERR("rmoveto stack"); + stbtt__csctx_rmove_to(c, s[sp-2], s[sp-1]); + break; + case 0x04: // vmoveto + in_header = 0; + if (sp < 1) return STBTT__CSERR("vmoveto stack"); + stbtt__csctx_rmove_to(c, 0, s[sp-1]); + break; + case 0x16: // hmoveto + in_header = 0; + if (sp < 1) return STBTT__CSERR("hmoveto stack"); + stbtt__csctx_rmove_to(c, s[sp-1], 0); + break; + + case 0x05: // rlineto + if (sp < 2) return STBTT__CSERR("rlineto stack"); + for (; i + 1 < sp; i += 2) + stbtt__csctx_rline_to(c, s[i], s[i+1]); + break; + + // hlineto/vlineto and vhcurveto/hvcurveto alternate horizontal and vertical + // starting from a different place. + + case 0x07: // vlineto + if (sp < 1) return STBTT__CSERR("vlineto stack"); + goto vlineto; + case 0x06: // hlineto + if (sp < 1) return STBTT__CSERR("hlineto stack"); + for (;;) { + if (i >= sp) break; + stbtt__csctx_rline_to(c, s[i], 0); + i++; + vlineto: + if (i >= sp) break; + stbtt__csctx_rline_to(c, 0, s[i]); + i++; + } + break; + + case 0x1F: // hvcurveto + if (sp < 4) return STBTT__CSERR("hvcurveto stack"); + goto hvcurveto; + case 0x1E: // vhcurveto + if (sp < 4) return STBTT__CSERR("vhcurveto stack"); + for (;;) { + if (i + 3 >= sp) break; + stbtt__csctx_rccurve_to(c, 0, s[i], s[i+1], s[i+2], s[i+3], (sp - i == 5) ? s[i + 4] : 0.0f); + i += 4; + hvcurveto: + if (i + 3 >= sp) break; + stbtt__csctx_rccurve_to(c, s[i], 0, s[i+1], s[i+2], (sp - i == 5) ? s[i+4] : 0.0f, s[i+3]); + i += 4; + } + break; + + case 0x08: // rrcurveto + if (sp < 6) return STBTT__CSERR("rcurveline stack"); + for (; i + 5 < sp; i += 6) + stbtt__csctx_rccurve_to(c, s[i], s[i+1], s[i+2], s[i+3], s[i+4], s[i+5]); + break; + + case 0x18: // rcurveline + if (sp < 8) return STBTT__CSERR("rcurveline stack"); + for (; i + 5 < sp - 2; i += 6) + stbtt__csctx_rccurve_to(c, s[i], s[i+1], s[i+2], s[i+3], s[i+4], s[i+5]); + if (i + 1 >= sp) return STBTT__CSERR("rcurveline stack"); + stbtt__csctx_rline_to(c, s[i], s[i+1]); + break; + + case 0x19: // rlinecurve + if (sp < 8) return STBTT__CSERR("rlinecurve stack"); + for (; i + 1 < sp - 6; i += 2) + stbtt__csctx_rline_to(c, s[i], s[i+1]); + if (i + 5 >= sp) return STBTT__CSERR("rlinecurve stack"); + stbtt__csctx_rccurve_to(c, s[i], s[i+1], s[i+2], s[i+3], s[i+4], s[i+5]); + break; + + case 0x1A: // vvcurveto + case 0x1B: // hhcurveto + if (sp < 4) return STBTT__CSERR("(vv|hh)curveto stack"); + f = 0.0; + if (sp & 1) { f = s[i]; i++; } + for (; i + 3 < sp; i += 4) { + if (b0 == 0x1B) + stbtt__csctx_rccurve_to(c, s[i], f, s[i+1], s[i+2], s[i+3], 0.0); + else + stbtt__csctx_rccurve_to(c, f, s[i], s[i+1], s[i+2], 0.0, s[i+3]); + f = 0.0; + } + break; + + case 0x0A: // callsubr + if (!has_subrs) { + if (info->fdselect.size) + subrs = stbtt__cid_get_glyph_subrs(info, glyph_index); + has_subrs = 1; + } + // fallthrough + case 0x1D: // callgsubr + if (sp < 1) return STBTT__CSERR("call(g|)subr stack"); + v = (int) s[--sp]; + if (subr_stack_height >= 10) return STBTT__CSERR("recursion limit"); + subr_stack[subr_stack_height++] = b; + b = stbtt__get_subr(b0 == 0x0A ? subrs : info->gsubrs, v); + if (b.size == 0) return STBTT__CSERR("subr not found"); + b.cursor = 0; + clear_stack = 0; + break; + + case 0x0B: // return + if (subr_stack_height <= 0) return STBTT__CSERR("return outside subr"); + b = subr_stack[--subr_stack_height]; + clear_stack = 0; + break; + + case 0x0E: // endchar + stbtt__csctx_close_shape(c); + return 1; + + case 0x0C: { // two-byte escape + float dx1, dx2, dx3, dx4, dx5, dx6, dy1, dy2, dy3, dy4, dy5, dy6; + float dx, dy; + int b1 = stbtt__buf_get8(&b); + switch (b1) { + // @TODO These "flex" implementations ignore the flex-depth and resolution, + // and always draw beziers. + case 0x22: // hflex + if (sp < 7) return STBTT__CSERR("hflex stack"); + dx1 = s[0]; + dx2 = s[1]; + dy2 = s[2]; + dx3 = s[3]; + dx4 = s[4]; + dx5 = s[5]; + dx6 = s[6]; + stbtt__csctx_rccurve_to(c, dx1, 0, dx2, dy2, dx3, 0); + stbtt__csctx_rccurve_to(c, dx4, 0, dx5, -dy2, dx6, 0); + break; + + case 0x23: // flex + if (sp < 13) return STBTT__CSERR("flex stack"); + dx1 = s[0]; + dy1 = s[1]; + dx2 = s[2]; + dy2 = s[3]; + dx3 = s[4]; + dy3 = s[5]; + dx4 = s[6]; + dy4 = s[7]; + dx5 = s[8]; + dy5 = s[9]; + dx6 = s[10]; + dy6 = s[11]; + //fd is s[12] + stbtt__csctx_rccurve_to(c, dx1, dy1, dx2, dy2, dx3, dy3); + stbtt__csctx_rccurve_to(c, dx4, dy4, dx5, dy5, dx6, dy6); + break; + + case 0x24: // hflex1 + if (sp < 9) return STBTT__CSERR("hflex1 stack"); + dx1 = s[0]; + dy1 = s[1]; + dx2 = s[2]; + dy2 = s[3]; + dx3 = s[4]; + dx4 = s[5]; + dx5 = s[6]; + dy5 = s[7]; + dx6 = s[8]; + stbtt__csctx_rccurve_to(c, dx1, dy1, dx2, dy2, dx3, 0); + stbtt__csctx_rccurve_to(c, dx4, 0, dx5, dy5, dx6, -(dy1+dy2+dy5)); + break; + + case 0x25: // flex1 + if (sp < 11) return STBTT__CSERR("flex1 stack"); + dx1 = s[0]; + dy1 = s[1]; + dx2 = s[2]; + dy2 = s[3]; + dx3 = s[4]; + dy3 = s[5]; + dx4 = s[6]; + dy4 = s[7]; + dx5 = s[8]; + dy5 = s[9]; + dx6 = dy6 = s[10]; + dx = dx1+dx2+dx3+dx4+dx5; + dy = dy1+dy2+dy3+dy4+dy5; + if (STBTT_fabs(dx) > STBTT_fabs(dy)) + dy6 = -dy; + else + dx6 = -dx; + stbtt__csctx_rccurve_to(c, dx1, dy1, dx2, dy2, dx3, dy3); + stbtt__csctx_rccurve_to(c, dx4, dy4, dx5, dy5, dx6, dy6); + break; + + default: + return STBTT__CSERR("unimplemented"); + } + } break; + + default: + if (b0 != 255 && b0 != 28 && (b0 < 32 || b0 > 254)) + return STBTT__CSERR("reserved operator"); + + // push immediate + if (b0 == 255) { + f = (float)(stbtt_int32)stbtt__buf_get32(&b) / 0x10000; + } else { + stbtt__buf_skip(&b, -1); + f = (float)(stbtt_int16)stbtt__cff_int(&b); + } + if (sp >= 48) return STBTT__CSERR("push stack overflow"); + s[sp++] = f; + clear_stack = 0; + break; + } + if (clear_stack) sp = 0; + } + return STBTT__CSERR("no endchar"); + +#undef STBTT__CSERR +} + +static int stbtt__GetGlyphShapeT2(const stbtt_fontinfo *info, int glyph_index, stbtt_vertex **pvertices) +{ + // runs the charstring twice, once to count and once to output (to avoid realloc) + stbtt__csctx count_ctx = STBTT__CSCTX_INIT(1); + stbtt__csctx output_ctx = STBTT__CSCTX_INIT(0); + if (stbtt__run_charstring(info, glyph_index, &count_ctx)) { + *pvertices = (stbtt_vertex*)STBTT_malloc(count_ctx.num_vertices*sizeof(stbtt_vertex), info->userdata); + output_ctx.pvertices = *pvertices; + if (stbtt__run_charstring(info, glyph_index, &output_ctx)) { + STBTT_assert(output_ctx.num_vertices == count_ctx.num_vertices); + return output_ctx.num_vertices; + } + } + *pvertices = NULL; + return 0; +} + +static int stbtt__GetGlyphInfoT2(const stbtt_fontinfo *info, int glyph_index, int *x0, int *y0, int *x1, int *y1) +{ + stbtt__csctx c = STBTT__CSCTX_INIT(1); + int r = stbtt__run_charstring(info, glyph_index, &c); + if (x0) *x0 = r ? c.min_x : 0; + if (y0) *y0 = r ? c.min_y : 0; + if (x1) *x1 = r ? c.max_x : 0; + if (y1) *y1 = r ? c.max_y : 0; + return r ? c.num_vertices : 0; +} + +STBTT_DEF int stbtt_GetGlyphShape(const stbtt_fontinfo *info, int glyph_index, stbtt_vertex **pvertices) +{ + if (!info->cff.size) + return stbtt__GetGlyphShapeTT(info, glyph_index, pvertices); + else + return stbtt__GetGlyphShapeT2(info, glyph_index, pvertices); +} + +STBTT_DEF void stbtt_GetGlyphHMetrics(const stbtt_fontinfo *info, int glyph_index, int *advanceWidth, int *leftSideBearing) +{ + stbtt_uint16 numOfLongHorMetrics = ttUSHORT(info->data+info->hhea + 34); + if (glyph_index < numOfLongHorMetrics) { + if (advanceWidth) *advanceWidth = ttSHORT(info->data + info->hmtx + 4*glyph_index); + if (leftSideBearing) *leftSideBearing = ttSHORT(info->data + info->hmtx + 4*glyph_index + 2); + } else { + if (advanceWidth) *advanceWidth = ttSHORT(info->data + info->hmtx + 4*(numOfLongHorMetrics-1)); + if (leftSideBearing) *leftSideBearing = ttSHORT(info->data + info->hmtx + 4*numOfLongHorMetrics + 2*(glyph_index - numOfLongHorMetrics)); + } +} + +STBTT_DEF int stbtt_GetKerningTableLength(const stbtt_fontinfo *info) +{ + stbtt_uint8 *data = info->data + info->kern; + + // we only look at the first table. it must be 'horizontal' and format 0. + if (!info->kern) + return 0; + if (ttUSHORT(data+2) < 1) // number of tables, need at least 1 + return 0; + if (ttUSHORT(data+8) != 1) // horizontal flag must be set in format + return 0; + + return ttUSHORT(data+10); +} + +STBTT_DEF int stbtt_GetKerningTable(const stbtt_fontinfo *info, stbtt_kerningentry* table, int table_length) +{ + stbtt_uint8 *data = info->data + info->kern; + int k, length; + + // we only look at the first table. it must be 'horizontal' and format 0. + if (!info->kern) + return 0; + if (ttUSHORT(data+2) < 1) // number of tables, need at least 1 + return 0; + if (ttUSHORT(data+8) != 1) // horizontal flag must be set in format + return 0; + + length = ttUSHORT(data+10); + if (table_length < length) + length = table_length; + + for (k = 0; k < length; k++) + { + table[k].glyph1 = ttUSHORT(data+18+(k*6)); + table[k].glyph2 = ttUSHORT(data+20+(k*6)); + table[k].advance = ttSHORT(data+22+(k*6)); + } + + return length; +} + +static int stbtt__GetGlyphKernInfoAdvance(const stbtt_fontinfo *info, int glyph1, int glyph2) +{ + stbtt_uint8 *data = info->data + info->kern; + stbtt_uint32 needle, straw; + int l, r, m; + + // we only look at the first table. it must be 'horizontal' and format 0. + if (!info->kern) + return 0; + if (ttUSHORT(data+2) < 1) // number of tables, need at least 1 + return 0; + if (ttUSHORT(data+8) != 1) // horizontal flag must be set in format + return 0; + + l = 0; + r = ttUSHORT(data+10) - 1; + needle = glyph1 << 16 | glyph2; + while (l <= r) { + m = (l + r) >> 1; + straw = ttULONG(data+18+(m*6)); // note: unaligned read + if (needle < straw) + r = m - 1; + else if (needle > straw) + l = m + 1; + else + return ttSHORT(data+22+(m*6)); + } + return 0; +} + +static stbtt_int32 stbtt__GetCoverageIndex(stbtt_uint8 *coverageTable, int glyph) +{ + stbtt_uint16 coverageFormat = ttUSHORT(coverageTable); + switch(coverageFormat) { + case 1: { + stbtt_uint16 glyphCount = ttUSHORT(coverageTable + 2); + + // Binary search. + stbtt_int32 l=0, r=glyphCount-1, m; + int straw, needle=glyph; + while (l <= r) { + stbtt_uint8 *glyphArray = coverageTable + 4; + stbtt_uint16 glyphID; + m = (l + r) >> 1; + glyphID = ttUSHORT(glyphArray + 2 * m); + straw = glyphID; + if (needle < straw) + r = m - 1; + else if (needle > straw) + l = m + 1; + else { + return m; + } + } + } break; + + case 2: { + stbtt_uint16 rangeCount = ttUSHORT(coverageTable + 2); + stbtt_uint8 *rangeArray = coverageTable + 4; + + // Binary search. + stbtt_int32 l=0, r=rangeCount-1, m; + int strawStart, strawEnd, needle=glyph; + while (l <= r) { + stbtt_uint8 *rangeRecord; + m = (l + r) >> 1; + rangeRecord = rangeArray + 6 * m; + strawStart = ttUSHORT(rangeRecord); + strawEnd = ttUSHORT(rangeRecord + 2); + if (needle < strawStart) + r = m - 1; + else if (needle > strawEnd) + l = m + 1; + else { + stbtt_uint16 startCoverageIndex = ttUSHORT(rangeRecord + 4); + return startCoverageIndex + glyph - strawStart; + } + } + } break; + + default: { + // There are no other cases. + STBTT_assert(0); + } break; + } + + return -1; +} + +static stbtt_int32 stbtt__GetGlyphClass(stbtt_uint8 *classDefTable, int glyph) +{ + stbtt_uint16 classDefFormat = ttUSHORT(classDefTable); + switch(classDefFormat) + { + case 1: { + stbtt_uint16 startGlyphID = ttUSHORT(classDefTable + 2); + stbtt_uint16 glyphCount = ttUSHORT(classDefTable + 4); + stbtt_uint8 *classDef1ValueArray = classDefTable + 6; + + if (glyph >= startGlyphID && glyph < startGlyphID + glyphCount) + return (stbtt_int32)ttUSHORT(classDef1ValueArray + 2 * (glyph - startGlyphID)); + + classDefTable = classDef1ValueArray + 2 * glyphCount; + } break; + + case 2: { + stbtt_uint16 classRangeCount = ttUSHORT(classDefTable + 2); + stbtt_uint8 *classRangeRecords = classDefTable + 4; + + // Binary search. + stbtt_int32 l=0, r=classRangeCount-1, m; + int strawStart, strawEnd, needle=glyph; + while (l <= r) { + stbtt_uint8 *classRangeRecord; + m = (l + r) >> 1; + classRangeRecord = classRangeRecords + 6 * m; + strawStart = ttUSHORT(classRangeRecord); + strawEnd = ttUSHORT(classRangeRecord + 2); + if (needle < strawStart) + r = m - 1; + else if (needle > strawEnd) + l = m + 1; + else + return (stbtt_int32)ttUSHORT(classRangeRecord + 4); + } + + classDefTable = classRangeRecords + 6 * classRangeCount; + } break; + + default: { + // There are no other cases. + STBTT_assert(0); + } break; + } + + return -1; +} + +// Define to STBTT_assert(x) if you want to break on unimplemented formats. +#define STBTT_GPOS_TODO_assert(x) + +static stbtt_int32 stbtt__GetGlyphGPOSInfoAdvance(const stbtt_fontinfo *info, int glyph1, int glyph2) +{ + stbtt_uint16 lookupListOffset; + stbtt_uint8 *lookupList; + stbtt_uint16 lookupCount; + stbtt_uint8 *data; + stbtt_int32 i; + + if (!info->gpos) return 0; + + data = info->data + info->gpos; + + if (ttUSHORT(data+0) != 1) return 0; // Major version 1 + if (ttUSHORT(data+2) != 0) return 0; // Minor version 0 + + lookupListOffset = ttUSHORT(data+8); + lookupList = data + lookupListOffset; + lookupCount = ttUSHORT(lookupList); + + for (i=0; i> 1; + pairValue = pairValueArray + (2 + valueRecordPairSizeInBytes) * m; + secondGlyph = ttUSHORT(pairValue); + straw = secondGlyph; + if (needle < straw) + r = m - 1; + else if (needle > straw) + l = m + 1; + else { + stbtt_int16 xAdvance = ttSHORT(pairValue + 2); + return xAdvance; + } + } + } break; + + case 2: { + stbtt_uint16 valueFormat1 = ttUSHORT(table + 4); + stbtt_uint16 valueFormat2 = ttUSHORT(table + 6); + + stbtt_uint16 classDef1Offset = ttUSHORT(table + 8); + stbtt_uint16 classDef2Offset = ttUSHORT(table + 10); + int glyph1class = stbtt__GetGlyphClass(table + classDef1Offset, glyph1); + int glyph2class = stbtt__GetGlyphClass(table + classDef2Offset, glyph2); + + stbtt_uint16 class1Count = ttUSHORT(table + 12); + stbtt_uint16 class2Count = ttUSHORT(table + 14); + STBTT_assert(glyph1class < class1Count); + STBTT_assert(glyph2class < class2Count); + + // TODO: Support more formats. + STBTT_GPOS_TODO_assert(valueFormat1 == 4); + if (valueFormat1 != 4) return 0; + STBTT_GPOS_TODO_assert(valueFormat2 == 0); + if (valueFormat2 != 0) return 0; + + if (glyph1class >= 0 && glyph1class < class1Count && glyph2class >= 0 && glyph2class < class2Count) { + stbtt_uint8 *class1Records = table + 16; + stbtt_uint8 *class2Records = class1Records + 2 * (glyph1class * class2Count); + stbtt_int16 xAdvance = ttSHORT(class2Records + 2 * glyph2class); + return xAdvance; + } + } break; + + default: { + // There are no other cases. + STBTT_assert(0); + break; + }; + } + } + break; + }; + + default: + // TODO: Implement other stuff. + break; + } + } + + return 0; +} + +STBTT_DEF int stbtt_GetGlyphKernAdvance(const stbtt_fontinfo *info, int g1, int g2) +{ + int xAdvance = 0; + + if (info->gpos) + xAdvance += stbtt__GetGlyphGPOSInfoAdvance(info, g1, g2); + else if (info->kern) + xAdvance += stbtt__GetGlyphKernInfoAdvance(info, g1, g2); + + return xAdvance; +} + +STBTT_DEF int stbtt_GetCodepointKernAdvance(const stbtt_fontinfo *info, int ch1, int ch2) +{ + if (!info->kern && !info->gpos) // if no kerning table, don't waste time looking up both codepoint->glyphs + return 0; + return stbtt_GetGlyphKernAdvance(info, stbtt_FindGlyphIndex(info,ch1), stbtt_FindGlyphIndex(info,ch2)); +} + +STBTT_DEF void stbtt_GetCodepointHMetrics(const stbtt_fontinfo *info, int codepoint, int *advanceWidth, int *leftSideBearing) +{ + stbtt_GetGlyphHMetrics(info, stbtt_FindGlyphIndex(info,codepoint), advanceWidth, leftSideBearing); +} + +STBTT_DEF void stbtt_GetFontVMetrics(const stbtt_fontinfo *info, int *ascent, int *descent, int *lineGap) +{ + if (ascent ) *ascent = ttSHORT(info->data+info->hhea + 4); + if (descent) *descent = ttSHORT(info->data+info->hhea + 6); + if (lineGap) *lineGap = ttSHORT(info->data+info->hhea + 8); +} + +STBTT_DEF int stbtt_GetFontVMetricsOS2(const stbtt_fontinfo *info, int *typoAscent, int *typoDescent, int *typoLineGap) +{ + int tab = stbtt__find_table(info->data, info->fontstart, "OS/2"); + if (!tab) + return 0; + if (typoAscent ) *typoAscent = ttSHORT(info->data+tab + 68); + if (typoDescent) *typoDescent = ttSHORT(info->data+tab + 70); + if (typoLineGap) *typoLineGap = ttSHORT(info->data+tab + 72); + return 1; +} + +STBTT_DEF void stbtt_GetFontBoundingBox(const stbtt_fontinfo *info, int *x0, int *y0, int *x1, int *y1) +{ + *x0 = ttSHORT(info->data + info->head + 36); + *y0 = ttSHORT(info->data + info->head + 38); + *x1 = ttSHORT(info->data + info->head + 40); + *y1 = ttSHORT(info->data + info->head + 42); +} + +STBTT_DEF float stbtt_ScaleForPixelHeight(const stbtt_fontinfo *info, float height) +{ + int fheight = ttSHORT(info->data + info->hhea + 4) - ttSHORT(info->data + info->hhea + 6); + return (float) height / fheight; +} + +STBTT_DEF float stbtt_ScaleForMappingEmToPixels(const stbtt_fontinfo *info, float pixels) +{ + int unitsPerEm = ttUSHORT(info->data + info->head + 18); + return pixels / unitsPerEm; +} + +STBTT_DEF void stbtt_FreeShape(const stbtt_fontinfo *info, stbtt_vertex *v) +{ + STBTT_free(v, info->userdata); +} + +STBTT_DEF stbtt_uint8 *stbtt_FindSVGDoc(const stbtt_fontinfo *info, int gl) +{ + int i; + stbtt_uint8 *data = info->data; + stbtt_uint8 *svg_doc_list = data + stbtt__get_svg((stbtt_fontinfo *) info); + + int numEntries = ttUSHORT(svg_doc_list); + stbtt_uint8 *svg_docs = svg_doc_list + 2; + + for(i=0; i= ttUSHORT(svg_doc)) && (gl <= ttUSHORT(svg_doc + 2))) + return svg_doc; + } + return 0; +} + +STBTT_DEF int stbtt_GetGlyphSVG(const stbtt_fontinfo *info, int gl, const char **svg) +{ + stbtt_uint8 *data = info->data; + stbtt_uint8 *svg_doc; + + if (info->svg == 0) + return 0; + + svg_doc = stbtt_FindSVGDoc(info, gl); + if (svg_doc != NULL) { + *svg = (char *) data + info->svg + ttULONG(svg_doc + 4); + return ttULONG(svg_doc + 8); + } else { + return 0; + } +} + +STBTT_DEF int stbtt_GetCodepointSVG(const stbtt_fontinfo *info, int unicode_codepoint, const char **svg) +{ + return stbtt_GetGlyphSVG(info, stbtt_FindGlyphIndex(info, unicode_codepoint), svg); +} + +////////////////////////////////////////////////////////////////////////////// +// +// antialiasing software rasterizer +// + +STBTT_DEF void stbtt_GetGlyphBitmapBoxSubpixel(const stbtt_fontinfo *font, int glyph, float scale_x, float scale_y,float shift_x, float shift_y, int *ix0, int *iy0, int *ix1, int *iy1) +{ + int x0=0,y0=0,x1,y1; // =0 suppresses compiler warning + if (!stbtt_GetGlyphBox(font, glyph, &x0,&y0,&x1,&y1)) { + // e.g. space character + if (ix0) *ix0 = 0; + if (iy0) *iy0 = 0; + if (ix1) *ix1 = 0; + if (iy1) *iy1 = 0; + } else { + // move to integral bboxes (treating pixels as little squares, what pixels get touched)? + if (ix0) *ix0 = STBTT_ifloor( x0 * scale_x + shift_x); + if (iy0) *iy0 = STBTT_ifloor(-y1 * scale_y + shift_y); + if (ix1) *ix1 = STBTT_iceil ( x1 * scale_x + shift_x); + if (iy1) *iy1 = STBTT_iceil (-y0 * scale_y + shift_y); + } +} + +STBTT_DEF void stbtt_GetGlyphBitmapBox(const stbtt_fontinfo *font, int glyph, float scale_x, float scale_y, int *ix0, int *iy0, int *ix1, int *iy1) +{ + stbtt_GetGlyphBitmapBoxSubpixel(font, glyph, scale_x, scale_y,0.0f,0.0f, ix0, iy0, ix1, iy1); +} + +STBTT_DEF void stbtt_GetCodepointBitmapBoxSubpixel(const stbtt_fontinfo *font, int codepoint, float scale_x, float scale_y, float shift_x, float shift_y, int *ix0, int *iy0, int *ix1, int *iy1) +{ + stbtt_GetGlyphBitmapBoxSubpixel(font, stbtt_FindGlyphIndex(font,codepoint), scale_x, scale_y,shift_x,shift_y, ix0,iy0,ix1,iy1); +} + +STBTT_DEF void stbtt_GetCodepointBitmapBox(const stbtt_fontinfo *font, int codepoint, float scale_x, float scale_y, int *ix0, int *iy0, int *ix1, int *iy1) +{ + stbtt_GetCodepointBitmapBoxSubpixel(font, codepoint, scale_x, scale_y,0.0f,0.0f, ix0,iy0,ix1,iy1); +} + +////////////////////////////////////////////////////////////////////////////// +// +// Rasterizer + +typedef struct stbtt__hheap_chunk +{ + struct stbtt__hheap_chunk *next; +} stbtt__hheap_chunk; + +typedef struct stbtt__hheap +{ + struct stbtt__hheap_chunk *head; + void *first_free; + int num_remaining_in_head_chunk; +} stbtt__hheap; + +static void *stbtt__hheap_alloc(stbtt__hheap *hh, size_t size, void *userdata) +{ + if (hh->first_free) { + void *p = hh->first_free; + hh->first_free = * (void **) p; + return p; + } else { + if (hh->num_remaining_in_head_chunk == 0) { + int count = (size < 32 ? 2000 : size < 128 ? 800 : 100); + stbtt__hheap_chunk *c = (stbtt__hheap_chunk *) STBTT_malloc(sizeof(stbtt__hheap_chunk) + size * count, userdata); + if (c == NULL) + return NULL; + c->next = hh->head; + hh->head = c; + hh->num_remaining_in_head_chunk = count; + } + --hh->num_remaining_in_head_chunk; + return (char *) (hh->head) + sizeof(stbtt__hheap_chunk) + size * hh->num_remaining_in_head_chunk; + } +} + +static void stbtt__hheap_free(stbtt__hheap *hh, void *p) +{ + *(void **) p = hh->first_free; + hh->first_free = p; +} + +static void stbtt__hheap_cleanup(stbtt__hheap *hh, void *userdata) +{ + stbtt__hheap_chunk *c = hh->head; + while (c) { + stbtt__hheap_chunk *n = c->next; + STBTT_free(c, userdata); + c = n; + } +} + +typedef struct stbtt__edge { + float x0,y0, x1,y1; + int invert; +} stbtt__edge; + + +typedef struct stbtt__active_edge +{ + struct stbtt__active_edge *next; + #if STBTT_RASTERIZER_VERSION==1 + int x,dx; + float ey; + int direction; + #elif STBTT_RASTERIZER_VERSION==2 + float fx,fdx,fdy; + float direction; + float sy; + float ey; + #else + #error "Unrecognized value of STBTT_RASTERIZER_VERSION" + #endif +} stbtt__active_edge; + +#if STBTT_RASTERIZER_VERSION == 1 +#define STBTT_FIXSHIFT 10 +#define STBTT_FIX (1 << STBTT_FIXSHIFT) +#define STBTT_FIXMASK (STBTT_FIX-1) + +static stbtt__active_edge *stbtt__new_active(stbtt__hheap *hh, stbtt__edge *e, int off_x, float start_point, void *userdata) +{ + stbtt__active_edge *z = (stbtt__active_edge *) stbtt__hheap_alloc(hh, sizeof(*z), userdata); + float dxdy = (e->x1 - e->x0) / (e->y1 - e->y0); + STBTT_assert(z != NULL); + if (!z) return z; + + // round dx down to avoid overshooting + if (dxdy < 0) + z->dx = -STBTT_ifloor(STBTT_FIX * -dxdy); + else + z->dx = STBTT_ifloor(STBTT_FIX * dxdy); + + z->x = STBTT_ifloor(STBTT_FIX * e->x0 + z->dx * (start_point - e->y0)); // use z->dx so when we offset later it's by the same amount + z->x -= off_x * STBTT_FIX; + + z->ey = e->y1; + z->next = 0; + z->direction = e->invert ? 1 : -1; + return z; +} +#elif STBTT_RASTERIZER_VERSION == 2 +static stbtt__active_edge *stbtt__new_active(stbtt__hheap *hh, stbtt__edge *e, int off_x, float start_point, void *userdata) +{ + stbtt__active_edge *z = (stbtt__active_edge *) stbtt__hheap_alloc(hh, sizeof(*z), userdata); + float dxdy = (e->x1 - e->x0) / (e->y1 - e->y0); + STBTT_assert(z != NULL); + //STBTT_assert(e->y0 <= start_point); + if (!z) return z; + z->fdx = dxdy; + z->fdy = dxdy != 0.0f ? (1.0f/dxdy) : 0.0f; + z->fx = e->x0 + dxdy * (start_point - e->y0); + z->fx -= off_x; + z->direction = e->invert ? 1.0f : -1.0f; + z->sy = e->y0; + z->ey = e->y1; + z->next = 0; + return z; +} +#else +#error "Unrecognized value of STBTT_RASTERIZER_VERSION" +#endif + +#if STBTT_RASTERIZER_VERSION == 1 +// note: this routine clips fills that extend off the edges... ideally this +// wouldn't happen, but it could happen if the truetype glyph bounding boxes +// are wrong, or if the user supplies a too-small bitmap +static void stbtt__fill_active_edges(unsigned char *scanline, int len, stbtt__active_edge *e, int max_weight) +{ + // non-zero winding fill + int x0=0, w=0; + + while (e) { + if (w == 0) { + // if we're currently at zero, we need to record the edge start point + x0 = e->x; w += e->direction; + } else { + int x1 = e->x; w += e->direction; + // if we went to zero, we need to draw + if (w == 0) { + int i = x0 >> STBTT_FIXSHIFT; + int j = x1 >> STBTT_FIXSHIFT; + + if (i < len && j >= 0) { + if (i == j) { + // x0,x1 are the same pixel, so compute combined coverage + scanline[i] = scanline[i] + (stbtt_uint8) ((x1 - x0) * max_weight >> STBTT_FIXSHIFT); + } else { + if (i >= 0) // add antialiasing for x0 + scanline[i] = scanline[i] + (stbtt_uint8) (((STBTT_FIX - (x0 & STBTT_FIXMASK)) * max_weight) >> STBTT_FIXSHIFT); + else + i = -1; // clip + + if (j < len) // add antialiasing for x1 + scanline[j] = scanline[j] + (stbtt_uint8) (((x1 & STBTT_FIXMASK) * max_weight) >> STBTT_FIXSHIFT); + else + j = len; // clip + + for (++i; i < j; ++i) // fill pixels between x0 and x1 + scanline[i] = scanline[i] + (stbtt_uint8) max_weight; + } + } + } + } + + e = e->next; + } +} + +static void stbtt__rasterize_sorted_edges(stbtt__bitmap *result, stbtt__edge *e, int n, int vsubsample, int off_x, int off_y, void *userdata) +{ + stbtt__hheap hh = { 0, 0, 0 }; + stbtt__active_edge *active = NULL; + int y,j=0; + int max_weight = (255 / vsubsample); // weight per vertical scanline + int s; // vertical subsample index + unsigned char scanline_data[512], *scanline; + + if (result->w > 512) + scanline = (unsigned char *) STBTT_malloc(result->w, userdata); + else + scanline = scanline_data; + + y = off_y * vsubsample; + e[n].y0 = (off_y + result->h) * (float) vsubsample + 1; + + while (j < result->h) { + STBTT_memset(scanline, 0, result->w); + for (s=0; s < vsubsample; ++s) { + // find center of pixel for this scanline + float scan_y = y + 0.5f; + stbtt__active_edge **step = &active; + + // update all active edges; + // remove all active edges that terminate before the center of this scanline + while (*step) { + stbtt__active_edge * z = *step; + if (z->ey <= scan_y) { + *step = z->next; // delete from list + STBTT_assert(z->direction); + z->direction = 0; + stbtt__hheap_free(&hh, z); + } else { + z->x += z->dx; // advance to position for current scanline + step = &((*step)->next); // advance through list + } + } + + // resort the list if needed + for(;;) { + int changed=0; + step = &active; + while (*step && (*step)->next) { + if ((*step)->x > (*step)->next->x) { + stbtt__active_edge *t = *step; + stbtt__active_edge *q = t->next; + + t->next = q->next; + q->next = t; + *step = q; + changed = 1; + } + step = &(*step)->next; + } + if (!changed) break; + } + + // insert all edges that start before the center of this scanline -- omit ones that also end on this scanline + while (e->y0 <= scan_y) { + if (e->y1 > scan_y) { + stbtt__active_edge *z = stbtt__new_active(&hh, e, off_x, scan_y, userdata); + if (z != NULL) { + // find insertion point + if (active == NULL) + active = z; + else if (z->x < active->x) { + // insert at front + z->next = active; + active = z; + } else { + // find thing to insert AFTER + stbtt__active_edge *p = active; + while (p->next && p->next->x < z->x) + p = p->next; + // at this point, p->next->x is NOT < z->x + z->next = p->next; + p->next = z; + } + } + } + ++e; + } + + // now process all active edges in XOR fashion + if (active) + stbtt__fill_active_edges(scanline, result->w, active, max_weight); + + ++y; + } + STBTT_memcpy(result->pixels + j * result->stride, scanline, result->w); + ++j; + } + + stbtt__hheap_cleanup(&hh, userdata); + + if (scanline != scanline_data) + STBTT_free(scanline, userdata); +} + +#elif STBTT_RASTERIZER_VERSION == 2 + +// the edge passed in here does not cross the vertical line at x or the vertical line at x+1 +// (i.e. it has already been clipped to those) +static void stbtt__handle_clipped_edge(float *scanline, int x, stbtt__active_edge *e, float x0, float y0, float x1, float y1) +{ + if (y0 == y1) return; + STBTT_assert(y0 < y1); + STBTT_assert(e->sy <= e->ey); + if (y0 > e->ey) return; + if (y1 < e->sy) return; + if (y0 < e->sy) { + x0 += (x1-x0) * (e->sy - y0) / (y1-y0); + y0 = e->sy; + } + if (y1 > e->ey) { + x1 += (x1-x0) * (e->ey - y1) / (y1-y0); + y1 = e->ey; + } + + if (x0 == x) + STBTT_assert(x1 <= x+1); + else if (x0 == x+1) + STBTT_assert(x1 >= x); + else if (x0 <= x) + STBTT_assert(x1 <= x); + else if (x0 >= x+1) + STBTT_assert(x1 >= x+1); + else + STBTT_assert(x1 >= x && x1 <= x+1); + + if (x0 <= x && x1 <= x) + scanline[x] += e->direction * (y1-y0); + else if (x0 >= x+1 && x1 >= x+1) + ; + else { + STBTT_assert(x0 >= x && x0 <= x+1 && x1 >= x && x1 <= x+1); + scanline[x] += e->direction * (y1-y0) * (1-((x0-x)+(x1-x))/2); // coverage = 1 - average x position + } +} + +static void stbtt__fill_active_edges_new(float *scanline, float *scanline_fill, int len, stbtt__active_edge *e, float y_top) +{ + float y_bottom = y_top+1; + + while (e) { + // brute force every pixel + + // compute intersection points with top & bottom + STBTT_assert(e->ey >= y_top); + + if (e->fdx == 0) { + float x0 = e->fx; + if (x0 < len) { + if (x0 >= 0) { + stbtt__handle_clipped_edge(scanline,(int) x0,e, x0,y_top, x0,y_bottom); + stbtt__handle_clipped_edge(scanline_fill-1,(int) x0+1,e, x0,y_top, x0,y_bottom); + } else { + stbtt__handle_clipped_edge(scanline_fill-1,0,e, x0,y_top, x0,y_bottom); + } + } + } else { + float x0 = e->fx; + float dx = e->fdx; + float xb = x0 + dx; + float x_top, x_bottom; + float sy0,sy1; + float dy = e->fdy; + STBTT_assert(e->sy <= y_bottom && e->ey >= y_top); + + // compute endpoints of line segment clipped to this scanline (if the + // line segment starts on this scanline. x0 is the intersection of the + // line with y_top, but that may be off the line segment. + if (e->sy > y_top) { + x_top = x0 + dx * (e->sy - y_top); + sy0 = e->sy; + } else { + x_top = x0; + sy0 = y_top; + } + if (e->ey < y_bottom) { + x_bottom = x0 + dx * (e->ey - y_top); + sy1 = e->ey; + } else { + x_bottom = xb; + sy1 = y_bottom; + } + + if (x_top >= 0 && x_bottom >= 0 && x_top < len && x_bottom < len) { + // from here on, we don't have to range check x values + + if ((int) x_top == (int) x_bottom) { + float height; + // simple case, only spans one pixel + int x = (int) x_top; + height = sy1 - sy0; + STBTT_assert(x >= 0 && x < len); + scanline[x] += e->direction * (1-((x_top - x) + (x_bottom-x))/2) * height; + scanline_fill[x] += e->direction * height; // everything right of this pixel is filled + } else { + int x,x1,x2; + float y_crossing, step, sign, area; + // covers 2+ pixels + if (x_top > x_bottom) { + // flip scanline vertically; signed area is the same + float t; + sy0 = y_bottom - (sy0 - y_top); + sy1 = y_bottom - (sy1 - y_top); + t = sy0, sy0 = sy1, sy1 = t; + t = x_bottom, x_bottom = x_top, x_top = t; + dx = -dx; + dy = -dy; + t = x0, x0 = xb, xb = t; + } + + x1 = (int) x_top; + x2 = (int) x_bottom; + // compute intersection with y axis at x1+1 + y_crossing = (x1+1 - x0) * dy + y_top; + + sign = e->direction; + // area of the rectangle covered from y0..y_crossing + area = sign * (y_crossing-sy0); + // area of the triangle (x_top,y0), (x+1,y0), (x+1,y_crossing) + scanline[x1] += area * (1-((x_top - x1)+(x1+1-x1))/2); + + step = sign * dy; + for (x = x1+1; x < x2; ++x) { + scanline[x] += area + step/2; + area += step; + } + y_crossing += dy * (x2 - (x1+1)); + + STBTT_assert(STBTT_fabs(area) <= 1.01f); + + scanline[x2] += area + sign * (1-((x2-x2)+(x_bottom-x2))/2) * (sy1-y_crossing); + + scanline_fill[x2] += sign * (sy1-sy0); + } + } else { + // if edge goes outside of box we're drawing, we require + // clipping logic. since this does not match the intended use + // of this library, we use a different, very slow brute + // force implementation + int x; + for (x=0; x < len; ++x) { + // cases: + // + // there can be up to two intersections with the pixel. any intersection + // with left or right edges can be handled by splitting into two (or three) + // regions. intersections with top & bottom do not necessitate case-wise logic. + // + // the old way of doing this found the intersections with the left & right edges, + // then used some simple logic to produce up to three segments in sorted order + // from top-to-bottom. however, this had a problem: if an x edge was epsilon + // across the x border, then the corresponding y position might not be distinct + // from the other y segment, and it might ignored as an empty segment. to avoid + // that, we need to explicitly produce segments based on x positions. + + // rename variables to clearly-defined pairs + float y0 = y_top; + float x1 = (float) (x); + float x2 = (float) (x+1); + float x3 = xb; + float y3 = y_bottom; + + // x = e->x + e->dx * (y-y_top) + // (y-y_top) = (x - e->x) / e->dx + // y = (x - e->x) / e->dx + y_top + float y1 = (x - x0) / dx + y_top; + float y2 = (x+1 - x0) / dx + y_top; + + if (x0 < x1 && x3 > x2) { // three segments descending down-right + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x1,y1); + stbtt__handle_clipped_edge(scanline,x,e, x1,y1, x2,y2); + stbtt__handle_clipped_edge(scanline,x,e, x2,y2, x3,y3); + } else if (x3 < x1 && x0 > x2) { // three segments descending down-left + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x2,y2); + stbtt__handle_clipped_edge(scanline,x,e, x2,y2, x1,y1); + stbtt__handle_clipped_edge(scanline,x,e, x1,y1, x3,y3); + } else if (x0 < x1 && x3 > x1) { // two segments across x, down-right + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x1,y1); + stbtt__handle_clipped_edge(scanline,x,e, x1,y1, x3,y3); + } else if (x3 < x1 && x0 > x1) { // two segments across x, down-left + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x1,y1); + stbtt__handle_clipped_edge(scanline,x,e, x1,y1, x3,y3); + } else if (x0 < x2 && x3 > x2) { // two segments across x+1, down-right + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x2,y2); + stbtt__handle_clipped_edge(scanline,x,e, x2,y2, x3,y3); + } else if (x3 < x2 && x0 > x2) { // two segments across x+1, down-left + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x2,y2); + stbtt__handle_clipped_edge(scanline,x,e, x2,y2, x3,y3); + } else { // one segment + stbtt__handle_clipped_edge(scanline,x,e, x0,y0, x3,y3); + } + } + } + } + e = e->next; + } +} + +// directly AA rasterize edges w/o supersampling +static void stbtt__rasterize_sorted_edges(stbtt__bitmap *result, stbtt__edge *e, int n, int vsubsample, int off_x, int off_y, void *userdata) +{ + stbtt__hheap hh = { 0, 0, 0 }; + stbtt__active_edge *active = NULL; + int y,j=0, i; + float scanline_data[129], *scanline, *scanline2; + + STBTT__NOTUSED(vsubsample); + + if (result->w > 64) + scanline = (float *) STBTT_malloc((result->w*2+1) * sizeof(float), userdata); + else + scanline = scanline_data; + + scanline2 = scanline + result->w; + + y = off_y; + e[n].y0 = (float) (off_y + result->h) + 1; + + while (j < result->h) { + // find center of pixel for this scanline + float scan_y_top = y + 0.0f; + float scan_y_bottom = y + 1.0f; + stbtt__active_edge **step = &active; + + STBTT_memset(scanline , 0, result->w*sizeof(scanline[0])); + STBTT_memset(scanline2, 0, (result->w+1)*sizeof(scanline[0])); + + // update all active edges; + // remove all active edges that terminate before the top of this scanline + while (*step) { + stbtt__active_edge * z = *step; + if (z->ey <= scan_y_top) { + *step = z->next; // delete from list + STBTT_assert(z->direction); + z->direction = 0; + stbtt__hheap_free(&hh, z); + } else { + step = &((*step)->next); // advance through list + } + } + + // insert all edges that start before the bottom of this scanline + while (e->y0 <= scan_y_bottom) { + if (e->y0 != e->y1) { + stbtt__active_edge *z = stbtt__new_active(&hh, e, off_x, scan_y_top, userdata); + if (z != NULL) { + if (j == 0 && off_y != 0) { + if (z->ey < scan_y_top) { + // this can happen due to subpixel positioning and some kind of fp rounding error i think + z->ey = scan_y_top; + } + } + STBTT_assert(z->ey >= scan_y_top); // if we get really unlucky a tiny bit of an edge can be out of bounds + // insert at front + z->next = active; + active = z; + } + } + ++e; + } + + // now process all active edges + if (active) + stbtt__fill_active_edges_new(scanline, scanline2+1, result->w, active, scan_y_top); + + { + float sum = 0; + for (i=0; i < result->w; ++i) { + float k; + int m; + sum += scanline2[i]; + k = scanline[i] + sum; + k = (float) STBTT_fabs(k)*255 + 0.5f; + m = (int) k; + if (m > 255) m = 255; + result->pixels[j*result->stride + i] = (unsigned char) m; + } + } + // advance all the edges + step = &active; + while (*step) { + stbtt__active_edge *z = *step; + z->fx += z->fdx; // advance to position for current scanline + step = &((*step)->next); // advance through list + } + + ++y; + ++j; + } + + stbtt__hheap_cleanup(&hh, userdata); + + if (scanline != scanline_data) + STBTT_free(scanline, userdata); +} +#else +#error "Unrecognized value of STBTT_RASTERIZER_VERSION" +#endif + +#define STBTT__COMPARE(a,b) ((a)->y0 < (b)->y0) + +static void stbtt__sort_edges_ins_sort(stbtt__edge *p, int n) +{ + int i,j; + for (i=1; i < n; ++i) { + stbtt__edge t = p[i], *a = &t; + j = i; + while (j > 0) { + stbtt__edge *b = &p[j-1]; + int c = STBTT__COMPARE(a,b); + if (!c) break; + p[j] = p[j-1]; + --j; + } + if (i != j) + p[j] = t; + } +} + +static void stbtt__sort_edges_quicksort(stbtt__edge *p, int n) +{ + /* threshold for transitioning to insertion sort */ + while (n > 12) { + stbtt__edge t; + int c01,c12,c,m,i,j; + + /* compute median of three */ + m = n >> 1; + c01 = STBTT__COMPARE(&p[0],&p[m]); + c12 = STBTT__COMPARE(&p[m],&p[n-1]); + /* if 0 >= mid >= end, or 0 < mid < end, then use mid */ + if (c01 != c12) { + /* otherwise, we'll need to swap something else to middle */ + int z; + c = STBTT__COMPARE(&p[0],&p[n-1]); + /* 0>mid && midn => n; 0 0 */ + /* 0n: 0>n => 0; 0 n */ + z = (c == c12) ? 0 : n-1; + t = p[z]; + p[z] = p[m]; + p[m] = t; + } + /* now p[m] is the median-of-three */ + /* swap it to the beginning so it won't move around */ + t = p[0]; + p[0] = p[m]; + p[m] = t; + + /* partition loop */ + i=1; + j=n-1; + for(;;) { + /* handling of equality is crucial here */ + /* for sentinels & efficiency with duplicates */ + for (;;++i) { + if (!STBTT__COMPARE(&p[i], &p[0])) break; + } + for (;;--j) { + if (!STBTT__COMPARE(&p[0], &p[j])) break; + } + /* make sure we haven't crossed */ + if (i >= j) break; + t = p[i]; + p[i] = p[j]; + p[j] = t; + + ++i; + --j; + } + /* recurse on smaller side, iterate on larger */ + if (j < (n-i)) { + stbtt__sort_edges_quicksort(p,j); + p = p+i; + n = n-i; + } else { + stbtt__sort_edges_quicksort(p+i, n-i); + n = j; + } + } +} + +static void stbtt__sort_edges(stbtt__edge *p, int n) +{ + stbtt__sort_edges_quicksort(p, n); + stbtt__sort_edges_ins_sort(p, n); +} + +typedef struct +{ + float x,y; +} stbtt__point; + +static void stbtt__rasterize(stbtt__bitmap *result, stbtt__point *pts, int *wcount, int windings, float scale_x, float scale_y, float shift_x, float shift_y, int off_x, int off_y, int invert, void *userdata) +{ + float y_scale_inv = invert ? -scale_y : scale_y; + stbtt__edge *e; + int n,i,j,k,m; +#if STBTT_RASTERIZER_VERSION == 1 + int vsubsample = result->h < 8 ? 15 : 5; +#elif STBTT_RASTERIZER_VERSION == 2 + int vsubsample = 1; +#else + #error "Unrecognized value of STBTT_RASTERIZER_VERSION" +#endif + // vsubsample should divide 255 evenly; otherwise we won't reach full opacity + + // now we have to blow out the windings into explicit edge lists + n = 0; + for (i=0; i < windings; ++i) + n += wcount[i]; + + e = (stbtt__edge *) STBTT_malloc(sizeof(*e) * (n+1), userdata); // add an extra one as a sentinel + if (e == 0) return; + n = 0; + + m=0; + for (i=0; i < windings; ++i) { + stbtt__point *p = pts + m; + m += wcount[i]; + j = wcount[i]-1; + for (k=0; k < wcount[i]; j=k++) { + int a=k,b=j; + // skip the edge if horizontal + if (p[j].y == p[k].y) + continue; + // add edge from j to k to the list + e[n].invert = 0; + if (invert ? p[j].y > p[k].y : p[j].y < p[k].y) { + e[n].invert = 1; + a=j,b=k; + } + e[n].x0 = p[a].x * scale_x + shift_x; + e[n].y0 = (p[a].y * y_scale_inv + shift_y) * vsubsample; + e[n].x1 = p[b].x * scale_x + shift_x; + e[n].y1 = (p[b].y * y_scale_inv + shift_y) * vsubsample; + ++n; + } + } + + // now sort the edges by their highest point (should snap to integer, and then by x) + //STBTT_sort(e, n, sizeof(e[0]), stbtt__edge_compare); + stbtt__sort_edges(e, n); + + // now, traverse the scanlines and find the intersections on each scanline, use xor winding rule + stbtt__rasterize_sorted_edges(result, e, n, vsubsample, off_x, off_y, userdata); + + STBTT_free(e, userdata); +} + +static void stbtt__add_point(stbtt__point *points, int n, float x, float y) +{ + if (!points) return; // during first pass, it's unallocated + points[n].x = x; + points[n].y = y; +} + +// tessellate until threshold p is happy... @TODO warped to compensate for non-linear stretching +static int stbtt__tesselate_curve(stbtt__point *points, int *num_points, float x0, float y0, float x1, float y1, float x2, float y2, float objspace_flatness_squared, int n) +{ + // midpoint + float mx = (x0 + 2*x1 + x2)/4; + float my = (y0 + 2*y1 + y2)/4; + // versus directly drawn line + float dx = (x0+x2)/2 - mx; + float dy = (y0+y2)/2 - my; + if (n > 16) // 65536 segments on one curve better be enough! + return 1; + if (dx*dx+dy*dy > objspace_flatness_squared) { // half-pixel error allowed... need to be smaller if AA + stbtt__tesselate_curve(points, num_points, x0,y0, (x0+x1)/2.0f,(y0+y1)/2.0f, mx,my, objspace_flatness_squared,n+1); + stbtt__tesselate_curve(points, num_points, mx,my, (x1+x2)/2.0f,(y1+y2)/2.0f, x2,y2, objspace_flatness_squared,n+1); + } else { + stbtt__add_point(points, *num_points,x2,y2); + *num_points = *num_points+1; + } + return 1; +} + +static void stbtt__tesselate_cubic(stbtt__point *points, int *num_points, float x0, float y0, float x1, float y1, float x2, float y2, float x3, float y3, float objspace_flatness_squared, int n) +{ + // @TODO this "flatness" calculation is just made-up nonsense that seems to work well enough + float dx0 = x1-x0; + float dy0 = y1-y0; + float dx1 = x2-x1; + float dy1 = y2-y1; + float dx2 = x3-x2; + float dy2 = y3-y2; + float dx = x3-x0; + float dy = y3-y0; + float longlen = (float) (STBTT_sqrt(dx0*dx0+dy0*dy0)+STBTT_sqrt(dx1*dx1+dy1*dy1)+STBTT_sqrt(dx2*dx2+dy2*dy2)); + float shortlen = (float) STBTT_sqrt(dx*dx+dy*dy); + float flatness_squared = longlen*longlen-shortlen*shortlen; + + if (n > 16) // 65536 segments on one curve better be enough! + return; + + if (flatness_squared > objspace_flatness_squared) { + float x01 = (x0+x1)/2; + float y01 = (y0+y1)/2; + float x12 = (x1+x2)/2; + float y12 = (y1+y2)/2; + float x23 = (x2+x3)/2; + float y23 = (y2+y3)/2; + + float xa = (x01+x12)/2; + float ya = (y01+y12)/2; + float xb = (x12+x23)/2; + float yb = (y12+y23)/2; + + float mx = (xa+xb)/2; + float my = (ya+yb)/2; + + stbtt__tesselate_cubic(points, num_points, x0,y0, x01,y01, xa,ya, mx,my, objspace_flatness_squared,n+1); + stbtt__tesselate_cubic(points, num_points, mx,my, xb,yb, x23,y23, x3,y3, objspace_flatness_squared,n+1); + } else { + stbtt__add_point(points, *num_points,x3,y3); + *num_points = *num_points+1; + } +} + +// returns number of contours +static stbtt__point *stbtt_FlattenCurves(stbtt_vertex *vertices, int num_verts, float objspace_flatness, int **contour_lengths, int *num_contours, void *userdata) +{ + stbtt__point *points=0; + int num_points=0; + + float objspace_flatness_squared = objspace_flatness * objspace_flatness; + int i,n=0,start=0, pass; + + // count how many "moves" there are to get the contour count + for (i=0; i < num_verts; ++i) + if (vertices[i].type == STBTT_vmove) + ++n; + + *num_contours = n; + if (n == 0) return 0; + + *contour_lengths = (int *) STBTT_malloc(sizeof(**contour_lengths) * n, userdata); + + if (*contour_lengths == 0) { + *num_contours = 0; + return 0; + } + + // make two passes through the points so we don't need to realloc + for (pass=0; pass < 2; ++pass) { + float x=0,y=0; + if (pass == 1) { + points = (stbtt__point *) STBTT_malloc(num_points * sizeof(points[0]), userdata); + if (points == NULL) goto error; + } + num_points = 0; + n= -1; + for (i=0; i < num_verts; ++i) { + switch (vertices[i].type) { + case STBTT_vmove: + // start the next contour + if (n >= 0) + (*contour_lengths)[n] = num_points - start; + ++n; + start = num_points; + + x = vertices[i].x, y = vertices[i].y; + stbtt__add_point(points, num_points++, x,y); + break; + case STBTT_vline: + x = vertices[i].x, y = vertices[i].y; + stbtt__add_point(points, num_points++, x, y); + break; + case STBTT_vcurve: + stbtt__tesselate_curve(points, &num_points, x,y, + vertices[i].cx, vertices[i].cy, + vertices[i].x, vertices[i].y, + objspace_flatness_squared, 0); + x = vertices[i].x, y = vertices[i].y; + break; + case STBTT_vcubic: + stbtt__tesselate_cubic(points, &num_points, x,y, + vertices[i].cx, vertices[i].cy, + vertices[i].cx1, vertices[i].cy1, + vertices[i].x, vertices[i].y, + objspace_flatness_squared, 0); + x = vertices[i].x, y = vertices[i].y; + break; + } + } + (*contour_lengths)[n] = num_points - start; + } + + return points; +error: + STBTT_free(points, userdata); + STBTT_free(*contour_lengths, userdata); + *contour_lengths = 0; + *num_contours = 0; + return NULL; +} + +STBTT_DEF void stbtt_Rasterize(stbtt__bitmap *result, float flatness_in_pixels, stbtt_vertex *vertices, int num_verts, float scale_x, float scale_y, float shift_x, float shift_y, int x_off, int y_off, int invert, void *userdata) +{ + float scale = scale_x > scale_y ? scale_y : scale_x; + int winding_count = 0; + int *winding_lengths = NULL; + stbtt__point *windings = stbtt_FlattenCurves(vertices, num_verts, flatness_in_pixels / scale, &winding_lengths, &winding_count, userdata); + if (windings) { + stbtt__rasterize(result, windings, winding_lengths, winding_count, scale_x, scale_y, shift_x, shift_y, x_off, y_off, invert, userdata); + STBTT_free(winding_lengths, userdata); + STBTT_free(windings, userdata); + } +} + +STBTT_DEF void stbtt_FreeBitmap(unsigned char *bitmap, void *userdata) +{ + STBTT_free(bitmap, userdata); +} + +STBTT_DEF unsigned char *stbtt_GetGlyphBitmapSubpixel(const stbtt_fontinfo *info, float scale_x, float scale_y, float shift_x, float shift_y, int glyph, int *width, int *height, int *xoff, int *yoff) +{ + int ix0,iy0,ix1,iy1; + stbtt__bitmap gbm; + stbtt_vertex *vertices; + int num_verts = stbtt_GetGlyphShape(info, glyph, &vertices); + + if (scale_x == 0) scale_x = scale_y; + if (scale_y == 0) { + if (scale_x == 0) { + STBTT_free(vertices, info->userdata); + return NULL; + } + scale_y = scale_x; + } + + stbtt_GetGlyphBitmapBoxSubpixel(info, glyph, scale_x, scale_y, shift_x, shift_y, &ix0,&iy0,&ix1,&iy1); + + // now we get the size + gbm.w = (ix1 - ix0); + gbm.h = (iy1 - iy0); + gbm.pixels = NULL; // in case we error + + if (width ) *width = gbm.w; + if (height) *height = gbm.h; + if (xoff ) *xoff = ix0; + if (yoff ) *yoff = iy0; + + if (gbm.w && gbm.h) { + gbm.pixels = (unsigned char *) STBTT_malloc(gbm.w * gbm.h, info->userdata); + if (gbm.pixels) { + gbm.stride = gbm.w; + + stbtt_Rasterize(&gbm, 0.35f, vertices, num_verts, scale_x, scale_y, shift_x, shift_y, ix0, iy0, 1, info->userdata); + } + } + STBTT_free(vertices, info->userdata); + return gbm.pixels; +} + +STBTT_DEF unsigned char *stbtt_GetGlyphBitmap(const stbtt_fontinfo *info, float scale_x, float scale_y, int glyph, int *width, int *height, int *xoff, int *yoff) +{ + return stbtt_GetGlyphBitmapSubpixel(info, scale_x, scale_y, 0.0f, 0.0f, glyph, width, height, xoff, yoff); +} + +STBTT_DEF void stbtt_MakeGlyphBitmapSubpixel(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int glyph) +{ + int ix0,iy0; + stbtt_vertex *vertices; + int num_verts = stbtt_GetGlyphShape(info, glyph, &vertices); + stbtt__bitmap gbm; + + stbtt_GetGlyphBitmapBoxSubpixel(info, glyph, scale_x, scale_y, shift_x, shift_y, &ix0,&iy0,0,0); + gbm.pixels = output; + gbm.w = out_w; + gbm.h = out_h; + gbm.stride = out_stride; + + if (gbm.w && gbm.h) + stbtt_Rasterize(&gbm, 0.35f, vertices, num_verts, scale_x, scale_y, shift_x, shift_y, ix0,iy0, 1, info->userdata); + + STBTT_free(vertices, info->userdata); +} + +STBTT_DEF void stbtt_MakeGlyphBitmap(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, int glyph) +{ + stbtt_MakeGlyphBitmapSubpixel(info, output, out_w, out_h, out_stride, scale_x, scale_y, 0.0f,0.0f, glyph); +} + +STBTT_DEF unsigned char *stbtt_GetCodepointBitmapSubpixel(const stbtt_fontinfo *info, float scale_x, float scale_y, float shift_x, float shift_y, int codepoint, int *width, int *height, int *xoff, int *yoff) +{ + return stbtt_GetGlyphBitmapSubpixel(info, scale_x, scale_y,shift_x,shift_y, stbtt_FindGlyphIndex(info,codepoint), width,height,xoff,yoff); +} + +STBTT_DEF void stbtt_MakeCodepointBitmapSubpixelPrefilter(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int oversample_x, int oversample_y, float *sub_x, float *sub_y, int codepoint) +{ + stbtt_MakeGlyphBitmapSubpixelPrefilter(info, output, out_w, out_h, out_stride, scale_x, scale_y, shift_x, shift_y, oversample_x, oversample_y, sub_x, sub_y, stbtt_FindGlyphIndex(info,codepoint)); +} + +STBTT_DEF void stbtt_MakeCodepointBitmapSubpixel(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int codepoint) +{ + stbtt_MakeGlyphBitmapSubpixel(info, output, out_w, out_h, out_stride, scale_x, scale_y, shift_x, shift_y, stbtt_FindGlyphIndex(info,codepoint)); +} + +STBTT_DEF unsigned char *stbtt_GetCodepointBitmap(const stbtt_fontinfo *info, float scale_x, float scale_y, int codepoint, int *width, int *height, int *xoff, int *yoff) +{ + return stbtt_GetCodepointBitmapSubpixel(info, scale_x, scale_y, 0.0f,0.0f, codepoint, width,height,xoff,yoff); +} + +STBTT_DEF void stbtt_MakeCodepointBitmap(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, int codepoint) +{ + stbtt_MakeCodepointBitmapSubpixel(info, output, out_w, out_h, out_stride, scale_x, scale_y, 0.0f,0.0f, codepoint); +} + +////////////////////////////////////////////////////////////////////////////// +// +// bitmap baking +// +// This is SUPER-CRAPPY packing to keep source code small + +static int stbtt_BakeFontBitmap_internal(unsigned char *data, int offset, // font location (use offset=0 for plain .ttf) + float pixel_height, // height of font in pixels + unsigned char *pixels, int pw, int ph, // bitmap to be filled in + int first_char, int num_chars, // characters to bake + stbtt_bakedchar *chardata) +{ + float scale; + int x,y,bottom_y, i; + stbtt_fontinfo f; + f.userdata = NULL; + if (!stbtt_InitFont(&f, data, offset)) + return -1; + STBTT_memset(pixels, 0, pw*ph); // background of 0 around pixels + x=y=1; + bottom_y = 1; + + scale = stbtt_ScaleForPixelHeight(&f, pixel_height); + + for (i=0; i < num_chars; ++i) { + int advance, lsb, x0,y0,x1,y1,gw,gh; + int g = stbtt_FindGlyphIndex(&f, first_char + i); + stbtt_GetGlyphHMetrics(&f, g, &advance, &lsb); + stbtt_GetGlyphBitmapBox(&f, g, scale,scale, &x0,&y0,&x1,&y1); + gw = x1-x0; + gh = y1-y0; + if (x + gw + 1 >= pw) + y = bottom_y, x = 1; // advance to next row + if (y + gh + 1 >= ph) // check if it fits vertically AFTER potentially moving to next row + return -i; + STBTT_assert(x+gw < pw); + STBTT_assert(y+gh < ph); + stbtt_MakeGlyphBitmap(&f, pixels+x+y*pw, gw,gh,pw, scale,scale, g); + chardata[i].x0 = (stbtt_int16) x; + chardata[i].y0 = (stbtt_int16) y; + chardata[i].x1 = (stbtt_int16) (x + gw); + chardata[i].y1 = (stbtt_int16) (y + gh); + chardata[i].xadvance = scale * advance; + chardata[i].xoff = (float) x0; + chardata[i].yoff = (float) y0; + x = x + gw + 1; + if (y+gh+1 > bottom_y) + bottom_y = y+gh+1; + } + return bottom_y; +} + +STBTT_DEF void stbtt_GetBakedQuad(const stbtt_bakedchar *chardata, int pw, int ph, int char_index, float *xpos, float *ypos, stbtt_aligned_quad *q, int opengl_fillrule) +{ + float d3d_bias = opengl_fillrule ? 0 : -0.5f; + float ipw = 1.0f / pw, iph = 1.0f / ph; + const stbtt_bakedchar *b = chardata + char_index; + int round_x = STBTT_ifloor((*xpos + b->xoff) + 0.5f); + int round_y = STBTT_ifloor((*ypos + b->yoff) + 0.5f); + + q->x0 = round_x + d3d_bias; + q->y0 = round_y + d3d_bias; + q->x1 = round_x + b->x1 - b->x0 + d3d_bias; + q->y1 = round_y + b->y1 - b->y0 + d3d_bias; + + q->s0 = b->x0 * ipw; + q->t0 = b->y0 * iph; + q->s1 = b->x1 * ipw; + q->t1 = b->y1 * iph; + + *xpos += b->xadvance; +} + +////////////////////////////////////////////////////////////////////////////// +// +// rectangle packing replacement routines if you don't have stb_rect_pack.h +// + +#ifndef STB_RECT_PACK_VERSION + +typedef int stbrp_coord; + +//////////////////////////////////////////////////////////////////////////////////// +// // +// // +// COMPILER WARNING ?!?!? // +// // +// // +// if you get a compile warning due to these symbols being defined more than // +// once, move #include "stb_rect_pack.h" before #include "stb_truetype.h" // +// // +//////////////////////////////////////////////////////////////////////////////////// + +typedef struct +{ + int width,height; + int x,y,bottom_y; +} stbrp_context; + +typedef struct +{ + unsigned char x; +} stbrp_node; + +struct stbrp_rect +{ + stbrp_coord x,y; + int id,w,h,was_packed; +}; + +static void stbrp_init_target(stbrp_context *con, int pw, int ph, stbrp_node *nodes, int num_nodes) +{ + con->width = pw; + con->height = ph; + con->x = 0; + con->y = 0; + con->bottom_y = 0; + STBTT__NOTUSED(nodes); + STBTT__NOTUSED(num_nodes); +} + +static void stbrp_pack_rects(stbrp_context *con, stbrp_rect *rects, int num_rects) +{ + int i; + for (i=0; i < num_rects; ++i) { + if (con->x + rects[i].w > con->width) { + con->x = 0; + con->y = con->bottom_y; + } + if (con->y + rects[i].h > con->height) + break; + rects[i].x = con->x; + rects[i].y = con->y; + rects[i].was_packed = 1; + con->x += rects[i].w; + if (con->y + rects[i].h > con->bottom_y) + con->bottom_y = con->y + rects[i].h; + } + for ( ; i < num_rects; ++i) + rects[i].was_packed = 0; +} +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// bitmap baking +// +// This is SUPER-AWESOME (tm Ryan Gordon) packing using stb_rect_pack.h. If +// stb_rect_pack.h isn't available, it uses the BakeFontBitmap strategy. + +STBTT_DEF int stbtt_PackBegin(stbtt_pack_context *spc, unsigned char *pixels, int pw, int ph, int stride_in_bytes, int padding, void *alloc_context) +{ + stbrp_context *context = (stbrp_context *) STBTT_malloc(sizeof(*context) ,alloc_context); + int num_nodes = pw - padding; + stbrp_node *nodes = (stbrp_node *) STBTT_malloc(sizeof(*nodes ) * num_nodes,alloc_context); + + if (context == NULL || nodes == NULL) { + if (context != NULL) STBTT_free(context, alloc_context); + if (nodes != NULL) STBTT_free(nodes , alloc_context); + return 0; + } + + spc->user_allocator_context = alloc_context; + spc->width = pw; + spc->height = ph; + spc->pixels = pixels; + spc->pack_info = context; + spc->nodes = nodes; + spc->padding = padding; + spc->stride_in_bytes = stride_in_bytes != 0 ? stride_in_bytes : pw; + spc->h_oversample = 1; + spc->v_oversample = 1; + spc->skip_missing = 0; + + stbrp_init_target(context, pw-padding, ph-padding, nodes, num_nodes); + + if (pixels) + STBTT_memset(pixels, 0, pw*ph); // background of 0 around pixels + + return 1; +} + +STBTT_DEF void stbtt_PackEnd (stbtt_pack_context *spc) +{ + STBTT_free(spc->nodes , spc->user_allocator_context); + STBTT_free(spc->pack_info, spc->user_allocator_context); +} + +STBTT_DEF void stbtt_PackSetOversampling(stbtt_pack_context *spc, unsigned int h_oversample, unsigned int v_oversample) +{ + STBTT_assert(h_oversample <= STBTT_MAX_OVERSAMPLE); + STBTT_assert(v_oversample <= STBTT_MAX_OVERSAMPLE); + if (h_oversample <= STBTT_MAX_OVERSAMPLE) + spc->h_oversample = h_oversample; + if (v_oversample <= STBTT_MAX_OVERSAMPLE) + spc->v_oversample = v_oversample; +} + +STBTT_DEF void stbtt_PackSetSkipMissingCodepoints(stbtt_pack_context *spc, int skip) +{ + spc->skip_missing = skip; +} + +#define STBTT__OVER_MASK (STBTT_MAX_OVERSAMPLE-1) + +static void stbtt__h_prefilter(unsigned char *pixels, int w, int h, int stride_in_bytes, unsigned int kernel_width) +{ + unsigned char buffer[STBTT_MAX_OVERSAMPLE]; + int safe_w = w - kernel_width; + int j; + STBTT_memset(buffer, 0, STBTT_MAX_OVERSAMPLE); // suppress bogus warning from VS2013 -analyze + for (j=0; j < h; ++j) { + int i; + unsigned int total; + STBTT_memset(buffer, 0, kernel_width); + + total = 0; + + // make kernel_width a constant in common cases so compiler can optimize out the divide + switch (kernel_width) { + case 2: + for (i=0; i <= safe_w; ++i) { + total += pixels[i] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i]; + pixels[i] = (unsigned char) (total / 2); + } + break; + case 3: + for (i=0; i <= safe_w; ++i) { + total += pixels[i] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i]; + pixels[i] = (unsigned char) (total / 3); + } + break; + case 4: + for (i=0; i <= safe_w; ++i) { + total += pixels[i] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i]; + pixels[i] = (unsigned char) (total / 4); + } + break; + case 5: + for (i=0; i <= safe_w; ++i) { + total += pixels[i] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i]; + pixels[i] = (unsigned char) (total / 5); + } + break; + default: + for (i=0; i <= safe_w; ++i) { + total += pixels[i] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i]; + pixels[i] = (unsigned char) (total / kernel_width); + } + break; + } + + for (; i < w; ++i) { + STBTT_assert(pixels[i] == 0); + total -= buffer[i & STBTT__OVER_MASK]; + pixels[i] = (unsigned char) (total / kernel_width); + } + + pixels += stride_in_bytes; + } +} + +static void stbtt__v_prefilter(unsigned char *pixels, int w, int h, int stride_in_bytes, unsigned int kernel_width) +{ + unsigned char buffer[STBTT_MAX_OVERSAMPLE]; + int safe_h = h - kernel_width; + int j; + STBTT_memset(buffer, 0, STBTT_MAX_OVERSAMPLE); // suppress bogus warning from VS2013 -analyze + for (j=0; j < w; ++j) { + int i; + unsigned int total; + STBTT_memset(buffer, 0, kernel_width); + + total = 0; + + // make kernel_width a constant in common cases so compiler can optimize out the divide + switch (kernel_width) { + case 2: + for (i=0; i <= safe_h; ++i) { + total += pixels[i*stride_in_bytes] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i*stride_in_bytes]; + pixels[i*stride_in_bytes] = (unsigned char) (total / 2); + } + break; + case 3: + for (i=0; i <= safe_h; ++i) { + total += pixels[i*stride_in_bytes] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i*stride_in_bytes]; + pixels[i*stride_in_bytes] = (unsigned char) (total / 3); + } + break; + case 4: + for (i=0; i <= safe_h; ++i) { + total += pixels[i*stride_in_bytes] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i*stride_in_bytes]; + pixels[i*stride_in_bytes] = (unsigned char) (total / 4); + } + break; + case 5: + for (i=0; i <= safe_h; ++i) { + total += pixels[i*stride_in_bytes] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i*stride_in_bytes]; + pixels[i*stride_in_bytes] = (unsigned char) (total / 5); + } + break; + default: + for (i=0; i <= safe_h; ++i) { + total += pixels[i*stride_in_bytes] - buffer[i & STBTT__OVER_MASK]; + buffer[(i+kernel_width) & STBTT__OVER_MASK] = pixels[i*stride_in_bytes]; + pixels[i*stride_in_bytes] = (unsigned char) (total / kernel_width); + } + break; + } + + for (; i < h; ++i) { + STBTT_assert(pixels[i*stride_in_bytes] == 0); + total -= buffer[i & STBTT__OVER_MASK]; + pixels[i*stride_in_bytes] = (unsigned char) (total / kernel_width); + } + + pixels += 1; + } +} + +static float stbtt__oversample_shift(int oversample) +{ + if (!oversample) + return 0.0f; + + // The prefilter is a box filter of width "oversample", + // which shifts phase by (oversample - 1)/2 pixels in + // oversampled space. We want to shift in the opposite + // direction to counter this. + return (float)-(oversample - 1) / (2.0f * (float)oversample); +} + +// rects array must be big enough to accommodate all characters in the given ranges +STBTT_DEF int stbtt_PackFontRangesGatherRects(stbtt_pack_context *spc, const stbtt_fontinfo *info, stbtt_pack_range *ranges, int num_ranges, stbrp_rect *rects) +{ + int i,j,k; + int missing_glyph_added = 0; + + k=0; + for (i=0; i < num_ranges; ++i) { + float fh = ranges[i].font_size; + float scale = fh > 0 ? stbtt_ScaleForPixelHeight(info, fh) : stbtt_ScaleForMappingEmToPixels(info, -fh); + ranges[i].h_oversample = (unsigned char) spc->h_oversample; + ranges[i].v_oversample = (unsigned char) spc->v_oversample; + for (j=0; j < ranges[i].num_chars; ++j) { + int x0,y0,x1,y1; + int codepoint = ranges[i].array_of_unicode_codepoints == NULL ? ranges[i].first_unicode_codepoint_in_range + j : ranges[i].array_of_unicode_codepoints[j]; + int glyph = stbtt_FindGlyphIndex(info, codepoint); + if (glyph == 0 && (spc->skip_missing || missing_glyph_added)) { + rects[k].w = rects[k].h = 0; + } else { + stbtt_GetGlyphBitmapBoxSubpixel(info,glyph, + scale * spc->h_oversample, + scale * spc->v_oversample, + 0,0, + &x0,&y0,&x1,&y1); + rects[k].w = (stbrp_coord) (x1-x0 + spc->padding + spc->h_oversample-1); + rects[k].h = (stbrp_coord) (y1-y0 + spc->padding + spc->v_oversample-1); + if (glyph == 0) + missing_glyph_added = 1; + } + ++k; + } + } + + return k; +} + +STBTT_DEF void stbtt_MakeGlyphBitmapSubpixelPrefilter(const stbtt_fontinfo *info, unsigned char *output, int out_w, int out_h, int out_stride, float scale_x, float scale_y, float shift_x, float shift_y, int prefilter_x, int prefilter_y, float *sub_x, float *sub_y, int glyph) +{ + stbtt_MakeGlyphBitmapSubpixel(info, + output, + out_w - (prefilter_x - 1), + out_h - (prefilter_y - 1), + out_stride, + scale_x, + scale_y, + shift_x, + shift_y, + glyph); + + if (prefilter_x > 1) + stbtt__h_prefilter(output, out_w, out_h, out_stride, prefilter_x); + + if (prefilter_y > 1) + stbtt__v_prefilter(output, out_w, out_h, out_stride, prefilter_y); + + *sub_x = stbtt__oversample_shift(prefilter_x); + *sub_y = stbtt__oversample_shift(prefilter_y); +} + +// rects array must be big enough to accommodate all characters in the given ranges +STBTT_DEF int stbtt_PackFontRangesRenderIntoRects(stbtt_pack_context *spc, const stbtt_fontinfo *info, stbtt_pack_range *ranges, int num_ranges, stbrp_rect *rects) +{ + int i,j,k, missing_glyph = -1, return_value = 1; + + // save current values + int old_h_over = spc->h_oversample; + int old_v_over = spc->v_oversample; + + k = 0; + for (i=0; i < num_ranges; ++i) { + float fh = ranges[i].font_size; + float scale = fh > 0 ? stbtt_ScaleForPixelHeight(info, fh) : stbtt_ScaleForMappingEmToPixels(info, -fh); + float recip_h,recip_v,sub_x,sub_y; + spc->h_oversample = ranges[i].h_oversample; + spc->v_oversample = ranges[i].v_oversample; + recip_h = 1.0f / spc->h_oversample; + recip_v = 1.0f / spc->v_oversample; + sub_x = stbtt__oversample_shift(spc->h_oversample); + sub_y = stbtt__oversample_shift(spc->v_oversample); + for (j=0; j < ranges[i].num_chars; ++j) { + stbrp_rect *r = &rects[k]; + if (r->was_packed && r->w != 0 && r->h != 0) { + stbtt_packedchar *bc = &ranges[i].chardata_for_range[j]; + int advance, lsb, x0,y0,x1,y1; + int codepoint = ranges[i].array_of_unicode_codepoints == NULL ? ranges[i].first_unicode_codepoint_in_range + j : ranges[i].array_of_unicode_codepoints[j]; + int glyph = stbtt_FindGlyphIndex(info, codepoint); + stbrp_coord pad = (stbrp_coord) spc->padding; + + // pad on left and top + r->x += pad; + r->y += pad; + r->w -= pad; + r->h -= pad; + stbtt_GetGlyphHMetrics(info, glyph, &advance, &lsb); + stbtt_GetGlyphBitmapBox(info, glyph, + scale * spc->h_oversample, + scale * spc->v_oversample, + &x0,&y0,&x1,&y1); + stbtt_MakeGlyphBitmapSubpixel(info, + spc->pixels + r->x + r->y*spc->stride_in_bytes, + r->w - spc->h_oversample+1, + r->h - spc->v_oversample+1, + spc->stride_in_bytes, + scale * spc->h_oversample, + scale * spc->v_oversample, + 0,0, + glyph); + + if (spc->h_oversample > 1) + stbtt__h_prefilter(spc->pixels + r->x + r->y*spc->stride_in_bytes, + r->w, r->h, spc->stride_in_bytes, + spc->h_oversample); + + if (spc->v_oversample > 1) + stbtt__v_prefilter(spc->pixels + r->x + r->y*spc->stride_in_bytes, + r->w, r->h, spc->stride_in_bytes, + spc->v_oversample); + + bc->x0 = (stbtt_int16) r->x; + bc->y0 = (stbtt_int16) r->y; + bc->x1 = (stbtt_int16) (r->x + r->w); + bc->y1 = (stbtt_int16) (r->y + r->h); + bc->xadvance = scale * advance; + bc->xoff = (float) x0 * recip_h + sub_x; + bc->yoff = (float) y0 * recip_v + sub_y; + bc->xoff2 = (x0 + r->w) * recip_h + sub_x; + bc->yoff2 = (y0 + r->h) * recip_v + sub_y; + + if (glyph == 0) + missing_glyph = j; + } else if (spc->skip_missing) { + return_value = 0; + } else if (r->was_packed && r->w == 0 && r->h == 0 && missing_glyph >= 0) { + ranges[i].chardata_for_range[j] = ranges[i].chardata_for_range[missing_glyph]; + } else { + return_value = 0; // if any fail, report failure + } + + ++k; + } + } + + // restore original values + spc->h_oversample = old_h_over; + spc->v_oversample = old_v_over; + + return return_value; +} + +STBTT_DEF void stbtt_PackFontRangesPackRects(stbtt_pack_context *spc, stbrp_rect *rects, int num_rects) +{ + stbrp_pack_rects((stbrp_context *) spc->pack_info, rects, num_rects); +} + +STBTT_DEF int stbtt_PackFontRanges(stbtt_pack_context *spc, const unsigned char *fontdata, int font_index, stbtt_pack_range *ranges, int num_ranges) +{ + stbtt_fontinfo info; + int i,j,n, return_value = 1; + //stbrp_context *context = (stbrp_context *) spc->pack_info; + stbrp_rect *rects; + + // flag all characters as NOT packed + for (i=0; i < num_ranges; ++i) + for (j=0; j < ranges[i].num_chars; ++j) + ranges[i].chardata_for_range[j].x0 = + ranges[i].chardata_for_range[j].y0 = + ranges[i].chardata_for_range[j].x1 = + ranges[i].chardata_for_range[j].y1 = 0; + + n = 0; + for (i=0; i < num_ranges; ++i) + n += ranges[i].num_chars; + + rects = (stbrp_rect *) STBTT_malloc(sizeof(*rects) * n, spc->user_allocator_context); + if (rects == NULL) + return 0; + + info.userdata = spc->user_allocator_context; + stbtt_InitFont(&info, fontdata, stbtt_GetFontOffsetForIndex(fontdata,font_index)); + + n = stbtt_PackFontRangesGatherRects(spc, &info, ranges, num_ranges, rects); + + stbtt_PackFontRangesPackRects(spc, rects, n); + + return_value = stbtt_PackFontRangesRenderIntoRects(spc, &info, ranges, num_ranges, rects); + + STBTT_free(rects, spc->user_allocator_context); + return return_value; +} + +STBTT_DEF int stbtt_PackFontRange(stbtt_pack_context *spc, const unsigned char *fontdata, int font_index, float font_size, + int first_unicode_codepoint_in_range, int num_chars_in_range, stbtt_packedchar *chardata_for_range) +{ + stbtt_pack_range range; + range.first_unicode_codepoint_in_range = first_unicode_codepoint_in_range; + range.array_of_unicode_codepoints = NULL; + range.num_chars = num_chars_in_range; + range.chardata_for_range = chardata_for_range; + range.font_size = font_size; + return stbtt_PackFontRanges(spc, fontdata, font_index, &range, 1); +} + +STBTT_DEF void stbtt_GetScaledFontVMetrics(const unsigned char *fontdata, int index, float size, float *ascent, float *descent, float *lineGap) +{ + int i_ascent, i_descent, i_lineGap; + float scale; + stbtt_fontinfo info; + stbtt_InitFont(&info, fontdata, stbtt_GetFontOffsetForIndex(fontdata, index)); + scale = size > 0 ? stbtt_ScaleForPixelHeight(&info, size) : stbtt_ScaleForMappingEmToPixels(&info, -size); + stbtt_GetFontVMetrics(&info, &i_ascent, &i_descent, &i_lineGap); + *ascent = (float) i_ascent * scale; + *descent = (float) i_descent * scale; + *lineGap = (float) i_lineGap * scale; +} + +STBTT_DEF void stbtt_GetPackedQuad(const stbtt_packedchar *chardata, int pw, int ph, int char_index, float *xpos, float *ypos, stbtt_aligned_quad *q, int align_to_integer) +{ + float ipw = 1.0f / pw, iph = 1.0f / ph; + const stbtt_packedchar *b = chardata + char_index; + + if (align_to_integer) { + float x = (float) STBTT_ifloor((*xpos + b->xoff) + 0.5f); + float y = (float) STBTT_ifloor((*ypos + b->yoff) + 0.5f); + q->x0 = x; + q->y0 = y; + q->x1 = x + b->xoff2 - b->xoff; + q->y1 = y + b->yoff2 - b->yoff; + } else { + q->x0 = *xpos + b->xoff; + q->y0 = *ypos + b->yoff; + q->x1 = *xpos + b->xoff2; + q->y1 = *ypos + b->yoff2; + } + + q->s0 = b->x0 * ipw; + q->t0 = b->y0 * iph; + q->s1 = b->x1 * ipw; + q->t1 = b->y1 * iph; + + *xpos += b->xadvance; +} + +////////////////////////////////////////////////////////////////////////////// +// +// sdf computation +// + +#define STBTT_min(a,b) ((a) < (b) ? (a) : (b)) +#define STBTT_max(a,b) ((a) < (b) ? (b) : (a)) + +static int stbtt__ray_intersect_bezier(float orig[2], float ray[2], float q0[2], float q1[2], float q2[2], float hits[2][2]) +{ + float q0perp = q0[1]*ray[0] - q0[0]*ray[1]; + float q1perp = q1[1]*ray[0] - q1[0]*ray[1]; + float q2perp = q2[1]*ray[0] - q2[0]*ray[1]; + float roperp = orig[1]*ray[0] - orig[0]*ray[1]; + + float a = q0perp - 2*q1perp + q2perp; + float b = q1perp - q0perp; + float c = q0perp - roperp; + + float s0 = 0., s1 = 0.; + int num_s = 0; + + if (a != 0.0) { + float discr = b*b - a*c; + if (discr > 0.0) { + float rcpna = -1 / a; + float d = (float) STBTT_sqrt(discr); + s0 = (b+d) * rcpna; + s1 = (b-d) * rcpna; + if (s0 >= 0.0 && s0 <= 1.0) + num_s = 1; + if (d > 0.0 && s1 >= 0.0 && s1 <= 1.0) { + if (num_s == 0) s0 = s1; + ++num_s; + } + } + } else { + // 2*b*s + c = 0 + // s = -c / (2*b) + s0 = c / (-2 * b); + if (s0 >= 0.0 && s0 <= 1.0) + num_s = 1; + } + + if (num_s == 0) + return 0; + else { + float rcp_len2 = 1 / (ray[0]*ray[0] + ray[1]*ray[1]); + float rayn_x = ray[0] * rcp_len2, rayn_y = ray[1] * rcp_len2; + + float q0d = q0[0]*rayn_x + q0[1]*rayn_y; + float q1d = q1[0]*rayn_x + q1[1]*rayn_y; + float q2d = q2[0]*rayn_x + q2[1]*rayn_y; + float rod = orig[0]*rayn_x + orig[1]*rayn_y; + + float q10d = q1d - q0d; + float q20d = q2d - q0d; + float q0rd = q0d - rod; + + hits[0][0] = q0rd + s0*(2.0f - 2.0f*s0)*q10d + s0*s0*q20d; + hits[0][1] = a*s0+b; + + if (num_s > 1) { + hits[1][0] = q0rd + s1*(2.0f - 2.0f*s1)*q10d + s1*s1*q20d; + hits[1][1] = a*s1+b; + return 2; + } else { + return 1; + } + } +} + +static int equal(float *a, float *b) +{ + return (a[0] == b[0] && a[1] == b[1]); +} + +static int stbtt__compute_crossings_x(float x, float y, int nverts, stbtt_vertex *verts) +{ + int i; + float orig[2], ray[2] = { 1, 0 }; + float y_frac; + int winding = 0; + + orig[0] = x; + orig[1] = y; + + // make sure y never passes through a vertex of the shape + y_frac = (float) STBTT_fmod(y, 1.0f); + if (y_frac < 0.01f) + y += 0.01f; + else if (y_frac > 0.99f) + y -= 0.01f; + orig[1] = y; + + // test a ray from (-infinity,y) to (x,y) + for (i=0; i < nverts; ++i) { + if (verts[i].type == STBTT_vline) { + int x0 = (int) verts[i-1].x, y0 = (int) verts[i-1].y; + int x1 = (int) verts[i ].x, y1 = (int) verts[i ].y; + if (y > STBTT_min(y0,y1) && y < STBTT_max(y0,y1) && x > STBTT_min(x0,x1)) { + float x_inter = (y - y0) / (y1 - y0) * (x1-x0) + x0; + if (x_inter < x) + winding += (y0 < y1) ? 1 : -1; + } + } + if (verts[i].type == STBTT_vcurve) { + int x0 = (int) verts[i-1].x , y0 = (int) verts[i-1].y ; + int x1 = (int) verts[i ].cx, y1 = (int) verts[i ].cy; + int x2 = (int) verts[i ].x , y2 = (int) verts[i ].y ; + int ax = STBTT_min(x0,STBTT_min(x1,x2)), ay = STBTT_min(y0,STBTT_min(y1,y2)); + int by = STBTT_max(y0,STBTT_max(y1,y2)); + if (y > ay && y < by && x > ax) { + float q0[2],q1[2],q2[2]; + float hits[2][2]; + q0[0] = (float)x0; + q0[1] = (float)y0; + q1[0] = (float)x1; + q1[1] = (float)y1; + q2[0] = (float)x2; + q2[1] = (float)y2; + if (equal(q0,q1) || equal(q1,q2)) { + x0 = (int)verts[i-1].x; + y0 = (int)verts[i-1].y; + x1 = (int)verts[i ].x; + y1 = (int)verts[i ].y; + if (y > STBTT_min(y0,y1) && y < STBTT_max(y0,y1) && x > STBTT_min(x0,x1)) { + float x_inter = (y - y0) / (y1 - y0) * (x1-x0) + x0; + if (x_inter < x) + winding += (y0 < y1) ? 1 : -1; + } + } else { + int num_hits = stbtt__ray_intersect_bezier(orig, ray, q0, q1, q2, hits); + if (num_hits >= 1) + if (hits[0][0] < 0) + winding += (hits[0][1] < 0 ? -1 : 1); + if (num_hits >= 2) + if (hits[1][0] < 0) + winding += (hits[1][1] < 0 ? -1 : 1); + } + } + } + } + return winding; +} + +static float stbtt__cuberoot( float x ) +{ + if (x<0) + return -(float) STBTT_pow(-x,1.0f/3.0f); + else + return (float) STBTT_pow( x,1.0f/3.0f); +} + +// x^3 + c*x^2 + b*x + a = 0 +static int stbtt__solve_cubic(float a, float b, float c, float* r) +{ + float s = -a / 3; + float p = b - a*a / 3; + float q = a * (2*a*a - 9*b) / 27 + c; + float p3 = p*p*p; + float d = q*q + 4*p3 / 27; + if (d >= 0) { + float z = (float) STBTT_sqrt(d); + float u = (-q + z) / 2; + float v = (-q - z) / 2; + u = stbtt__cuberoot(u); + v = stbtt__cuberoot(v); + r[0] = s + u + v; + return 1; + } else { + float u = (float) STBTT_sqrt(-p/3); + float v = (float) STBTT_acos(-STBTT_sqrt(-27/p3) * q / 2) / 3; // p3 must be negative, since d is negative + float m = (float) STBTT_cos(v); + float n = (float) STBTT_cos(v-3.141592/2)*1.732050808f; + r[0] = s + u * 2 * m; + r[1] = s - u * (m + n); + r[2] = s - u * (m - n); + + //STBTT_assert( STBTT_fabs(((r[0]+a)*r[0]+b)*r[0]+c) < 0.05f); // these asserts may not be safe at all scales, though they're in bezier t parameter units so maybe? + //STBTT_assert( STBTT_fabs(((r[1]+a)*r[1]+b)*r[1]+c) < 0.05f); + //STBTT_assert( STBTT_fabs(((r[2]+a)*r[2]+b)*r[2]+c) < 0.05f); + return 3; + } +} + +STBTT_DEF unsigned char * stbtt_GetGlyphSDF(const stbtt_fontinfo *info, float scale, int glyph, int padding, unsigned char onedge_value, float pixel_dist_scale, int *width, int *height, int *xoff, int *yoff) +{ + float scale_x = scale, scale_y = scale; + int ix0,iy0,ix1,iy1; + int w,h; + unsigned char *data; + + if (scale == 0) return NULL; + + stbtt_GetGlyphBitmapBoxSubpixel(info, glyph, scale, scale, 0.0f,0.0f, &ix0,&iy0,&ix1,&iy1); + + // if empty, return NULL + if (ix0 == ix1 || iy0 == iy1) + return NULL; + + ix0 -= padding; + iy0 -= padding; + ix1 += padding; + iy1 += padding; + + w = (ix1 - ix0); + h = (iy1 - iy0); + + if (width ) *width = w; + if (height) *height = h; + if (xoff ) *xoff = ix0; + if (yoff ) *yoff = iy0; + + // invert for y-downwards bitmaps + scale_y = -scale_y; + + { + int x,y,i,j; + float *precompute; + stbtt_vertex *verts; + int num_verts = stbtt_GetGlyphShape(info, glyph, &verts); + data = (unsigned char *) STBTT_malloc(w * h, info->userdata); + precompute = (float *) STBTT_malloc(num_verts * sizeof(float), info->userdata); + + for (i=0,j=num_verts-1; i < num_verts; j=i++) { + if (verts[i].type == STBTT_vline) { + float x0 = verts[i].x*scale_x, y0 = verts[i].y*scale_y; + float x1 = verts[j].x*scale_x, y1 = verts[j].y*scale_y; + float dist = (float) STBTT_sqrt((x1-x0)*(x1-x0) + (y1-y0)*(y1-y0)); + precompute[i] = (dist == 0) ? 0.0f : 1.0f / dist; + } else if (verts[i].type == STBTT_vcurve) { + float x2 = verts[j].x *scale_x, y2 = verts[j].y *scale_y; + float x1 = verts[i].cx*scale_x, y1 = verts[i].cy*scale_y; + float x0 = verts[i].x *scale_x, y0 = verts[i].y *scale_y; + float bx = x0 - 2*x1 + x2, by = y0 - 2*y1 + y2; + float len2 = bx*bx + by*by; + if (len2 != 0.0f) + precompute[i] = 1.0f / (bx*bx + by*by); + else + precompute[i] = 0.0f; + } else + precompute[i] = 0.0f; + } + + for (y=iy0; y < iy1; ++y) { + for (x=ix0; x < ix1; ++x) { + float val; + float min_dist = 999999.0f; + float sx = (float) x + 0.5f; + float sy = (float) y + 0.5f; + float x_gspace = (sx / scale_x); + float y_gspace = (sy / scale_y); + + int winding = stbtt__compute_crossings_x(x_gspace, y_gspace, num_verts, verts); // @OPTIMIZE: this could just be a rasterization, but needs to be line vs. non-tesselated curves so a new path + + for (i=0; i < num_verts; ++i) { + float x0 = verts[i].x*scale_x, y0 = verts[i].y*scale_y; + + // check against every point here rather than inside line/curve primitives -- @TODO: wrong if multiple 'moves' in a row produce a garbage point, and given culling, probably more efficient to do within line/curve + float dist2 = (x0-sx)*(x0-sx) + (y0-sy)*(y0-sy); + if (dist2 < min_dist*min_dist) + min_dist = (float) STBTT_sqrt(dist2); + + if (verts[i].type == STBTT_vline) { + float x1 = verts[i-1].x*scale_x, y1 = verts[i-1].y*scale_y; + + // coarse culling against bbox + //if (sx > STBTT_min(x0,x1)-min_dist && sx < STBTT_max(x0,x1)+min_dist && + // sy > STBTT_min(y0,y1)-min_dist && sy < STBTT_max(y0,y1)+min_dist) + float dist = (float) STBTT_fabs((x1-x0)*(y0-sy) - (y1-y0)*(x0-sx)) * precompute[i]; + STBTT_assert(i != 0); + if (dist < min_dist) { + // check position along line + // x' = x0 + t*(x1-x0), y' = y0 + t*(y1-y0) + // minimize (x'-sx)*(x'-sx)+(y'-sy)*(y'-sy) + float dx = x1-x0, dy = y1-y0; + float px = x0-sx, py = y0-sy; + // minimize (px+t*dx)^2 + (py+t*dy)^2 = px*px + 2*px*dx*t + t^2*dx*dx + py*py + 2*py*dy*t + t^2*dy*dy + // derivative: 2*px*dx + 2*py*dy + (2*dx*dx+2*dy*dy)*t, set to 0 and solve + float t = -(px*dx + py*dy) / (dx*dx + dy*dy); + if (t >= 0.0f && t <= 1.0f) + min_dist = dist; + } + } else if (verts[i].type == STBTT_vcurve) { + float x2 = verts[i-1].x *scale_x, y2 = verts[i-1].y *scale_y; + float x1 = verts[i ].cx*scale_x, y1 = verts[i ].cy*scale_y; + float box_x0 = STBTT_min(STBTT_min(x0,x1),x2); + float box_y0 = STBTT_min(STBTT_min(y0,y1),y2); + float box_x1 = STBTT_max(STBTT_max(x0,x1),x2); + float box_y1 = STBTT_max(STBTT_max(y0,y1),y2); + // coarse culling against bbox to avoid computing cubic unnecessarily + if (sx > box_x0-min_dist && sx < box_x1+min_dist && sy > box_y0-min_dist && sy < box_y1+min_dist) { + int num=0; + float ax = x1-x0, ay = y1-y0; + float bx = x0 - 2*x1 + x2, by = y0 - 2*y1 + y2; + float mx = x0 - sx, my = y0 - sy; + float res[3],px,py,t,it; + float a_inv = precompute[i]; + if (a_inv == 0.0) { // if a_inv is 0, it's 2nd degree so use quadratic formula + float a = 3*(ax*bx + ay*by); + float b = 2*(ax*ax + ay*ay) + (mx*bx+my*by); + float c = mx*ax+my*ay; + if (a == 0.0) { // if a is 0, it's linear + if (b != 0.0) { + res[num++] = -c/b; + } + } else { + float discriminant = b*b - 4*a*c; + if (discriminant < 0) + num = 0; + else { + float root = (float) STBTT_sqrt(discriminant); + res[0] = (-b - root)/(2*a); + res[1] = (-b + root)/(2*a); + num = 2; // don't bother distinguishing 1-solution case, as code below will still work + } + } + } else { + float b = 3*(ax*bx + ay*by) * a_inv; // could precompute this as it doesn't depend on sample point + float c = (2*(ax*ax + ay*ay) + (mx*bx+my*by)) * a_inv; + float d = (mx*ax+my*ay) * a_inv; + num = stbtt__solve_cubic(b, c, d, res); + } + if (num >= 1 && res[0] >= 0.0f && res[0] <= 1.0f) { + t = res[0], it = 1.0f - t; + px = it*it*x0 + 2*t*it*x1 + t*t*x2; + py = it*it*y0 + 2*t*it*y1 + t*t*y2; + dist2 = (px-sx)*(px-sx) + (py-sy)*(py-sy); + if (dist2 < min_dist * min_dist) + min_dist = (float) STBTT_sqrt(dist2); + } + if (num >= 2 && res[1] >= 0.0f && res[1] <= 1.0f) { + t = res[1], it = 1.0f - t; + px = it*it*x0 + 2*t*it*x1 + t*t*x2; + py = it*it*y0 + 2*t*it*y1 + t*t*y2; + dist2 = (px-sx)*(px-sx) + (py-sy)*(py-sy); + if (dist2 < min_dist * min_dist) + min_dist = (float) STBTT_sqrt(dist2); + } + if (num >= 3 && res[2] >= 0.0f && res[2] <= 1.0f) { + t = res[2], it = 1.0f - t; + px = it*it*x0 + 2*t*it*x1 + t*t*x2; + py = it*it*y0 + 2*t*it*y1 + t*t*y2; + dist2 = (px-sx)*(px-sx) + (py-sy)*(py-sy); + if (dist2 < min_dist * min_dist) + min_dist = (float) STBTT_sqrt(dist2); + } + } + } + } + if (winding == 0) + min_dist = -min_dist; // if outside the shape, value is negative + val = onedge_value + pixel_dist_scale * min_dist; + if (val < 0) + val = 0; + else if (val > 255) + val = 255; + data[(y-iy0)*w+(x-ix0)] = (unsigned char) val; + } + } + STBTT_free(precompute, info->userdata); + STBTT_free(verts, info->userdata); + } + return data; +} + +STBTT_DEF unsigned char * stbtt_GetCodepointSDF(const stbtt_fontinfo *info, float scale, int codepoint, int padding, unsigned char onedge_value, float pixel_dist_scale, int *width, int *height, int *xoff, int *yoff) +{ + return stbtt_GetGlyphSDF(info, scale, stbtt_FindGlyphIndex(info, codepoint), padding, onedge_value, pixel_dist_scale, width, height, xoff, yoff); +} + +STBTT_DEF void stbtt_FreeSDF(unsigned char *bitmap, void *userdata) +{ + STBTT_free(bitmap, userdata); +} + +////////////////////////////////////////////////////////////////////////////// +// +// font name matching -- recommended not to use this +// + +// check if a utf8 string contains a prefix which is the utf16 string; if so return length of matching utf8 string +static stbtt_int32 stbtt__CompareUTF8toUTF16_bigendian_prefix(stbtt_uint8 *s1, stbtt_int32 len1, stbtt_uint8 *s2, stbtt_int32 len2) +{ + stbtt_int32 i=0; + + // convert utf16 to utf8 and compare the results while converting + while (len2) { + stbtt_uint16 ch = s2[0]*256 + s2[1]; + if (ch < 0x80) { + if (i >= len1) return -1; + if (s1[i++] != ch) return -1; + } else if (ch < 0x800) { + if (i+1 >= len1) return -1; + if (s1[i++] != 0xc0 + (ch >> 6)) return -1; + if (s1[i++] != 0x80 + (ch & 0x3f)) return -1; + } else if (ch >= 0xd800 && ch < 0xdc00) { + stbtt_uint32 c; + stbtt_uint16 ch2 = s2[2]*256 + s2[3]; + if (i+3 >= len1) return -1; + c = ((ch - 0xd800) << 10) + (ch2 - 0xdc00) + 0x10000; + if (s1[i++] != 0xf0 + (c >> 18)) return -1; + if (s1[i++] != 0x80 + ((c >> 12) & 0x3f)) return -1; + if (s1[i++] != 0x80 + ((c >> 6) & 0x3f)) return -1; + if (s1[i++] != 0x80 + ((c ) & 0x3f)) return -1; + s2 += 2; // plus another 2 below + len2 -= 2; + } else if (ch >= 0xdc00 && ch < 0xe000) { + return -1; + } else { + if (i+2 >= len1) return -1; + if (s1[i++] != 0xe0 + (ch >> 12)) return -1; + if (s1[i++] != 0x80 + ((ch >> 6) & 0x3f)) return -1; + if (s1[i++] != 0x80 + ((ch ) & 0x3f)) return -1; + } + s2 += 2; + len2 -= 2; + } + return i; +} + +static int stbtt_CompareUTF8toUTF16_bigendian_internal(char *s1, int len1, char *s2, int len2) +{ + return len1 == stbtt__CompareUTF8toUTF16_bigendian_prefix((stbtt_uint8*) s1, len1, (stbtt_uint8*) s2, len2); +} + +// returns results in whatever encoding you request... but note that 2-byte encodings +// will be BIG-ENDIAN... use stbtt_CompareUTF8toUTF16_bigendian() to compare +STBTT_DEF const char *stbtt_GetFontNameString(const stbtt_fontinfo *font, int *length, int platformID, int encodingID, int languageID, int nameID) +{ + stbtt_int32 i,count,stringOffset; + stbtt_uint8 *fc = font->data; + stbtt_uint32 offset = font->fontstart; + stbtt_uint32 nm = stbtt__find_table(fc, offset, "name"); + if (!nm) return NULL; + + count = ttUSHORT(fc+nm+2); + stringOffset = nm + ttUSHORT(fc+nm+4); + for (i=0; i < count; ++i) { + stbtt_uint32 loc = nm + 6 + 12 * i; + if (platformID == ttUSHORT(fc+loc+0) && encodingID == ttUSHORT(fc+loc+2) + && languageID == ttUSHORT(fc+loc+4) && nameID == ttUSHORT(fc+loc+6)) { + *length = ttUSHORT(fc+loc+8); + return (const char *) (fc+stringOffset+ttUSHORT(fc+loc+10)); + } + } + return NULL; +} + +static int stbtt__matchpair(stbtt_uint8 *fc, stbtt_uint32 nm, stbtt_uint8 *name, stbtt_int32 nlen, stbtt_int32 target_id, stbtt_int32 next_id) +{ + stbtt_int32 i; + stbtt_int32 count = ttUSHORT(fc+nm+2); + stbtt_int32 stringOffset = nm + ttUSHORT(fc+nm+4); + + for (i=0; i < count; ++i) { + stbtt_uint32 loc = nm + 6 + 12 * i; + stbtt_int32 id = ttUSHORT(fc+loc+6); + if (id == target_id) { + // find the encoding + stbtt_int32 platform = ttUSHORT(fc+loc+0), encoding = ttUSHORT(fc+loc+2), language = ttUSHORT(fc+loc+4); + + // is this a Unicode encoding? + if (platform == 0 || (platform == 3 && encoding == 1) || (platform == 3 && encoding == 10)) { + stbtt_int32 slen = ttUSHORT(fc+loc+8); + stbtt_int32 off = ttUSHORT(fc+loc+10); + + // check if there's a prefix match + stbtt_int32 matchlen = stbtt__CompareUTF8toUTF16_bigendian_prefix(name, nlen, fc+stringOffset+off,slen); + if (matchlen >= 0) { + // check for target_id+1 immediately following, with same encoding & language + if (i+1 < count && ttUSHORT(fc+loc+12+6) == next_id && ttUSHORT(fc+loc+12) == platform && ttUSHORT(fc+loc+12+2) == encoding && ttUSHORT(fc+loc+12+4) == language) { + slen = ttUSHORT(fc+loc+12+8); + off = ttUSHORT(fc+loc+12+10); + if (slen == 0) { + if (matchlen == nlen) + return 1; + } else if (matchlen < nlen && name[matchlen] == ' ') { + ++matchlen; + if (stbtt_CompareUTF8toUTF16_bigendian_internal((char*) (name+matchlen), nlen-matchlen, (char*)(fc+stringOffset+off),slen)) + return 1; + } + } else { + // if nothing immediately following + if (matchlen == nlen) + return 1; + } + } + } + + // @TODO handle other encodings + } + } + return 0; +} + +static int stbtt__matches(stbtt_uint8 *fc, stbtt_uint32 offset, stbtt_uint8 *name, stbtt_int32 flags) +{ + stbtt_int32 nlen = (stbtt_int32) STBTT_strlen((char *) name); + stbtt_uint32 nm,hd; + if (!stbtt__isfont(fc+offset)) return 0; + + // check italics/bold/underline flags in macStyle... + if (flags) { + hd = stbtt__find_table(fc, offset, "head"); + if ((ttUSHORT(fc+hd+44) & 7) != (flags & 7)) return 0; + } + + nm = stbtt__find_table(fc, offset, "name"); + if (!nm) return 0; + + if (flags) { + // if we checked the macStyle flags, then just check the family and ignore the subfamily + if (stbtt__matchpair(fc, nm, name, nlen, 16, -1)) return 1; + if (stbtt__matchpair(fc, nm, name, nlen, 1, -1)) return 1; + if (stbtt__matchpair(fc, nm, name, nlen, 3, -1)) return 1; + } else { + if (stbtt__matchpair(fc, nm, name, nlen, 16, 17)) return 1; + if (stbtt__matchpair(fc, nm, name, nlen, 1, 2)) return 1; + if (stbtt__matchpair(fc, nm, name, nlen, 3, -1)) return 1; + } + + return 0; +} + +static int stbtt_FindMatchingFont_internal(unsigned char *font_collection, char *name_utf8, stbtt_int32 flags) +{ + stbtt_int32 i; + for (i=0;;++i) { + stbtt_int32 off = stbtt_GetFontOffsetForIndex(font_collection, i); + if (off < 0) return off; + if (stbtt__matches((stbtt_uint8 *) font_collection, off, (stbtt_uint8*) name_utf8, flags)) + return off; + } +} + +#if defined(__GNUC__) || defined(__clang__) +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wcast-qual" +#endif + +STBTT_DEF int stbtt_BakeFontBitmap(const unsigned char *data, int offset, + float pixel_height, unsigned char *pixels, int pw, int ph, + int first_char, int num_chars, stbtt_bakedchar *chardata) +{ + return stbtt_BakeFontBitmap_internal((unsigned char *) data, offset, pixel_height, pixels, pw, ph, first_char, num_chars, chardata); +} + +STBTT_DEF int stbtt_GetFontOffsetForIndex(const unsigned char *data, int index) +{ + return stbtt_GetFontOffsetForIndex_internal((unsigned char *) data, index); +} + +STBTT_DEF int stbtt_GetNumberOfFonts(const unsigned char *data) +{ + return stbtt_GetNumberOfFonts_internal((unsigned char *) data); +} + +STBTT_DEF int stbtt_InitFont(stbtt_fontinfo *info, const unsigned char *data, int offset) +{ + return stbtt_InitFont_internal(info, (unsigned char *) data, offset); +} + +STBTT_DEF int stbtt_FindMatchingFont(const unsigned char *fontdata, const char *name, int flags) +{ + return stbtt_FindMatchingFont_internal((unsigned char *) fontdata, (char *) name, flags); +} + +STBTT_DEF int stbtt_CompareUTF8toUTF16_bigendian(const char *s1, int len1, const char *s2, int len2) +{ + return stbtt_CompareUTF8toUTF16_bigendian_internal((char *) s1, len1, (char *) s2, len2); +} + +#if defined(__GNUC__) || defined(__clang__) +#pragma GCC diagnostic pop +#endif + +#endif // STB_TRUETYPE_IMPLEMENTATION + + +// FULL VERSION HISTORY +// +// 1.19 (2018-02-11) OpenType GPOS kerning (horizontal only), STBTT_fmod +// 1.18 (2018-01-29) add missing function +// 1.17 (2017-07-23) make more arguments const; doc fix +// 1.16 (2017-07-12) SDF support +// 1.15 (2017-03-03) make more arguments const +// 1.14 (2017-01-16) num-fonts-in-TTC function +// 1.13 (2017-01-02) support OpenType fonts, certain Apple fonts +// 1.12 (2016-10-25) suppress warnings about casting away const with -Wcast-qual +// 1.11 (2016-04-02) fix unused-variable warning +// 1.10 (2016-04-02) allow user-defined fabs() replacement +// fix memory leak if fontsize=0.0 +// fix warning from duplicate typedef +// 1.09 (2016-01-16) warning fix; avoid crash on outofmem; use alloc userdata for PackFontRanges +// 1.08 (2015-09-13) document stbtt_Rasterize(); fixes for vertical & horizontal edges +// 1.07 (2015-08-01) allow PackFontRanges to accept arrays of sparse codepoints; +// allow PackFontRanges to pack and render in separate phases; +// fix stbtt_GetFontOFfsetForIndex (never worked for non-0 input?); +// fixed an assert() bug in the new rasterizer +// replace assert() with STBTT_assert() in new rasterizer +// 1.06 (2015-07-14) performance improvements (~35% faster on x86 and x64 on test machine) +// also more precise AA rasterizer, except if shapes overlap +// remove need for STBTT_sort +// 1.05 (2015-04-15) fix misplaced definitions for STBTT_STATIC +// 1.04 (2015-04-15) typo in example +// 1.03 (2015-04-12) STBTT_STATIC, fix memory leak in new packing, various fixes +// 1.02 (2014-12-10) fix various warnings & compile issues w/ stb_rect_pack, C++ +// 1.01 (2014-12-08) fix subpixel position when oversampling to exactly match +// non-oversampled; STBTT_POINT_SIZE for packed case only +// 1.00 (2014-12-06) add new PackBegin etc. API, w/ support for oversampling +// 0.99 (2014-09-18) fix multiple bugs with subpixel rendering (ryg) +// 0.9 (2014-08-07) support certain mac/iOS fonts without an MS platformID +// 0.8b (2014-07-07) fix a warning +// 0.8 (2014-05-25) fix a few more warnings +// 0.7 (2013-09-25) bugfix: subpixel glyph bug fixed in 0.5 had come back +// 0.6c (2012-07-24) improve documentation +// 0.6b (2012-07-20) fix a few more warnings +// 0.6 (2012-07-17) fix warnings; added stbtt_ScaleForMappingEmToPixels, +// stbtt_GetFontBoundingBox, stbtt_IsGlyphEmpty +// 0.5 (2011-12-09) bugfixes: +// subpixel glyph renderer computed wrong bounding box +// first vertex of shape can be off-curve (FreeSans) +// 0.4b (2011-12-03) fixed an error in the font baking example +// 0.4 (2011-12-01) kerning, subpixel rendering (tor) +// bugfixes for: +// codepoint-to-glyph conversion using table fmt=12 +// codepoint-to-glyph conversion using table fmt=4 +// stbtt_GetBakedQuad with non-square texture (Zer) +// updated Hello World! sample to use kerning and subpixel +// fixed some warnings +// 0.3 (2009-06-24) cmap fmt=12, compound shapes (MM) +// userdata, malloc-from-userdata, non-zero fill (stb) +// 0.2 (2009-03-11) Fix unsigned/signed char warnings +// 0.1 (2009-03-09) First public release +// + +/* +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person obtaining a copy of +this software and associated documentation files (the "Software"), to deal in +the Software without restriction, including without limitation the rights to +use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies +of the Software, and to permit persons to whom the Software is furnished to do +so, subject to the following conditions: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ diff --git a/src/ui.c b/src/ui.c new file mode 100644 index 0000000..c717d51 --- /dev/null +++ b/src/ui.c @@ -0,0 +1,1332 @@ +/************************************************************//** +* +* @file: ui.c +* @author: Martin Fouilleul +* @date: 08/08/2022 +* @revision: +* +*****************************************************************/ +#include"memory.h" +#include"hash.h" +#include"platform_clock.h" +#include"ui.h" + +#define LOG_SUBSYSTEM "UI" + +//----------------------------------------------------------------------------- +// context +//----------------------------------------------------------------------------- + +const u32 UI_MAX_INPUT_CHAR_PER_FRAME = 64; + +typedef struct ui_input_text +{ + u8 count; + utf32 codePoints[UI_MAX_INPUT_CHAR_PER_FRAME]; + +} ui_input_text; + +typedef struct ui_stack_elt ui_stack_elt; +struct ui_stack_elt +{ + ui_stack_elt* parent; + union + { + ui_box* box; + ui_size size; + mp_rect clip; + ui_style style; + }; +}; + +const u64 UI_BOX_MAP_BUCKET_COUNT = 1024; + +typedef struct ui_context +{ + bool init; + + f32 width; + f32 height; + + u64 frameCounter; + f64 frameTime; + f64 lastFrameDuration; + + mem_arena frameArena; + mem_pool boxPool; + list_info boxMap[UI_BOX_MAP_BUCKET_COUNT]; + + ui_box* root; + ui_box* overlay; + ui_stack_elt* boxStack; + ui_stack_elt* sizeStack[UI_AXIS_COUNT]; + ui_stack_elt* styleStack[UI_STYLE_SELECTOR_COUNT]; + ui_stack_elt* clipStack; + + u32 z; + ui_box* hovered; + +} ui_context; + +__thread ui_context __uiContext = {0}; + +ui_context* ui_get_context() +{ + return(&__uiContext); +} + +//----------------------------------------------------------------------------- +// stacks +//----------------------------------------------------------------------------- +ui_stack_elt* ui_stack_push(ui_context* ui, ui_stack_elt** stack) +{ + ui_stack_elt* elt = mem_arena_alloc_type(&ui->frameArena, ui_stack_elt); + memset(elt, 0, sizeof(ui_stack_elt)); + elt->parent = *stack; + *stack = elt; + return(elt); +} + +void ui_stack_pop(ui_stack_elt** stack) +{ + if(*stack) + { + *stack = (*stack)->parent; + } + else + { + LOG_ERROR("ui stack underflow\n"); + } +} + +void ui_size_push(ui_axis axis, ui_size_kind kind, f32 value, f32 strictness) +{ + ui_context* ui = ui_get_context(); + ui_stack_elt* elt = ui_stack_push(ui, &ui->sizeStack[axis]); + elt->size = (ui_size){.kind = kind, .value = value, .strictness = strictness}; +} + +void ui_size_pop(ui_axis axis) +{ + ui_context* ui = ui_get_context(); + ui_stack_pop(&ui->sizeStack[axis]); +} + +ui_size ui_size_top(ui_axis axis) +{ + ui_context* ui = ui_get_context(); + ui_size size = {0}; + if(ui->sizeStack[axis]) + { + size = ui->sizeStack[axis]->size; + } + return(size); +} + +ui_stack_elt** ui_style_stack_select(ui_context* ui, ui_style_selector selector) +{ + DEBUG_ASSERT(selector < UI_STYLE_SELECTOR_COUNT); + ui_stack_elt** stack = &ui->styleStack[selector]; + return(stack); +} + +void ui_style_push(ui_style_selector selector, ui_style style) +{ + ui_context* ui = ui_get_context(); + ui_stack_elt** stack = ui_style_stack_select(ui, selector); + ui_stack_elt* elt = ui_stack_push(ui, stack); + elt->style = style; +} + +void ui_style_pop(ui_style_selector selector) +{ + ui_context* ui = ui_get_context(); + ui_stack_elt** stack = ui_style_stack_select(ui, selector); + ui_stack_pop(stack); +} + +ui_style* ui_style_top(ui_style_selector selector) +{ + ui_context* ui = ui_get_context(); + ui_stack_elt* stack = *ui_style_stack_select(ui, selector); + ui_style* style = 0; + if(stack) + { + style = &stack->style; + } + return(style); +} + +mp_rect ui_intersect_rects(mp_rect lhs, mp_rect rhs) +{ + //NOTE(martin): intersect with current clip + f32 x0 = maximum(lhs.x, rhs.x); + f32 y0 = maximum(lhs.y, rhs.y); + f32 x1 = minimum(lhs.x + lhs.w, rhs.x + rhs.w); + f32 y1 = minimum(lhs.y + lhs.h, rhs.y + rhs.h); + mp_rect r = {x0, y0, maximum(0, x1-x0), maximum(0, y1-y0)}; + return(r); +} + +mp_rect ui_clip_top() +{ + mp_rect r = {-FLT_MAX/2, -FLT_MAX/2, FLT_MAX, FLT_MAX}; + ui_context* ui = ui_get_context(); + ui_stack_elt* elt = ui->clipStack; + if(elt) + { + r = elt->clip; + } + return(r); +} + +void ui_clip_push(mp_rect clip) +{ + ui_context* ui = ui_get_context(); + mp_rect current = ui_clip_top(); + ui_stack_elt* elt = ui_stack_push(ui, &ui->clipStack); + elt->clip = ui_intersect_rects(current, clip); +} + +void ui_clip_pop() +{ + ui_context* ui = ui_get_context(); + ui_stack_pop(&ui->clipStack); +} + +ui_box* ui_box_top() +{ + ui_context* ui = ui_get_context(); + ui_stack_elt* elt = ui->boxStack; + ui_box* box = elt ? elt->box : 0; + return(box); +} + +void ui_box_push(ui_box* box) +{ + ui_context* ui = ui_get_context(); + ui_stack_elt* elt = ui_stack_push(ui, &ui->boxStack); + elt->box = box; + if(box->flags & UI_FLAG_CLIP) + { + ui_clip_push(box->rect); + } +} + +void ui_box_pop() +{ + ui_context* ui = ui_get_context(); + ui_box* box = ui_box_top(); + if(box->flags & UI_FLAG_CLIP) + { + ui_clip_pop(); + } + + ui_stack_pop(&ui->boxStack); +} + +//----------------------------------------------------------------------------- +// key hashing and caching +//----------------------------------------------------------------------------- +ui_key ui_key_from_string(str8 string) +{ + ui_context* ui = ui_get_context(); + u64 seed = 0; + ui_box* parent = ui_box_top(); + if(parent) + { + seed = parent->key.hash; + } + + ui_key key = {}; + key.hash = mp_hash_aes_string_seed(string, seed); + return(key); +} + +bool ui_key_equal(ui_key a, ui_key b) +{ + return(a.hash == b.hash); +} + +void ui_box_cache(ui_context* ui, ui_box* box) +{ + u64 index = box->key.hash & (UI_BOX_MAP_BUCKET_COUNT-1); + ListAppend(&(ui->boxMap[index]), &box->bucketElt); +} + +ui_box* ui_box_lookup(ui_context* ui, ui_key key) +{ + u64 index = key.hash & (UI_BOX_MAP_BUCKET_COUNT-1); + + for_each_in_list(&ui->boxMap[index], box, ui_box, bucketElt) + { + if(ui_key_equal(key, box->key)) + { + return(box); + } + } + return(0); +} + +//----------------------------------------------------------------------------- +// ui boxes +//----------------------------------------------------------------------------- + +bool ui_rect_hit(mp_rect r, vec2 p) +{ + return( (p.x > r.x) + &&(p.x < r.x + r.w) + &&(p.y > r.y) + &&(p.y < r.y + r.h)); +} + +bool ui_box_hovering(ui_box* box, vec2 p) +{ + ui_context* ui = ui_get_context(); + + mp_rect clip = ui_clip_top(); + mp_rect rect = ui_intersect_rects(clip, box->rect); + bool hit = ui_rect_hit(rect, p); + bool result = hit && (!ui->hovered || box->z >= ui->hovered->z); + return(result); +} + +vec2 ui_mouse_position() +{ + ui_context* ui = ui_get_context(); + + vec2 mousePos = mp_input_mouse_position(); + mousePos.y = ui->height - mousePos.y; + return(mousePos); +} + +vec2 ui_mouse_delta() +{ + ui_context* ui = ui_get_context(); + vec2 delta = mp_input_mouse_delta(); + delta.y *= -1.; + return(delta); +} + +vec2 ui_mouse_wheel() +{ + ui_context* ui = ui_get_context(); + vec2 delta = mp_input_mouse_wheel(); + delta.y *= -1.; + return(delta); +} + +void ui_box_compute_signals(ui_context* ui, ui_box* box) +{ + ui_sig* sig = mem_arena_alloc_type(&ui->frameArena, ui_sig); + memset(sig, 0, sizeof(ui_sig)); + + if(!box->closed && !box->parentClosed) + { + vec2 mousePos = ui_mouse_position(); + + sig->hovering = ui_box_hovering(box, mousePos); + + if(box->flags & UI_FLAG_CLICKABLE) + { + if(sig->hovering) + { + sig->pressed = mp_input_mouse_pressed(MP_MOUSE_LEFT); + if(sig->pressed) + { + box->dragging = true; + } + + sig->clicked = mp_input_mouse_clicked(MP_MOUSE_LEFT); + sig->doubleClicked = mp_input_mouse_clicked(MP_MOUSE_LEFT); + } + + sig->released = mp_input_mouse_released(MP_MOUSE_LEFT); + if(sig->released) + { + if(box->dragging && sig->hovering) + { + sig->triggered = true; + } + } + + if(!mp_input_mouse_down(MP_MOUSE_LEFT)) + { + box->dragging = false; + } + + sig->dragging = box->dragging; + } + + sig->mouse = (vec2){mousePos.x - box->rect.x, mousePos.y - box->rect.y}; + sig->delta = ui_mouse_delta(); + sig->wheel = ui_mouse_wheel(); + } + box->sig = sig; +} + +ui_box* ui_box_make_str8(str8 string, ui_flags flags) +{ + ui_context* ui = ui_get_context(); + + ui_key key = ui_key_from_string(string); + ui_box* box = ui_box_lookup(ui, key); + + if(!box) + { + box = mem_pool_alloc_type(&ui->boxPool, ui_box); + memset(box, 0, sizeof(ui_box)); + + box->key = key; + ui_box_cache(ui, box); + } + + //NOTE: setup hierarchy + ListInit(&box->children); + box->parent = ui_box_top(); + if(box->parent) + { + ListAppend(&box->parent->children, &box->listElt); + box->parentClosed = box->parent->closed || box->parent->parentClosed; + } + + //NOTE: setup per-frame state + box->frameCounter = ui->frameCounter; + box->string = str8_push_copy(&ui->frameArena, string); + box->flags = flags; + + box->floating[UI_AXIS_X] = false; + box->floating[UI_AXIS_Y] = false; + + box->desiredSize[UI_AXIS_X] = ui_size_top(UI_AXIS_X); + box->desiredSize[UI_AXIS_Y] = ui_size_top(UI_AXIS_Y); + box->layout = box->parent ? box->parent->layout : (ui_layout){0}; + + for(int i=0; istyles[i] = ui_style_top(i); + } + box->styleSelector = UI_STYLE_NORMAL; + + //NOTE: compute input signals + ui_box_compute_signals(ui, box); + + return(box); +} + +ui_box* ui_box_make(const char* cstring, ui_flags flags) +{ + str8 string = str8_from_cstring((char*)cstring); + return(ui_box_make_str8(string, flags)); +} + +ui_box* ui_box_begin_str8(str8 string, ui_flags flags) +{ + ui_context* ui = ui_get_context(); + ui_box* box = ui_box_make_str8(string, flags); + ui_box_push(box); + return(box); +} + +ui_box* ui_box_begin(const char* cstring, ui_flags flags) +{ + str8 string = str8_from_cstring((char*)cstring); + return(ui_box_begin_str8(string, flags)); +} + +ui_box* ui_box_end() +{ + ui_context* ui = ui_get_context(); + ui_box* box = ui_box_top(); + DEBUG_ASSERT(box, "box stack underflow"); + + ui_box_pop(); + return(box); +} + +void ui_box_set_layout(ui_box* box, ui_axis axis, ui_align alignX, ui_align alignY) +{ + box->layout = (ui_layout){axis, {alignX, alignY}}; +} + +void ui_box_set_size(ui_box* box, ui_axis axis, ui_size_kind kind, f32 value, f32 strictness) +{ + box->desiredSize[axis] = (ui_size){kind, value, strictness}; +} + +void ui_box_set_floating(ui_box* box, ui_axis axis, f32 pos) +{ + box->floating[axis] = true; + box->rect.c[axis] = pos; +} + +void ui_box_set_style_selector(ui_box* box, ui_style_selector selector) +{ + box->styleSelector = selector; +} + +void ui_box_set_closed(ui_box* box, bool closed) +{ + box->closed = closed; +} + +bool ui_box_closed(ui_box* box) +{ + return(box->closed); +} + +ui_sig ui_box_sig(ui_box* box) +{ + return(*box->sig); +} + +bool ui_box_hidden(ui_box* box) +{ + return(box->closed || box->parentClosed); +} + +//----------------------------------------------------------------------------- +// Auto-layout +//----------------------------------------------------------------------------- + +void ui_animate_f32(ui_context* ui, f32* value, f32 target, f32 animationTime) +{ + if( animationTime < 1e-6 + || fabs(*value - target) < 0.001) + { + *value = target; + } + else + { + + /*NOTE: + we use the euler approximation for df/dt = alpha(target - f) + the implicit form is f(t) = target*(1-e^(-alpha*t)) for the rising front, + and f(t) = e^(-alpha*t) for the falling front (e.g. classic RC circuit charge/discharge) + + Here we bake alpha = 1/tau = -ln(0.05)/tr, with tr the rise time to 95% of target + */ + f32 alpha = 3/animationTime; + f32 dt = ui->lastFrameDuration; + + *value += (target - *value)*alpha*dt; + } +} + +void ui_animate_color(ui_context* ui, mg_color* color, mg_color target, f32 animationTime) +{ + for(int i=0; i<4; i++) + { + ui_animate_f32(ui, &color->c[i], target.c[i], animationTime); + } +} + +void ui_box_compute_styling(ui_context* ui, ui_box* box) +{ + ui_style* targetStyle = box->styles[box->styleSelector]; + if(!targetStyle) + { + targetStyle = box->styles[UI_STYLE_NORMAL]; + } + DEBUG_ASSERT(targetStyle); + + f32 animationTime = targetStyle->animationTime; + + /* + f32 transitionValue = 0; + + //NOTE: update transition values + if(box->styleSelector == UI_STYLE_ACTIVE) + { + //NOTE: transition from normal or hot to active + if(box->hotTransition) + { + //NOTE: transition from hot to active + baseStyle = box->styles[UI_STYLE_HOT]; + } + else + { + //NOTE: transition from normal to active + baseStyle = box->styles[UI_STYLE_NORMAL]; + } + + box->hotTransition = ui_update_transition(ui, box->activeTransition, 1, animationTime); + box->activeTransition = ui_update_transition(ui, box->activeTransition, 1, animationTime); + transitionValue = box->activeTransition; + } + else if(box->styleSelector == UI_STYLE_HOT) + { + box->activeTransition = ui_update_transition(ui, box->activeTransition, 0, animationTime); + box->hotTransition = ui_update_transition(ui, box->hotTransition, 1, animationTime); + + if(box->activeTransition) + { + //NOTE: transition from active to hot + transitionValue = box->activeTransition; + baseStyle = box->styles[UI_STYLE_ACTIVE]; + } + else + { + //NOTE: transition from normal to hot + transitionValue = box->hotTransition; + baseStyle = box->styles[UI_STYLE_NORMAL]; + } + } + else + { + //NOTE: transition from hot or active to normal + box->activeTransition = 0; + box->hotTransition = ui_update_transition(ui, box->hotTransition, 0, animationTime); + transitionValue = box->hotTransition; + } + */ + + //TODO: interpolate based on transition values + ui_animate_color(ui, &box->computedStyle.backgroundColor, targetStyle->backgroundColor, animationTime); + ui_animate_color(ui, &box->computedStyle.foregroundColor, targetStyle->foregroundColor, animationTime); + ui_animate_color(ui, &box->computedStyle.borderColor, targetStyle->borderColor, animationTime); + ui_animate_color(ui, &box->computedStyle.textColor, targetStyle->textColor, animationTime); + + box->computedStyle.font = targetStyle->font; + ui_animate_f32(ui, &box->computedStyle.fontSize, targetStyle->fontSize, animationTime); + ui_animate_f32(ui, &box->computedStyle.borderSize, targetStyle->borderSize, animationTime); + ui_animate_f32(ui, &box->computedStyle.roundness, targetStyle->roundness, animationTime); +} + +void ui_layout_prepass(ui_context* ui, ui_box* box) +{ + if(ui_box_hidden(box)) + { + return; + } + + //NOTE: compute styling and static sizes + ui_box_compute_styling(ui, box); + + ui_style* style = &box->computedStyle; + + mp_rect textBox = {}; + if( box->desiredSize[UI_AXIS_X].kind == UI_SIZE_TEXT + ||box->desiredSize[UI_AXIS_Y].kind == UI_SIZE_TEXT) + { + textBox = mg_text_bounding_box(style->font, style->fontSize, box->string); + } + + for(int i=0; idesiredSize[i]; + + if(size.kind == UI_SIZE_TEXT) + { + box->rect.c[2+i] = textBox.c[2+i]; + } + else if(size.kind == UI_SIZE_PIXELS) + { + box->rect.c[2+i] = size.value; + } + } + + for_each_in_list(&box->children, child, ui_box, listElt) + { + ui_layout_prepass(ui, child); + } +} + +void ui_layout_upward_dependent_size(ui_context* ui, ui_box* box, int axis) +{ + if(ui_box_hidden(box)) + { + return; + } + + ui_size* size = &box->desiredSize[axis]; + + if(size->kind == UI_SIZE_PARENT_RATIO) + { + ui_box* parent = box->parent; + if( parent + && parent->desiredSize[axis].kind != UI_SIZE_CHILDREN) + { + box->rect.c[2+axis] = parent->rect.c[2+axis] * size->value; + } + //TODO else? + } + + for_each_in_list(&box->children, child, ui_box, listElt) + { + ui_layout_upward_dependent_size(ui, child, axis); + } +} + +void ui_layout_downward_dependent_size(ui_context* ui, ui_box* box, int axis) +{ + f32 sum = 0; + if(box->layout.axis == axis) + { + for_each_in_list(&box->children, child, ui_box, listElt) + { + if(!ui_box_hidden(child)) + { + ui_layout_downward_dependent_size(ui, child, axis); + if(!child->floating[axis]) + { + sum += child->rect.c[2+axis]; + } + } + } + } + else + { + for_each_in_list(&box->children, child, ui_box, listElt) + { + if(!ui_box_hidden(child)) + { + ui_layout_downward_dependent_size(ui, child, axis); + if(!child->floating[axis]) + { + sum = maximum(sum, child->rect.c[2+axis]); + } + } + } + } + + box->childrenSum[axis] = sum; + + ui_size* size = &box->desiredSize[axis]; + if(size->kind == UI_SIZE_CHILDREN) + { + box->rect.c[2+axis] = sum; + } +} + +void ui_layout_solve_conflicts(ui_context* ui, ui_box* box, int axis) +{ + //TODO +} + +void ui_layout_compute_rect(ui_context* ui, ui_box* box, vec2 pos) +{ + if(ui_box_hidden(box)) + { + return; + } + + box->rect.x = pos.x; + box->rect.y = pos.y; + box->z = ui->z; + ui->z++; + + ui_axis layoutAxis = box->layout.axis; + ui_axis secondAxis = (layoutAxis == UI_AXIS_X) ? UI_AXIS_Y : UI_AXIS_X; + ui_align* align = box->layout.align; + + vec2 origin = {box->rect.x - box->scroll.x, + box->rect.y - box->scroll.y}; + vec2 currentPos = origin; + + vec2 contentsSize = {maximum(box->rect.w, box->childrenSum[UI_AXIS_X]), + maximum(box->rect.h, box->childrenSum[UI_AXIS_Y])}; + + for(int i=0; ichildrenSum[i]; + } + } + if(align[layoutAxis] == UI_ALIGN_CENTER) + { + currentPos.c[layoutAxis] += 0.5*(contentsSize.c[layoutAxis] - box->childrenSum[layoutAxis]); + } + + for_each_in_list(&box->children, child, ui_box, listElt) + { + if(align[secondAxis] == UI_ALIGN_CENTER) + { + currentPos.c[secondAxis] = origin.c[secondAxis] + 0.5*(contentsSize.c[secondAxis] - child->rect.c[2+secondAxis]); + } + + vec2 childPos = currentPos; + for(int i=0; ifloating[i]) + { + childPos.c[i] = origin.c[i] + child->rect.c[i]; + } + } + + ui_layout_compute_rect(ui, child, childPos); + + if(!child->floating[layoutAxis]) + { + currentPos.c[layoutAxis] += child->rect.c[2+layoutAxis]; + } + } +} + +void ui_layout_find_next_hovered_recursive(ui_context* ui, ui_box* box, vec2 p) +{ + if(ui_box_hidden(box)) + { + return; + } + + bool hit = ui_rect_hit(box->rect, p); + if(hit && (box->flags & UI_FLAG_BLOCK_MOUSE)) + { + ui->hovered = box; + } + if(hit || !(box->flags & UI_FLAG_CLIP)) + { + for_each_in_list(&box->children, child, ui_box, listElt) + { + ui_layout_find_next_hovered_recursive(ui, child, p); + } + } +} + +void ui_layout_find_next_hovered(ui_context* ui, vec2 p) +{ + ui->hovered = 0; + ui_layout_find_next_hovered_recursive(ui, ui->root, p); +} + +void ui_solve_layout(ui_context* ui) +{ + ui_layout_prepass(ui, ui->root); + + for(int axis=0; axisroot, axis); + ui_layout_downward_dependent_size(ui, ui->root, axis); + ui_layout_solve_conflicts(ui, ui->root, axis); + } + ui_layout_compute_rect(ui, ui->root, (vec2){0, 0}); + + vec2 p = ui_mouse_position(); + ui_layout_find_next_hovered(ui, p); +} + +//----------------------------------------------------------------------------- +// Drawing +//----------------------------------------------------------------------------- + +void ui_rectangle_fill(mg_canvas canvas, mp_rect rect, f32 roundness) +{ + if(roundness) + { + mg_rounded_rectangle_fill(canvas, rect.x, rect.y, rect.w, rect.h, roundness); + } + else + { + mg_rectangle_fill(canvas, rect.x, rect.y, rect.w, rect.h); + } +} + +void ui_rectangle_stroke(mg_canvas canvas, mp_rect rect, f32 roundness) +{ + if(roundness) + { + mg_rounded_rectangle_stroke(canvas, rect.x, rect.y, rect.w, rect.h, roundness); + } + else + { + mg_rectangle_stroke(canvas, rect.x, rect.y, rect.w, rect.h); + } +} + +void ui_draw_box(mg_canvas canvas, ui_box* box) +{ + if(ui_box_hidden(box)) + { + return; + } + + ui_style* style = &box->computedStyle; + + if(box->flags & UI_FLAG_CLIP) + { + mg_clip_push(canvas, box->rect.x, box->rect.y, box->rect.w, box->rect.h); + } + + if(box->flags & UI_FLAG_DRAW_BACKGROUND) + { + mg_set_color(canvas, style->backgroundColor); + ui_rectangle_fill(canvas, box->rect, style->roundness); + } + + for_each_in_list(&box->children, child, ui_box, listElt) + { + ui_draw_box(canvas, child); + } + + if(box->flags & UI_FLAG_DRAW_FOREGROUND) + { + mg_set_color(canvas, style->foregroundColor); + ui_rectangle_fill(canvas, box->rect, style->roundness); + } + + if(box->flags & UI_FLAG_DRAW_TEXT) + { + mp_rect textBox = mg_text_bounding_box(style->font, style->fontSize, box->string); + + f32 x = box->rect.x + 0.5*(box->rect.w - textBox.w); + f32 y = box->rect.y + 0.5*(box->rect.h - textBox.h) - textBox.y; + + mg_set_font(canvas, style->font); + mg_set_font_size(canvas, style->fontSize); + mg_set_color(canvas, style->textColor); + + mg_move_to(canvas, x, y); + mg_text_outlines(canvas, box->string); + mg_fill(canvas); + } + + if(box->flags & UI_FLAG_CLIP) + { + mg_clip_pop(canvas); + } + + if(box->flags & UI_FLAG_DRAW_BORDER) + { + mg_set_width(canvas, style->borderSize); + mg_set_color(canvas, style->borderColor); + ui_rectangle_stroke(canvas, box->rect, style->roundness); + } +} + +void ui_draw(mg_canvas canvas) +{ + ui_context* ui = ui_get_context(); + + //NOTE: draw + mg_mat2x3 transform = {1, 0, 0, + 0, -1, ui->height}; + + bool oldTextFlip = mg_get_text_flip(canvas); + mg_set_text_flip(canvas, true); + + mg_matrix_push(canvas, transform); + ui_draw_box(canvas, ui->root); + mg_matrix_pop(canvas); + + mg_set_text_flip(canvas, oldTextFlip); + + //TODO: restore flip?? +} + +//----------------------------------------------------------------------------- +// frame begin/end +//----------------------------------------------------------------------------- + +void ui_begin_frame(u32 width, u32 height, ui_style defaultStyle) +{ + ui_context* ui = ui_get_context(); + + mem_arena_clear(&ui->frameArena); + + ui->width = width; + ui->height = height; + ui->frameCounter++; + f64 time = mp_get_time(MP_CLOCK_MONOTONIC); + ui->lastFrameDuration = time - ui->frameTime; + ui->frameTime = time; + + ui->boxStack = 0; + ui->sizeStack[UI_AXIS_X] = 0; + ui->sizeStack[UI_AXIS_Y] = 0; + for(int i=0; istyleStack[i] = 0; + } + + ui->clipStack = 0; + ui->z = 0; + + ui_style_push(UI_STYLE_NORMAL, defaultStyle); + + ui_size_push(UI_AXIS_X, UI_SIZE_PIXELS, width, 0); + ui_size_push(UI_AXIS_Y, UI_SIZE_PIXELS, height, 0); + + ui->root = ui_box_begin("_root_", 0); + + ui_box* contents = ui_box_make("_contents_", 0); + ui_box_set_floating(contents, UI_AXIS_X, 0); + ui_box_set_floating(contents, UI_AXIS_Y, 0); + ui_box_set_layout(contents, UI_AXIS_Y, UI_ALIGN_START, UI_ALIGN_START); + + ui->overlay = ui_box_make("_overlay_", 0); + ui_box_set_floating(ui->overlay, UI_AXIS_X, 0); + ui_box_set_floating(ui->overlay, UI_AXIS_Y, 0); + + ui_size_pop(UI_AXIS_X); + ui_size_pop(UI_AXIS_Y); + + ui_box_push(contents); +} + +void ui_end_frame() +{ + ui_context* ui = ui_get_context(); + + ui_box_pop(); + + ui_box* box = ui_box_end(); + DEBUG_ASSERT(box == ui->root, "unbalanced box stack"); + + ui_style_pop(UI_STYLE_NORMAL); + + //NOTE: layout + ui_solve_layout(ui); + + //NOTE: prune unused boxes + for(int i=0; iboxMap[i], box, ui_box, bucketElt) + { + if(box->frameCounter < ui->frameCounter) + { + ListRemove(&ui->boxMap[i], &box->bucketElt); + } + } + } +} + +//----------------------------------------------------------------------------- +// Init / cleanup +//----------------------------------------------------------------------------- +void ui_init() +{ + ui_context* ui = ui_get_context(); + if(!ui->init) + { + memset(ui, 0, sizeof(ui_context)); + mem_arena_init(&ui->frameArena); + mem_pool_init(&ui->boxPool, sizeof(ui_box)); + ui->init = true; + } +} + +void ui_cleanup() +{ + ui_context* ui = ui_get_context(); + mem_arena_release(&ui->frameArena); + mem_pool_release(&ui->boxPool); + ui->init = false; +} + + +//----------------------------------------------------------------------------- +// Basic helpers +//----------------------------------------------------------------------------- + +ui_sig ui_label(const char* label) +{ + ui_flags flags = UI_FLAG_CLIP + | UI_FLAG_DRAW_TEXT; + ui_box* box = ui_box_make(label, flags); + ui_box_set_size(box, UI_AXIS_X, UI_SIZE_TEXT, 0, 0); + ui_box_set_size(box, UI_AXIS_Y, UI_SIZE_TEXT, 0, 0); + + ui_sig sig = ui_box_sig(box); + return(sig); +} + +ui_sig ui_button(const char* label) +{ + ui_flags flags = UI_FLAG_CLICKABLE + | UI_FLAG_CLIP + | UI_FLAG_DRAW_FOREGROUND + | UI_FLAG_DRAW_BORDER + | UI_FLAG_DRAW_TEXT + | UI_FLAG_HOT_ANIMATION + | UI_FLAG_ACTIVE_ANIMATION; + + ui_box* box = ui_box_make(label, flags); + ui_sig sig = ui_box_sig(box); + + if(sig.hovering) + { + ui_box_set_style_selector(box, UI_STYLE_HOT); + if(sig.dragging) + { + ui_box_set_style_selector(box, UI_STYLE_ACTIVE); + } + } + + return(sig); +} + +ui_box* ui_scrollbar(const char* label, f32 thumbRatio, f32* scrollValue) +{ + ui_box* frame = ui_box_begin(label, 0); + { + ui_axis trackAxis = (frame->rect.w > frame->rect.h) ? UI_AXIS_X : UI_AXIS_Y; + ui_axis secondAxis = (trackAxis == UI_AXIS_Y) ? UI_AXIS_X : UI_AXIS_Y; + + f32 roundness = 0.5*frame->rect.c[2+secondAxis]; + f32 animationTime = 0.5; + + ui_style style[UI_STYLE_SELECTOR_COUNT] = { + [UI_STYLE_NORMAL] = {.backgroundColor = {0, 0, 0, 0}, + .foregroundColor = {0, 0, 0, 0}, + .roundness = roundness, + .animationTime = animationTime}, + [UI_STYLE_HOT] = {.backgroundColor = {0, 0, 0, 0.5}, + .foregroundColor = {0, 0, 0, 0.7}, + .roundness = roundness, + .animationTime = animationTime}, + [UI_STYLE_ACTIVE] = {.backgroundColor = {0, 0, 0, 0.5}, + .foregroundColor = {0, 0, 0, 0.7}, + .roundness = roundness, + .animationTime = animationTime}}; + + for(int i=0; irect.c[2+trackAxis] - thumb->rect.c[2+trackAxis]; + f32 delta = thumbSig.delta.c[trackAxis]/trackExtents; + f32 oldValue = *scrollValue; + + *scrollValue += delta; + *scrollValue = Clamp(*scrollValue, 0, 1); + } + + ui_sig trackSig = ui_box_sig(track); + if(trackSig.hovering) + { + ui_box_set_style_selector(track, UI_STYLE_HOT); + ui_box_set_style_selector(thumb, UI_STYLE_HOT); + } + if(thumbSig.dragging) + { + ui_box_set_style_selector(track, UI_STYLE_ACTIVE); + ui_box_set_style_selector(thumb, UI_STYLE_ACTIVE); + } + + } ui_box_end(); + + return(frame); +} + +void ui_panel_begin(const char* name) +{ + ui_flags panelFlags = UI_FLAG_DRAW_BACKGROUND + | UI_FLAG_DRAW_BORDER + | UI_FLAG_CLIP + | UI_FLAG_BLOCK_MOUSE; + ui_box* panel = ui_box_begin(name, panelFlags); + + ui_box* innerView = ui_box_begin(name, 0); +} + +void ui_panel_end() +{ + ui_box* innerView = ui_box_top(); + ui_box_end(); + + ui_box* panel = ui_box_top(); + + f32 contentsW = ClampLowBound(innerView->childrenSum[0], innerView->rect.w); + f32 contentsH = ClampLowBound(innerView->childrenSum[1], innerView->rect.h); + + contentsW = ClampLowBound(contentsW, 1); + contentsH = ClampLowBound(contentsH, 1); + + if(contentsW > innerView->rect.w) + { + f32 thumbRatioX = innerView->rect.w / contentsW; + f32 sliderX = innerView->scroll.x /(contentsW - innerView->rect.w); + + ui_box* scrollBarX = ui_scrollbar("scrollerX", thumbRatioX, &sliderX); + ui_box_set_size(scrollBarX, UI_AXIS_X, UI_SIZE_PARENT_RATIO, 1., 0); + ui_box_set_size(scrollBarX, UI_AXIS_Y, UI_SIZE_PIXELS, 10, 0); + ui_box_set_floating(scrollBarX, UI_AXIS_X, 0); + ui_box_set_floating(scrollBarX, UI_AXIS_Y, panel->rect.h - 12); + + innerView->scroll.x = sliderX * (contentsW - innerView->rect.w); + } + + if(contentsH > innerView->rect.h) + { + f32 thumbRatioY = innerView->rect.h / contentsH; + f32 sliderY = innerView->scroll.y /(contentsH - innerView->rect.h); + + ui_box* scrollBarY = ui_scrollbar("scrollerY", thumbRatioY, &sliderY); + ui_box_set_size(scrollBarY, UI_AXIS_X, UI_SIZE_PIXELS, 10, 0); + ui_box_set_size(scrollBarY, UI_AXIS_Y, UI_SIZE_PARENT_RATIO, 1., 0); + ui_box_set_floating(scrollBarY, UI_AXIS_X, panel->rect.w - 12); + ui_box_set_floating(scrollBarY, UI_AXIS_Y, 0); + + innerView->scroll.y = sliderY * (contentsH - innerView->rect.h); + } + + ui_box_end(); +} + +ui_sig ui_tooltip_begin(const char* name) +{ + ui_context* ui = ui_get_context(); + ui_box_push(ui->overlay); + + vec2 p = ui_mouse_position(); + + ui_flags flags = UI_FLAG_DRAW_BACKGROUND + | UI_FLAG_DRAW_BORDER; + + ui_box* tooltip = ui_box_make(name, flags); + ui_box_set_size(tooltip, UI_AXIS_X, UI_SIZE_CHILDREN, 0, 0); + ui_box_set_size(tooltip, UI_AXIS_Y, UI_SIZE_CHILDREN, 0, 0); + ui_box_set_floating(tooltip, UI_AXIS_X, p.x); + ui_box_set_floating(tooltip, UI_AXIS_Y, p.y); + ui_box_push(tooltip); + + return(ui_box_sig(tooltip)); +} + +void ui_tooltip_end() +{ + ui_box_pop(); // tooltip + ui_box_pop(); // ui->overlay +} + +void ui_box_activate(ui_box* box) +{ + box->active = true; +} + +void ui_box_deactivate(ui_box* box) +{ + box->active = false; +} + +bool ui_box_active(ui_box* box) +{ + return(box->active); +} + +void ui_menu_bar_begin(const char* name) +{ + ui_box* bar = ui_box_begin(name, UI_FLAG_DRAW_BACKGROUND); + ui_box_set_size(bar, UI_AXIS_X, UI_SIZE_PARENT_RATIO, 1., 0); + ui_box_set_size(bar, UI_AXIS_Y, UI_SIZE_CHILDREN, 0, 0); + ui_box_set_layout(bar, UI_AXIS_X, UI_ALIGN_START, UI_ALIGN_START); + + ui_size_push(UI_AXIS_X, UI_SIZE_TEXT, 0, 0); + ui_size_push(UI_AXIS_Y, UI_SIZE_TEXT, 0, 0); + + + ui_sig sig = ui_box_sig(bar); + if(!sig.hovering && mp_input_mouse_released(MP_MOUSE_LEFT)) + { + ui_box_deactivate(bar); + } +} + +void ui_menu_bar_end() +{ + ui_size_pop(UI_AXIS_X); + ui_size_pop(UI_AXIS_Y); + ui_box_end(); // menu bar +} + +void ui_menu_begin(const char* label) +{ + ui_box* button = ui_box_make(label, UI_FLAG_CLICKABLE | UI_FLAG_DRAW_TEXT); + ui_box* bar = button->parent; + + ui_sig sig = ui_box_sig(button); + ui_sig barSig = ui_box_sig(bar); + + ui_context* ui = ui_get_context(); + ui_box_push(ui->overlay); + + ui_flags flags = UI_FLAG_DRAW_BACKGROUND + | UI_FLAG_DRAW_BORDER; + + ui_box* menu = ui_box_make(label, flags); + ui_box_set_size(menu, UI_AXIS_X, UI_SIZE_CHILDREN, 0, 0); + ui_box_set_size(menu, UI_AXIS_Y, UI_SIZE_CHILDREN, 0, 0); + ui_box_set_floating(menu, UI_AXIS_X, button->rect.x); + ui_box_set_floating(menu, UI_AXIS_Y, button->rect.y + button->rect.h); + ui_box_set_layout(menu, UI_AXIS_Y, UI_ALIGN_START, UI_ALIGN_START); + + if(ui_box_active(bar)) + { + if(sig.hovering) + { + ui_box_activate(button); + } + else if(barSig.hovering) + { + ui_box_deactivate(button); + } + } + else + { + ui_box_deactivate(button); + if(sig.pressed) + { + ui_box_activate(bar); + ui_box_activate(button); + } + } + + ui_box_set_closed(menu, !ui_box_active(button)); + ui_box_push(menu); +} + +void ui_menu_end() +{ + ui_box_pop(); // menu + ui_box_pop(); // overlay; +} + + + + +#undef LOG_SUBSYSTEM diff --git a/src/ui.h b/src/ui.h new file mode 100644 index 0000000..a0290c6 --- /dev/null +++ b/src/ui.h @@ -0,0 +1,212 @@ +/************************************************************//** +* +* @file: ui.h +* @author: Martin Fouilleul +* @date: 08/08/2022 +* @revision: +* +*****************************************************************/ +#ifndef __UI_H_ +#define __UI_H_ + +#include"typedefs.h" +#include"lists.h" +#include"graphics.h" + +#ifdef __cplusplus +extern "C" { +#endif + +typedef enum +{ + UI_FLAG_CLICKABLE = (1<<0), + UI_FLAG_SCROLLABLE = (1<<1), + UI_FLAG_BLOCK_MOUSE = (1<<2), + UI_FLAG_HOT_ANIMATION = (1<<3), + UI_FLAG_ACTIVE_ANIMATION = (1<<4), + UI_FLAG_CLIP = (1<<5), + UI_FLAG_DRAW_BACKGROUND = (1<<6), + UI_FLAG_DRAW_FOREGROUND = (1<<7), + UI_FLAG_DRAW_BORDER = (1<<8), + UI_FLAG_DRAW_TEXT = (1<<9), + +} ui_flags; + +typedef struct ui_key +{ + u64 hash; +} ui_key; + +typedef enum +{ + UI_AXIS_X, + UI_AXIS_Y, + UI_AXIS_COUNT +} ui_axis; + +typedef enum +{ + UI_ALIGN_START, + UI_ALIGN_END, + UI_ALIGN_CENTER, +} ui_align; + +typedef struct ui_layout +{ + ui_axis axis; + ui_align align[UI_AXIS_COUNT]; + +} ui_layout; + +typedef enum +{ + UI_SIZE_TEXT, + UI_SIZE_PIXELS, + UI_SIZE_CHILDREN, + UI_SIZE_PARENT_RATIO, + +} ui_size_kind; + +typedef struct ui_size +{ + ui_size_kind kind; + f32 value; + f32 strictness; +} ui_size; + +typedef enum { UI_STYLE_NORMAL, + UI_STYLE_HOT, + UI_STYLE_ACTIVE, + UI_STYLE_SELECTOR_COUNT } ui_style_selector; + +typedef struct ui_style +{ + mg_color backgroundColor; + mg_color foregroundColor; + mg_color borderColor; + mg_color textColor; + mg_font font; + f32 fontSize; + f32 borderSize; + f32 roundness; + f32 animationTime; +} ui_style; + +typedef struct ui_box ui_box; + +typedef struct ui_sig +{ + ui_box* box; + + vec2 mouse; + vec2 delta; + vec2 wheel; + + bool pressed; + bool released; + bool triggered; + bool clicked; + bool doubleClicked; + bool rightClicked; + bool dragging; + bool hovering; + +} ui_sig; + +struct ui_box +{ + // hierarchy + list_elt listElt; + list_info children; + ui_box* parent; + + // keying and caching + list_elt bucketElt; + ui_key key; + u64 frameCounter; + + // builder-provided info + ui_flags flags; + str8 string; + + // layout + u32 z; + bool floating[UI_AXIS_COUNT]; + ui_size desiredSize[UI_AXIS_COUNT]; + ui_layout layout; + mp_rect rect; + f32 childrenSum[2]; + + // styling + ui_style* styles[UI_STYLE_SELECTOR_COUNT]; + ui_style computedStyle; + ui_style_selector styleSelector; + + // signals + ui_sig* sig; + + // stateful behaviour + bool closed; + bool parentClosed; + bool dragging; + bool active; + vec2 scroll; + + // animation data + f32 hotTransition; + f32 activeTransition; +}; + +void ui_init(); + +void ui_begin_frame(u32 width, u32 height, ui_style defaultStyle); +void ui_end_frame(); +void ui_draw(mg_canvas canvas); + +ui_box* ui_box_make(const char* string, ui_flags flags); +ui_box* ui_box_begin(const char* string, ui_flags flags); +ui_box* ui_box_make_str8(str8 string, ui_flags flags); +ui_box* ui_box_begin_str8(str8 string, ui_flags flags); +ui_box* ui_box_end(); +#define ui_container(name, flags) defer_loop(ui_box_begin(name, flags), ui_box_end()) + +void ui_box_set_layout(ui_box* box, ui_axis axis, ui_align alignX, ui_align alignY); +void ui_box_set_size(ui_box* box, ui_axis axis, ui_size_kind kind, f32 value, f32 strictness); +void ui_box_set_floating(ui_box* box, ui_axis axis, f32 pos); + +void ui_box_set_style_selector(ui_box* box, ui_style_selector selector); + +ui_sig ui_box_sig(ui_box* box); + +void ui_size_push(ui_axis axis, ui_size_kind kind, f32 value, f32 strictness); +void ui_size_pop(ui_axis axis); +void ui_style_push(ui_style_selector selector, ui_style style); +void ui_style_pop(ui_style_selector selector); + +// Basic helpers +ui_sig ui_label(const char* label); + +ui_sig ui_button(const char* label); +ui_box* ui_scrollbar(const char* label, f32 thumbRatio, f32* scrollValue); + +void ui_panel_begin(const char* name); +void ui_panel_end(); +#define ui_panel(name) defer_loop(ui_panel_begin(name), ui_panel_end()) + +ui_sig ui_tooltip_begin(const char* name); +void ui_tooltip_end(); +#define ui_tooltip(name) defer_loop(ui_tooltip_begin(name), ui_tooltip_end()) + +void ui_menu_bar_begin(const char* label); +void ui_menu_bar_end(); +#define ui_menu_bar(name) defer_loop(ui_menu_bar_begin(name), ui_menu_bar_end()) + +void ui_menu_begin(const char* label); +void ui_menu_end(); +#define ui_menu(name) defer_loop(ui_menu_begin(name), ui_menu_end()) + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__UI_H_ diff --git a/src/util/debug_log.c b/src/util/debug_log.c new file mode 100644 index 0000000..ce2caaf --- /dev/null +++ b/src/util/debug_log.c @@ -0,0 +1,123 @@ +/************************************************************//** +* +* @file: debug_log.c +* @author: Martin Fouilleul +* @date: 22/10/2020 +* @revision: +* +*****************************************************************/ +#include +#include +#include"debug_log.h" + +static const char* LOG_HEADINGS[LOG_LEVEL_COUNT] = { + "Error", + "Warning", + "Message", + "Debug"}; + +static const char* LOG_FORMATS[LOG_LEVEL_COUNT] = { + "\e[38;5;9m\e[1m", + "\e[38;5;13m\e[1m", + "\e[38;5;10m\e[1m", + "\e[38;5;14m\e[1m" }; + +static const char* LOG_FORMAT_STOP = "\e[m"; + +const int LOG_SUBSYSTEM_MAX_COUNT = 16; + +typedef struct log_config +{ + FILE* out; + log_level level; + const char* subsystemNames[LOG_SUBSYSTEM_MAX_COUNT]; + log_level subsystemLevels[LOG_SUBSYSTEM_MAX_COUNT]; + +} log_config; + +static log_config __log_config = {.out = 0, + .level = LOG_DEFAULT_LEVEL, + .subsystemNames = {}, + .subsystemLevels = {}}; + +int LogFindSubsystem(const char* subsystem) +{ + for(int i=0; i= 0)? __log_config.subsystemLevels[subsystemIndex] : __log_config.level; + + if(level <= filterLevel) + { + if(!__log_config.out) + { + __log_config.out = LOG_DEFAULT_OUTPUT; + } + fprintf(__log_config.out, + "%s%s:%s [%s] %s() in %s:%i: ", + LOG_FORMATS[level], + LOG_HEADINGS[level], + LOG_FORMAT_STOP, + subsystem, + functionName, + fileName, + line); + + va_list ap; + va_start(ap, msg); + vfprintf(__log_config.out, msg, ap); + va_end(ap); + } +} + +void LogOutput(FILE* output) +{ + __log_config.out = output; +} + +void LogLevel(log_level level) +{ + __log_config.level = level; +} + +void LogFilter(const char* subsystem, log_level level) +{ + int firstNull = -1; + for(int i=0; i +#include"typedefs.h" +#include"macro_helpers.h" + +#ifdef __cplusplus +extern "C" { +#endif +//NOTE(martin): the default logging level can be adjusted by defining LOG_DEFAULT_LEVEL. As the name suggest, it is the default, but it +// can be adjusted at runtime with LogLevel() +#ifndef LOG_DEFAULT_LEVEL + #define LOG_DEFAULT_LEVEL LOG_LEVEL_WARNING +#endif + +//NOTE(martin): the default output can be adjusted by defining LOG_DEFAULT_OUTPUT. It can be adjusted at runtime with LogOutput() +#ifndef LOG_DEFAULT_OUTPUT + #define LOG_DEFAULT_OUTPUT stdout +#endif + +//NOTE(martin): LOG_SUBSYSTEM can be defined in each compilation unit to associate it with a subsystem, like this: +// #define LOG_SUBSYSTEM "name" + +typedef enum { LOG_LEVEL_ERROR, + LOG_LEVEL_WARNING, + LOG_LEVEL_MESSAGE, + LOG_LEVEL_DEBUG, + LOG_LEVEL_COUNT } log_level; + +void LogGeneric(log_level level, + const char* subsystem, + const char* functionName, + const char* fileName, + u32 line, + const char* msg, + ...); + +void LogOutput(FILE* output); +void LogLevel(log_level level); +void LogFilter(const char* subsystem, log_level level); + +#define LOG_GENERIC(level, func, file, line, msg, ...) LogGeneric(level, LOG_SUBSYSTEM, func, file, line, msg, ##__VA_ARGS__ ) + +#define LOG_ERROR(msg, ...) LOG_GENERIC(LOG_LEVEL_ERROR, __FUNCTION__, __FILE__, __LINE__, msg, ##__VA_ARGS__ ) + +//NOTE(martin): warnings, messages, and debug info can be enabled in debug mode by defining LOG_COMPILE_XXX, XXX being the max desired log level +// error logging is always compiled +#if defined(LOG_COMPILE_WARNING) || defined(LOG_COMPILE_MESSAGE) || defined(LOG_COMPILE_DEBUG) + #define LOG_WARNING(msg, ...) LOG_GENERIC(LOG_LEVEL_WARNING, __FUNCTION__, __FILE__, __LINE__, msg, ##__VA_ARGS__ ) + + #if defined(LOG_COMPILE_MESSAGE) || defined(LOG_COMPILE_DEBUG) + #define LOG_MESSAGE(msg, ...) LOG_GENERIC(LOG_LEVEL_MESSAGE, __FUNCTION__, __FILE__, __LINE__, msg, ##__VA_ARGS__ ) + + #if defined(LOG_COMPILE_DEBUG) + #define LOG_DEBUG(msg, ...) LOG_GENERIC(LOG_LEVEL_DEBUG, __FUNCTION__, __FILE__, __LINE__, msg, ##__VA_ARGS__ ) + #else + #define LOG_DEBUG(msg, ...) + #endif + #else + #define LOG_MESSAGE(msg, ...) + #define LOG_DEBUG(msg, ...) + #endif +#else + #define LOG_WARNING(msg, ...) + #define LOG_MESSAGE(msg, ...) + #define LOG_DEBUG(msg, ...) +#endif + +#ifndef NO_ASSERT + #include + #define _ASSERT_(x, msg) assert(x && msg) + #define ASSERT(x, ...) _ASSERT_(x, #__VA_ARGS__) + + #ifdef DEBUG + #define DEBUG_ASSERT(x, ...) ASSERT(x, ##__VA_ARGS__ ) + #else + #define DEBUG_ASSERT(x, ...) + #endif +#else + #define ASSERT(x, ...) + #define DEBUG_ASSERT(x, ...) +#endif + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__DEBUG_LOG_H_ diff --git a/src/util/hash.c b/src/util/hash.c new file mode 100644 index 0000000..7e704d8 --- /dev/null +++ b/src/util/hash.c @@ -0,0 +1,124 @@ +/************************************************************//** +* +* @file: hash.cpp +* @author: Martin Fouilleul +* @date: 08/08/2022 +* @revision: +* +*****************************************************************/ +#include +#include"hash.h" + +u64 mp_hash_aes_u64(u64 x) +{ + u8 seed[16] = { + 0xaa, 0x9b, 0xbd, 0xb8, + 0xa1, 0x98, 0xac, 0x3f, + 0x1f, 0x94, 0x07, 0xb3, + 0x8c, 0x27, 0x93, 0x69 }; + + __m128i hash = _mm_set_epi64x(0L, x); + __m128i key = _mm_loadu_si128((__m128i*)seed); + hash = _mm_aesdec_si128(hash, key); + hash = _mm_aesdec_si128(hash, key); + u64 result = _mm_extract_epi64(hash, 0); + + return(result); +} + +u64 mp_hash_aes_u64_x2(u64 x, u64 y) +{ + u8 seed[16] = { + 0xaa, 0x9b, 0xbd, 0xb8, + 0xa1, 0x98, 0xac, 0x3f, + 0x1f, 0x94, 0x07, 0xb3, + 0x8c, 0x27, 0x93, 0x69 }; + + __m128i hash = _mm_set_epi64x(x, y); + __m128i key = _mm_loadu_si128((__m128i*)seed); + hash = _mm_aesdec_si128(hash, key); + hash = _mm_aesdec_si128(hash, key); + u64 result = _mm_extract_epi64(hash, 0); + + return(result); +} + +u64 mp_hash_aes_string(str8 string) +{ + u8 seed[16] = { + 0xaa, 0x9b, 0xbd, 0xb8, + 0xa1, 0x98, 0xac, 0x3f, + 0x1f, 0x94, 0x07, 0xb3, + 0x8c, 0x27, 0x93, 0x69 }; + + __m128i hash = _mm_loadu_si128((__m128i*)seed); + + u64 chunkCount = string.len / 16; + char* at = string.ptr; + + while(chunkCount--) + { + __m128i in = _mm_loadu_si128((__m128i*)at); + at += 16; + + hash = _mm_xor_si128(hash, in); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + } + + u64 restCount = string.len % 16; + char tmp[16]; + memset(tmp, 0, 16); + memmove(tmp, at, restCount); + + __m128i in = _mm_loadu_si128((__m128i*)tmp); + hash = _mm_xor_si128(hash, in); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + + u64 result = _mm_extract_epi64(hash, 0); + return(result); +} + +u64 mp_hash_aes_string_seed(str8 string, u64 seed) +{ + u8 seed16[16]; + memcpy(seed16, &seed, 8); + memcpy(seed16+8, &seed, 8); + + __m128i hash = _mm_loadu_si64(&seed16); + + u64 chunkCount = string.len / 16; + char* at = string.ptr; + + while(chunkCount--) + { + __m128i in = _mm_loadu_si128((__m128i*)at); + at += 16; + + hash = _mm_xor_si128(hash, in); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + } + + u64 restCount = string.len % 16; + char tmp[16]; + memset(tmp, 0, 16); + memmove(tmp, at, restCount); + + __m128i in = _mm_loadu_si128((__m128i*)tmp); + hash = _mm_xor_si128(hash, in); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + hash = _mm_aesdec_si128(hash, _mm_setzero_si128()); + + u64 result = _mm_extract_epi64(hash, 0); + return(result); +} diff --git a/src/util/hash.h b/src/util/hash.h new file mode 100644 index 0000000..b482a4c --- /dev/null +++ b/src/util/hash.h @@ -0,0 +1,29 @@ +/************************************************************//** +* +* @file: hash.h +* @author: Martin Fouilleul +* @date: 08/08/2022 +* @revision: +* +*****************************************************************/ +#ifndef __HASH_H_ +#define __HASH_H_ + +#include"typedefs.h" +#include"strings.h" + +#ifdef __cplusplus +extern "C" { +#endif + +u64 mp_hash_aes_u64(u64 x); +u64 mp_hash_aes_u64_x2(u64 x, u64 y); +u64 mp_hash_aes_string(str8 string); +u64 mp_hash_aes_string_seed(str8 string, u64 seed); + +#ifdef __cplusplus +} // extern "C" +#endif + + +#endif //__HASH_H_ diff --git a/src/util/lists.h b/src/util/lists.h new file mode 100644 index 0000000..605089f --- /dev/null +++ b/src/util/lists.h @@ -0,0 +1,390 @@ +/************************************************************//** +* +* @file: lists.h +* @author: Martin Fouilleul +* @date: 22/11/2017 +* @revision: 28/04/2019 : deleted containers which are not used by BLITz +* @brief: Implements generic intrusive linked list and dynamic array +* +****************************************************************/ +#ifndef __CONTAINERS_H_ +#define __CONTAINERS_H_ + +#include"debug_log.h" + +#ifdef __cplusplus +extern "C" { +#endif + +#define OFFSET_OF_CONTAINER(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER) + +#ifdef __cplusplus +#define CONTAINER_OF(ptr, type, member) ({ \ + const decltype( ((type *)0)->member ) *__mptr = (ptr); \ + (type *)( (char *)__mptr - OFFSET_OF_CONTAINER(type,member) );}) +#else +#define CONTAINER_OF(ptr, type, member) ({ \ + const char *__mptr = (char*)(ptr); \ + (type *)(__mptr - OFFSET_OF_CONTAINER(type,member) );}) +#endif + +//------------------------------------------------------------------------- +// Intrusive linked lists +//------------------------------------------------------------------------- + +#define ListEntry(ptr, type, member) \ + CONTAINER_OF(ptr, type, member) + +#define ListNext(elt) (elt)->next +#define ListPrev(elt) (elt)->prev + +#define ListNextEntry(list, elt, type, member) \ + ((elt->member.next != ListEnd(list)) ? ListEntry(elt->member.next, type, member) : 0) + +#define ListPrevEntry(list, elt, type, member) \ + ((elt->member.prev != ListEnd(list)) ? ListEntry(elt->member.prev, type, member) : 0) + +#define ListCheckedEntry(list, type, member) \ + (((list) != 0) ? ListEntry(list, type, member) : 0) + +#define ListFirstEntry(list, type, member) \ + (ListCheckedEntry(ListBegin(list), type, member)) + +#define ListLastEntry(list, type, member) \ + (ListCheckedEntry(ListLast(list), type, member)) + +#define for_each_in_list(list, elt, type, member) \ + for(type* elt = ListCheckedEntry(ListBegin(list), type, member); \ + elt != 0; \ + elt = ListCheckedEntry(elt->member.next, type, member)) \ + +#define for_each_in_list_reverse(list, elt, type, member) \ + for(type* elt = ListCheckedEntry(ListLast(list), type, member); \ + elt != 0; \ + elt = ListCheckedEntry(elt->member.prev, type, member)) \ + +#define for_each_in_list_safe(list, elt, type, member) \ + for(type* elt = ListCheckedEntry(ListBegin(list), type, member), \ + *__tmp = elt ? ListCheckedEntry(elt->member.next, type, member) : 0 ; \ + elt != 0; \ + elt = __tmp, \ + __tmp = elt ? ListCheckedEntry(elt->member.next, type, member) : 0) \ + +#define ListPopEntry(list, type, member) (ListEmpty(list) ? 0 : ListEntry(ListPop(list), type, member)) + +typedef struct list_elt list_elt; +struct list_elt +{ + list_elt* next; + list_elt* prev; +}; + +typedef struct list_info +{ + list_elt* first; + list_elt* last; +} list_info; + +static inline void ListInit(list_info* list) +{ + list->first = list->last = 0; +} + +static inline list_elt* ListBegin(list_info* list) +{ + return(list->first); +} +static inline list_elt* ListEnd(list_info* list) +{ + return(0); +} + +static inline list_elt* ListLast(list_info* list) +{ + return(list->last); +} + +static inline void ListInsert(list_info* list, list_elt* afterElt, list_elt* elt) +{ + elt->prev = afterElt; + elt->next = afterElt->next; + if(afterElt->next) + { + afterElt->next->prev = elt; + } + else + { + list->last = elt; + } + afterElt->next = elt; + + ASSERT(elt->next != elt, "ListInsert(): can't insert an element into itself"); +} + +static inline void ListInsertBefore(list_info* list, list_elt* beforeElt, list_elt* elt) +{ + elt->next = beforeElt; + elt->prev = beforeElt->prev; + + if(beforeElt->prev) + { + beforeElt->prev->next = elt; + } + else + { + list->first = elt; + } + beforeElt->prev = elt; + + ASSERT(elt->next != elt, "ListInsertBefore(): can't insert an element into itself"); +} + +static inline void ListRemove(list_info* list, list_elt* elt) +{ + if(elt->prev) + { + elt->prev->next = elt->next; + } + else + { + DEBUG_ASSERT(list->first == elt); + list->first = elt->next; + } + if(elt->next) + { + elt->next->prev = elt->prev; + } + else + { + DEBUG_ASSERT(list->last == elt); + list->last = elt->prev; + } + elt->prev = elt->next = 0; +} + +static inline void ListPush(list_info* list, list_elt* elt) +{ + elt->next = list->first; + elt->prev = 0; + if(list->first) + { + list->first->prev = elt; + } + else + { + list->last = elt; + } + list->first = elt; +} + +static inline list_elt* ListPop(list_info* list) +{ + list_elt* elt = ListBegin(list); + if(elt != ListEnd(list)) + { + ListRemove(list, elt); + return(elt); + } + else + { + return(0); + } +} + +static inline void ListPushBack(list_info* list, list_elt* elt) +{ + elt->prev = list->last; + elt->next = 0; + if(list->last) + { + list->last->next = elt; + } + else + { + list->first = elt; + } + list->last = elt; +} +#define ListAppend(a, b) ListPushBack(a, b) + + +static inline list_elt* ListPopBack(list_info* list) +{ + list_elt* elt = ListLast(list); + if(elt != ListEnd(list)) + { + ListRemove(list, elt); + return(elt); + } + else + { + return(0); + } +} + +static inline bool ListEmpty(list_info* list) +{ + return(list->first == 0 || list->last == 0); +} + + +//------------------------------------------------------------------------- +// Circular Intrusive linked lists +//------------------------------------------------------------------------- + +#define CListEntry(ptr, type, member) ListEntry(ptr, type, member) +#define CListNext(elt) ListNext(elt) +#define CListPrev(elt) ListPrev(elt) + +#define CListNextEntry(head, elt, type, member) \ + ((elt->member.next != CListEnd(head)) ? CListEntry(elt->member.next, type, member) : 0) + +#define CListPrevEntry(head, elt, type, member) \ + ((elt->member.prev != CListEnd(head)) ? CListEntry(elt->member.prev, type, member) : 0) + +#define CListCheckedEntry(head, info, type, member) \ + ((info != CListEnd(head)) ? CListEntry(info, type, member) : 0) + +#define CListFirstEntry(head, type, member) \ + (CListCheckedEntry(head, CListBegin(head), type, member)) + +#define CListLastEntry(head, type, member) \ + (CListCheckedEntry(head, CListLast(head), type, member)) + +#define for_each_in_clist(list, elt, type, member) \ + for(type* elt = CListEntry(CListBegin(list), type, member); \ + &elt->member != CListEnd(list); \ + elt = CListEntry(elt->member.next, type, member)) \ + + +#define for_each_in_clist_reverse(list, elt, type, member) \ + for(type* elt = CListEntry(CListLast(list), type, member); \ + &elt->member != CListEnd(list); \ + elt = CListEntry(elt->member.prev, type, member)) \ + + +#define for_each_in_clist_safe(list, elt, type, member) \ + for(type* elt = CListEntry(CListBegin(list), type, member), \ + *__tmp = CListEntry(elt->member.next, type, member); \ + &elt->member != CListEnd(list); \ + elt = CListEntry(&__tmp->member, type, member), \ + __tmp = CListEntry(elt->member.next, type, member)) \ + + +#define CListPush(a, b) CListInsert(a, b) +#define CListInsertBefore(a, b) CListAppend(a, b) + +#define CListPopEntry(list, type, member) (CListEmpty(list) ? 0 : CListEntry(CListPop(list), type, member)) + +static inline void CListInit(list_elt* info) +{ + info->next = info->prev = info; +} + +static inline list_elt* CListBegin(list_elt* head) +{ + return(head->next ? head->next : head ); +} +static inline list_elt* CListEnd(list_elt* head) +{ + return(head); +} + +static inline list_elt* CListLast(list_elt* head) +{ + return(head->prev ? head->prev : head); +} + +static inline void CListInsert(list_elt* head, list_elt* elt) +{ + elt->prev = head; + elt->next = head->next; + if(head->next) + { + head->next->prev = elt; + } + else + { + head->prev = elt; + } + head->next = elt; + + ASSERT(elt->next != elt, "CListInsert(): can't insert an element into itself"); +} + +static inline void CListAppend(list_elt* head, list_elt* elt) +{ + CListInsert(head->prev, elt); +} + +static inline void CListCat(list_elt* head, list_elt* list) +{ + if(head->prev) + { + head->prev->next = list->next; + } + if(head->prev && head->prev->next) + { + head->prev->next->prev = head->prev; + } + head->prev = list->prev; + if(head->prev) + { + head->prev->next = head; + } + CListInit(list); +} + +static inline void CListRemove(list_elt* elt) +{ + if(elt->prev) + { + elt->prev->next = elt->next; + } + if(elt->next) + { + elt->next->prev = elt->prev; + } + elt->prev = elt->next = 0; +} + +static inline list_elt* CListPop(list_elt* head) +{ + list_elt* it = CListBegin(head); + if(it != CListEnd(head)) + { + CListRemove(it); + return(it); + } + else + { + return(0); + } + +} + +static inline list_elt* CListPopBack(list_elt* head) +{ + list_elt* it = CListLast(head); + if(it != CListEnd(head)) + { + CListRemove(it); + return(it); + } + else + { + return(0); + } +} + +static inline bool CListEmpty(list_elt* head) +{ + return(head->next == 0 || head->next == head); +} + + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__CONTAINERS_H_ diff --git a/src/util/macro_helpers.h b/src/util/macro_helpers.h new file mode 100644 index 0000000..a38c0b3 --- /dev/null +++ b/src/util/macro_helpers.h @@ -0,0 +1,129 @@ +/************************************************************//** +* +* @file: macro_helpers.h +* @author: Martin Fouilleul +* @date: 27/03/2020 +* @revision: +* +*****************************************************************/ +#ifndef __MACRO_HELPERS_H_ +#define __MACRO_HELPERS_H_ + +//NOTE(martin): macro concatenation +#define _cat2_(a, b) a##b +#define _cat3_(a, b, c) a##b##c + + +//NOTE(martin): inline, but still generate code +// (eg. use the inline version inside a library, but still exports the function for client code) +//TODO(martin): this is a compiler-specific attribute, recognized by clang and gcc. See if there's a more portable approach +//#define INLINE_GEN __attribute__((used)) static inline + + +//NOTE(martin): typed and array mallocs +#define malloc_type(type) ((type*)malloc(sizeof(type))) +#define malloc_array(type, count) ((type*)malloc(sizeof(type)*count)) + +//NOTE(martin): 'hygienic' templates, to replace macros and avoid multiple evaluation problems. +#ifdef __cplusplus + //NOTE(martin): in C++ we use templates and decltype/declval + // (overloaded functions would be ambiguous because of the + // overload resolution and conversion/promotion rules) + + #include + + template + inline decltype(std::declval()+std::declval()) minimum_safe(Ta a, Tb b) + { + return(a < b ? a : b); + } + + template + inline decltype(std::declval()+std::declval()) maximum_safe(Ta a, Tb b) + { + return(a > b ? a : b); + } + + template + inline T square_safe(T a) {return(a*a);} + + template + inline T cube_safe(T a) {return(a*a*a);} + +#else // (__cplusplus not defined) + + //NOTE(martin): Type generic arithmetic functions helpers + // this macros helps generate variants of a generic 'template' for all arithmetic types. + // the def parameter must be a macro that take a type, and optional arguments + #define tga_generate_variants(def, ...) \ + def(u8, ##__VA_ARGS__) def(i8, ##__VA_ARGS__ ) def(u16, ##__VA_ARGS__) def(i16, ##__VA_ARGS__) \ + def(u32, ##__VA_ARGS__) def(i32, ##__VA_ARGS__) def(u64, ##__VA_ARGS__) def(i64, ##__VA_ARGS__) \ + def(f32, ##__VA_ARGS__) def(f64, ##__VA_ARGS__) + + // This macro generates the name of a typed variant + #define tga_variant_name(name, type) _cat3_(name, _, type) + + // This macro generates a _Generic association between a type and its variant + #define tga_variant_association(type, name) , type: tga_variant_name(name, type) + + // This macros selects the appropriate variant for a 2 parameters functions + #define tga_select_binary(name, a, b) \ + _Generic((a+b) tga_generate_variants(tga_variant_association, name))(a, b) + + // This macros selects the appropriate variant for a 1 parameters functions + #define tga_select_unary(name, a) \ + _Generic((a) tga_generate_variants(tga_variant_association, name))(a) + + //NOTE(martin): type generic templates + #define minimum_def(type) static inline type tga_variant_name(minimum_safe, type)(type a, type b) {return(a < b ? a : b);} + #define maximum_def(type) static inline type tga_variant_name(maximum_safe, type)(type a, type b) {return(a > b ? a : b);} + #define square_def(type) static inline type tga_variant_name(square_safe, type)(type a) {return(a*a);} + #define cube_def(type) static inline type tga_variant_name(cube_safe, type)(type a) {return(a*a*a);} + + //NOTE(martin): instantiante our templates for all arithmetic types + tga_generate_variants(minimum_def) + tga_generate_variants(maximum_def) + tga_generate_variants(square_def) + tga_generate_variants(cube_def) + + //NOTE(martin): select the correct variant according to the argument types + #define minimum_safe(a, b) tga_select_binary(minimum_safe, a, b) + #define maximum_safe(a, b) tga_select_binary(maximum_safe, a, b) + #define square_safe(a) tga_select_unary(square_safe, a) + #define cube_safe(a) tga_select_unary(cube_safe, a) + +#endif // __cplusplus else branch + + +//NOTE(martin): these macros are calling the safe functions defined above, so they don't evaluate their +// arguments twice + +#define minimum(a, b) minimum_safe(a, b) +#define maximum(a, b) maximum_safe(a, b) + +#define ClampLowBound(a, low) (maximum((a), (low))) +#define ClampHighBound(a, high) (minimum((a), (high))) +#define Clamp(a, low, high) (ClampLowBound(ClampHighBound((a), (high)), (low))) + +#define Square(a) square_safe(a) +#define Cube(a) cube_safe(a) + +#define AlignUpOnPow2(x, a) (((x) + (a) - 1) & ~((a)-1)) +#define AlignDownOnPow2(x, a) ((x) & ~((a)-1)) + +static inline u64 next_pow2_u64(u64 x) +{ + x--; + x |= x>>1; + x |= x>>2; + x |= x>>4; + x |= x>>8; + x |= x>>16; + x |= x>>32; + x++; + return(x); +} + +#define defer_loop(begin, end) begin; for(int __i__=0; __i__<1; __i__++, end) + +#endif //__MACRO_HELPERS_H_ diff --git a/src/util/memory.c b/src/util/memory.c new file mode 100644 index 0000000..25b0157 --- /dev/null +++ b/src/util/memory.c @@ -0,0 +1,129 @@ +/************************************************************//** +* +* @file: memory.c +* @author: Martin Fouilleul +* @date: 24/10/2019 +* @revision: +* +*****************************************************************/ +#include // memset + +#include"memory.h" +#include"platform_base_allocator.h" +#include"macro_helpers.h" + +static const u32 MEM_ARENA_DEFAULT_RESERVE_SIZE = 1<<30; +static const u32 MEM_ARENA_COMMIT_ALIGNMENT = 1<<20; + +//-------------------------------------------------------------------------------- +//NOTE(martin): memory arena +//-------------------------------------------------------------------------------- +void mem_arena_init(mem_arena* arena) +{ + mem_arena_init_with_options(arena, &(mem_arena_options){}); +} + +void mem_arena_init_with_options(mem_arena* arena, mem_arena_options* options) +{ + arena->base = options->base ? options->base : mem_base_allocator_default(); + arena->cap = options->reserve ? options->reserve : MEM_ARENA_DEFAULT_RESERVE_SIZE; + + arena->ptr = mem_base_reserve(arena->base, arena->cap); + arena->committed = 0; + arena->offset = 0; +} + +void mem_arena_release(mem_arena* arena) +{ + mem_base_release(arena->base, arena->ptr, arena->cap); + memset(arena, 0, sizeof(mem_arena)); +} + +void* mem_arena_alloc(mem_arena* arena, u64 size) +{ + u64 nextOffset = arena->offset + size; + ASSERT(nextOffset <= arena->cap); + + if(nextOffset > arena->committed) + { + u64 nextCommitted = AlignUpOnPow2(nextOffset, MEM_ARENA_COMMIT_ALIGNMENT); + nextCommitted = ClampHighBound(nextCommitted, arena->cap); + u64 commitSize = nextCommitted - arena->committed; + mem_base_commit(arena->base, arena->ptr + arena->committed, commitSize); + arena->committed = nextCommitted; + } + char* p = arena->ptr + arena->offset; + arena->offset += size; + + return(p); +} + +void mem_arena_clear(mem_arena* arena) +{ + arena->offset = 0; +} + +//-------------------------------------------------------------------------------- +//NOTE(martin): memory pool +//-------------------------------------------------------------------------------- +void mem_pool_init(mem_pool* pool, u64 blockSize) +{ + mem_pool_init_with_options(pool, blockSize, &(mem_pool_options){}); +} +void mem_pool_init_with_options(mem_pool* pool, u64 blockSize, mem_pool_options* options) +{ + mem_arena_init_with_options(&pool->arena, &(mem_arena_options){.base = options->base, .reserve = options->reserve}); + pool->blockSize = ClampLowBound(blockSize, sizeof(list_info)); + ListInit(&pool->freeList); +} + +void mem_pool_release(mem_pool* pool) +{ + mem_arena_release(&pool->arena); + memset(pool, 0, sizeof(mem_pool)); +} + +void* mem_pool_alloc(mem_pool* pool) +{ + if(ListEmpty(&pool->freeList)) + { + return(mem_arena_alloc(&pool->arena, pool->blockSize)); + } + else + { + return(ListPop(&pool->freeList)); + } +} + +void mem_pool_recycle(mem_pool* pool, void* ptr) +{ + ASSERT((((char*)ptr) >= pool->arena.ptr) && (((char*)ptr) < (pool->arena.ptr + pool->arena.offset))); + ListPush(&pool->freeList, (list_elt*)ptr); +} + +void mem_pool_clear(mem_pool* pool) +{ + mem_arena_clear(&pool->arena); + ListInit(&pool->freeList); +} + + +//-------------------------------------------------------------------------------- +//NOTE(martin): per-thread scratch arena +//-------------------------------------------------------------------------------- + +__thread mem_arena __scratchArena = {}; + +mem_arena* mem_scratch() +{ + if(__scratchArena.ptr == 0) + { + mem_arena_init(&__scratchArena); + } + return(&__scratchArena); +} + +void mem_scratch_clear() +{ + mem_arena_clear(mem_scratch()); +} diff --git a/src/util/memory.h b/src/util/memory.h new file mode 100644 index 0000000..4e04e16 --- /dev/null +++ b/src/util/memory.h @@ -0,0 +1,106 @@ +/************************************************************//** +* +* @file: memory.h +* @author: Martin Fouilleul +* @date: 24/10/2019 +* @revision: +* +*****************************************************************/ +#ifndef __MEMORY_H_ +#define __MEMORY_H_ + +#include"typedefs.h" +#include"lists.h" + +#ifdef __cplusplus +extern "C" { +#endif + +//-------------------------------------------------------------------------------- +//NOTE(martin): base allocator +//-------------------------------------------------------------------------------- +typedef void*(*mem_reserve_function)(void* context, u64 size); +typedef void(*mem_modify_function)(void* context, void* ptr, u64 size); + +typedef struct mem_base_allocator +{ + mem_reserve_function reserve; + mem_modify_function commit; + mem_modify_function decommit; + mem_modify_function release; + void* context; + +} mem_base_allocator; + +mem_base_allocator* mem_base_allocator_default(); + +#define mem_base_reserve(base, size) base->reserve(base->context, size) +#define mem_base_commit(base, ptr, size) base->commit(base->context, ptr, size) +#define mem_base_decommit(base, ptr, size) base->decommit(base->context, ptr, size) +#define mem_base_release(base, ptr, size) base->release(base->context, ptr, size) + +//-------------------------------------------------------------------------------- +//NOTE(martin): memory arena +//-------------------------------------------------------------------------------- +typedef struct mem_arena +{ + mem_base_allocator* base; + char* ptr; + u64 offset; + u64 committed; + u64 cap; +} mem_arena; + +typedef struct mem_arena_options +{ + mem_base_allocator* base; + u64 reserve; +} mem_arena_options; + +void mem_arena_init(mem_arena* arena); +void mem_arena_init_with_options(mem_arena* arena, mem_arena_options* options); +void mem_arena_release(mem_arena* arena); + +void* mem_arena_alloc(mem_arena* arena, u64 size); +void mem_arena_clear(mem_arena* arena); + +#define mem_arena_alloc_type(arena, type) ((type*)mem_arena_alloc(arena, sizeof(type))) +#define mem_arena_alloc_array(arena, type, count) ((type*)mem_arena_alloc(arena, sizeof(type)*(count))) + +//-------------------------------------------------------------------------------- +//NOTE(martin): memory pool +//-------------------------------------------------------------------------------- +typedef struct mem_pool +{ + mem_arena arena; + list_info freeList; + u64 blockSize; +} mem_pool; + +typedef struct mem_pool_options +{ + mem_base_allocator* base; + u64 reserve; +} mem_pool_options; + +void mem_pool_init(mem_pool* pool, u64 blockSize); +void mem_pool_init_with_options(mem_pool* pool, u64 blockSize, mem_pool_options* options); +void mem_pool_release(mem_pool* pool); + +void* mem_pool_alloc(mem_pool* pool); +void mem_pool_recycle(mem_pool* pool, void* ptr); +void mem_pool_clear(mem_pool* pool); + +#define mem_pool_alloc_type(arena, type) ((type*)mem_pool_alloc(arena)) + +//-------------------------------------------------------------------------------- +//NOTE(martin): per-thread implicit scratch arena +//-------------------------------------------------------------------------------- +void mem_scratch_clear(); +mem_arena* mem_scratch(); + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__MEMORY_H_ diff --git a/src/util/ringbuffer.c b/src/util/ringbuffer.c new file mode 100644 index 0000000..6252063 --- /dev/null +++ b/src/util/ringbuffer.c @@ -0,0 +1,94 @@ +/************************************************************//** +* +* @file: ringbuffer.cpp +* @author: Martin Fouilleul +* @date: 31/07/2022 +* @revision: +* +*****************************************************************/ +#include // malloc, free +#include"ringbuffer.h" + +void ringbuffer_init(ringbuffer* ring, u8 capExp) +{ + u32 cap = 1<mask = cap - 1; + ring->readIndex = 0; + ring->writeIndex = 0; + ring->buffer = (u8*)malloc(cap); +} + +void ringbuffer_cleanup(ringbuffer* ring) +{ + free(ring->buffer); +} + +u32 ringbuffer_read_available(ringbuffer* ring) +{ + return((ring->writeIndex - ring->readIndex) & ring->mask); +} + +u32 ringbuffer_write_available(ringbuffer* ring) +{ + //NOTE(martin): we keep one sentinel byte between write index and read index, + // when the buffer is full, to avoid overrunning read index. + // Hence, available write space is size - 1 - available read space. + return(ring->mask - ringbuffer_read_available(ring)); +} + +u32 ringbuffer_write(ringbuffer* ring, u32 size, u8* data) +{ + u32 read = ring->readIndex; + u32 write = ring->writeIndex; + + u32 writeAvailable = ringbuffer_write_available(ring); + if(size > writeAvailable) + { + DEBUG_ASSERT("not enough space available"); + size = writeAvailable; + } + + if(read <= write) + { + u32 copyCount = minimum(size, ring->mask + 1 - write); + memcpy(ring->buffer + write, data, copyCount); + + data += copyCount; + copyCount = size - copyCount; + memcpy(ring->buffer, data, copyCount); + } + else + { + memcpy(ring->buffer + write, data, size); + } + ring->writeIndex = (write + size) & ring->mask; + return(size); +} + +u32 ringbuffer_read(ringbuffer* ring, u32 size, u8* data) +{ + u32 read = ring->readIndex; + u32 write = ring->writeIndex; + + u32 readAvailable = ringbuffer_read_available(ring); + if(size > readAvailable) + { + size = readAvailable; + } + + if(read <= write) + { + memcpy(data, ring->buffer + read, size); + } + else + { + u32 copyCount = minimum(size, ring->mask + 1 - read); + memcpy(data, ring->buffer + read, copyCount); + + data += copyCount; + copyCount = size - copyCount; + memcpy(data, ring->buffer, copyCount); + } + ring->readIndex = (read + size) & ring->mask; + return(size); +} diff --git a/src/util/ringbuffer.h b/src/util/ringbuffer.h new file mode 100644 index 0000000..a35740f --- /dev/null +++ b/src/util/ringbuffer.h @@ -0,0 +1,40 @@ +/************************************************************//** +* +* @file: ringbuffer.h +* @author: Martin Fouilleul +* @date: 31/07/2022 +* @revision: +* +*****************************************************************/ +#ifndef __RINGBUFFER_H_ +#define __RINGBUFFER_H_ + +#include +#include"typedefs.h" + +#ifdef __cplusplus +extern "C" { +#endif + +typedef struct ringbuffer +{ + u32 mask; + _Atomic(u32) readIndex; + _Atomic(u32) writeIndex; + + u8* buffer; + +} ringbuffer; + +void ringbuffer_init(ringbuffer* ring, u8 capExp); +void ringbuffer_cleanup(ringbuffer* ring); +u32 ringbuffer_read_available(ringbuffer* ring); +u32 ringbuffer_write_available(ringbuffer* ring); +u32 ringbuffer_write(ringbuffer* ring, u32 size, u8* data); +u32 ringbuffer_read(ringbuffer* ring, u32 size, u8* data); + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__RINGBUFFER_H_ diff --git a/src/util/strings.c b/src/util/strings.c new file mode 100644 index 0000000..fa8261d --- /dev/null +++ b/src/util/strings.c @@ -0,0 +1,298 @@ +/************************************************************//** +* +* @file: strings.c +* @author: Martin Fouilleul +* @date: 29/05/2021 +* @revision: +* +*****************************************************************/ +#include // strlen(), strcpy(), etc. +#include // va_list() etc. +#include"debug_log.h" +#include"strings.h" + +#define LOG_SUBSYSTEM "Strings" + +//---------------------------------------------------------------------------------- +// string slices as values +//---------------------------------------------------------------------------------- + +str8 str8_from_buffer(u64 len, char* buffer) +{ + return((str8){.len = len, .ptr = buffer}); +} + +str8 str8_from_cstring(char* str) +{ + return(str8_from_buffer(strlen(str), (char*)str)); +} + +str8 str8_slice(str8 s, u64 start, u64 end) +{ + ASSERT(start <= end && start <= s.len && end <= s.len); + return((str8){.len = end - start, .ptr = s.ptr + start}); +} + +str8 str8_push_buffer(mem_arena* arena, u64 len, char* buffer) +{ + str8 str = {}; + str.len = len; + str.ptr = mem_arena_alloc_array(arena, char, len); + memcpy(str.ptr, buffer, len); + return(str); +} + +str8 str8_push_cstring(mem_arena* arena, const char* str) +{ + return(str8_push_buffer(arena, strlen(str), (char*)str)); +} + +str8 str8_push_copy(mem_arena* arena, str8 s) +{ + return(str8_push_buffer(arena, str8_unbox(s))); +} + +str8 str8_push_slice(mem_arena* arena, str8 s, u64 start, u64 end) +{ + str8 slice = str8_slice(s, start, end); + return(str8_push_copy(arena, slice)); +} + +str8 str8_pushfv(mem_arena* arena, const char* format, va_list args) +{ + //NOTE(martin): + // We first compute the number of characters to write passing a size of 0. + // then we allocate len+1 (since vsnprint always terminates with a '\0'). + // We could call vsprintf since we know the size, but I'm not sure there's a hard + // guarantee that vsprintf and vsnprintf write the same number of characters in all + // and every case, and that would be a difficult bug to spot, so it seems better to + // waste one byte and be safe. + char dummy; + str8 str = {}; + va_list argCopy; + va_copy(argCopy, args); + str.len = vsnprintf(&dummy, 0, format, argCopy); + va_end(argCopy); + + str.ptr = mem_arena_alloc_array(arena, char, str.len + 1); + vsnprintf((char*)str.ptr, str.len + 1, format, args); + return(str); +} + +str8 str8_pushf(mem_arena* arena, const char* format, ...) +{ + va_list args; + va_start(args, format); + str8 str = str8_pushfv(arena, format, args); + va_end(args); + return(str); +} + +int str8_cmp(str8 s1, str8 s2) +{ + int res = strncmp(s1.ptr, s2.ptr, minimum(s1.len, s2.len)); + if(!res) + { + res = (s1.len < s2.len)? -1 : ((s1.len == s2.len)? 0 : 1); + } + return(res); +} + +char* str8_to_cstring(mem_arena* arena, str8 string) +{ + char* cstr = mem_arena_alloc_array(arena, char, string.len+1); + memcpy(cstr, string.ptr, string.len); + cstr[string.len] = '\0'; + return(cstr); +} + +//---------------------------------------------------------------------------------- +// string lists +//---------------------------------------------------------------------------------- + +void str8_list_init(str8_list* list) +{ + ListInit(&list->list); + list->eltCount = 0; + list->len = 0; +} + +void str8_list_push(mem_arena* arena, str8_list* list, str8 str) +{ + str8_elt* elt = mem_arena_alloc_type(arena, str8_elt); + elt->string = str; + ListAppend(&list->list, &elt->listElt); + list->eltCount++; + list->len += str.len; +} + +void str8_list_pushf(mem_arena* arena, str8_list* list, const char* format, ...) +{ + va_list args; + va_start(args, format); + str8 str = str8_pushfv(arena, format, args); + va_end(args); + str8_list_push(arena, list, str); +} + +str8 str8_list_collate(mem_arena* arena, str8_list list, str8 prefix, str8 separator, str8 postfix) +{ + str8 str = {}; + str.len = prefix.len + list.len + list.eltCount*separator.len + postfix.len; + str.ptr = mem_arena_alloc_array(arena, char, str.len); + char* dst = str.ptr; + memcpy(dst, prefix.ptr, prefix.len); + dst += prefix.len; + + str8_elt* elt = ListFirstEntry(&list.list, str8_elt, listElt); + if(elt) + { + memcpy(dst, elt->string.ptr, elt->string.len); + dst += elt->string.len; + elt = ListNextEntry(&list.list, elt, str8_elt, listElt); + } + + for( ; elt != 0; elt = ListNextEntry(&list.list, elt, str8_elt, listElt)) + { + memcpy(dst, separator.ptr, separator.len); + dst += separator.len; + memcpy(dst, elt->string.ptr, elt->string.len); + dst += elt->string.len; + } + memcpy(dst, postfix.ptr, postfix.len); + return(str); +} + +str8 str8_list_join(mem_arena* arena, str8_list list) +{ + str8 empty = {.len = 0, .ptr = 0}; + return(str8_list_collate(arena, list, empty, empty, empty)); +} + +str8_list str8_split(mem_arena* arena, str8 str, str8_list separators) +{ + str8_list list = {}; + ListInit(&list.list); + + char* ptr = str.ptr; + char* end = str.ptr + str.len; + char* subStart = ptr; + for(; ptr < end; ptr++) + { + //NOTE(martin): search all separators and try to match them to the current ptr + str8* foundSep = 0; + for_each_in_list(&separators.list, elt, str8_elt, listElt) + { + str8* separator = &elt->string; + bool equal = true; + for(u64 offset = 0; + (offset < separator->len) && (ptr+offset < end); + offset++) + { + if(separator->ptr[offset] != ptr[offset]) + { + equal = false; + break; + } + } + if(equal) + { + foundSep = separator; + break; + } + } + if(foundSep) + { + //NOTE(martin): we found a separator. If the start of the current substring is != ptr, + // the current substring is not empty and we emit the substring + if(ptr != subStart) + { + str8 sub = str8_from_buffer(ptr-subStart, subStart); + str8_list_push(arena, &list, sub); + } + ptr += foundSep->len - 1; //NOTE(martin): ptr is incremented at the end of the loop + subStart = ptr+1; + } + } + //NOTE(martin): emit the last substring + if(ptr != subStart) + { + str8 sub = str8_from_buffer(ptr-subStart, subStart); + str8_list_push(arena, &list, sub); + } + return(list); +} + +//---------------------------------------------------------------------------------- +// u32 strings +//---------------------------------------------------------------------------------- +str32 str32_from_buffer(u64 len, u32* buffer) +{ + return((str32){.len = len, .ptr = buffer}); +} + +str32 str32_slice(str32 s, u64 start, u64 end) +{ + ASSERT(start <= end && start <= s.len && end <= s.len); + return((str32){.len = end - start, .ptr = s.ptr + start}); +} + +str32 str32_push_buffer(mem_arena* arena, u64 len, u32* buffer) +{ + str32 str = {}; + str.len = len; + str.ptr = mem_arena_alloc_array(arena, u32, len); + memcpy(str.ptr, buffer, len*sizeof(u32)); + return(str); +} + +str32 str32_push_copy(mem_arena* arena, str32 s) +{ + return(str32_push_buffer(arena, s.len, s.ptr)); +} + +str32 str32_push_slice(mem_arena* arena, str32 s, u64 start, u64 end) +{ + str32 slice = str32_slice(s, start, end); + return(str32_push_copy(arena, slice)); +} + +//---------------------------------------------------------------------------------- +// Paths helpers +//---------------------------------------------------------------------------------- +str8 mp_path_directory(str8 fullPath) +{ + i64 lastSlashIndex = -1; + + for(i64 i = fullPath.len-1; i >= 0; i--) + { + if(fullPath.ptr[i] == '/') + { + lastSlashIndex = i; + break; + } + } + str8 directory = str8_slice(fullPath, 0, lastSlashIndex+1); + return(directory); +} + +str8 mp_path_base_name(str8 fullPath) +{ + i64 lastSlashIndex = -1; + + for(i64 i = fullPath.len-1; i >= 0; i--) + { + if(fullPath.ptr[i] == '/') + { + lastSlashIndex = i; + break; + } + } + + str8 basename = str8_slice(fullPath, lastSlashIndex+1, fullPath.len); + return(basename); +} + + + +#undef LOG_SUBSYSTEM diff --git a/src/util/strings.h b/src/util/strings.h new file mode 100644 index 0000000..612efa4 --- /dev/null +++ b/src/util/strings.h @@ -0,0 +1,96 @@ +/************************************************************//** +* +* @file: strings.h +* @author: Martin Fouilleul +* @date: 29/05/2021 +* @revision: +* +*****************************************************************/ +#ifndef __STRINGS_H_ +#define __STRINGS_H_ + +#include"typedefs.h" +#include"lists.h" +#include"memory.h" + +#ifdef __cplusplus +extern "C" { +#endif + +//---------------------------------------------------------------------------------- +// string slices as values +//---------------------------------------------------------------------------------- +typedef struct str8 +{ + u64 len; + char* ptr; +} str8; + +#define str8_lit(s) ((str8){.len = sizeof(s)-1, .ptr = (char*)(s)}) +#define str8_unbox(s) (int)((s).len), ((s).ptr) + +str8 str8_from_buffer(u64 len, char* buffer); +str8 str8_from_cstring(char* str); +str8 str8_slice(str8 s, u64 start, u64 end); + +str8 str8_push_buffer(mem_arena* arena, u64 len, char* buffer); +str8 str8_push_cstring(mem_arena* arena, const char* str); +str8 str8_push_copy(mem_arena* arena, str8 s); +str8 str8_push_slice(mem_arena* arena, str8 s, u64 start, u64 end); + +str8 str8_pushfv(mem_arena* arena, const char* format, va_list args); +str8 str8_pushf(mem_arena* arena, const char* format, ...); + +int str8_cmp(str8 s1, str8 s2); + +char* str8_to_cstring(mem_arena* arena, str8 string); +//---------------------------------------------------------------------------------- +// string lists +//---------------------------------------------------------------------------------- +typedef struct str8_elt +{ + list_elt listElt; + str8 string; +} str8_elt; + +typedef struct str8_list +{ + list_info list; + u64 eltCount; + u64 len; +} str8_list; + +void str8_list_push(mem_arena* arena, str8_list* list, str8 str); +void str8_list_pushf(mem_arena* arena, str8_list* list, const char* format, ...); + +str8 str8_list_join(mem_arena* arena, str8_list list); +str8_list str8_split(mem_arena* arena, str8 str, str8_list separators); + +//---------------------------------------------------------------------------------- +// u32 strings +//---------------------------------------------------------------------------------- +typedef struct str32 +{ + u64 len; + u32* ptr; +} str32; + +str32 str32_from_buffer(u64 len, u32* buffer); +str32 str32_slice(str32 s, u64 start, u64 end); + +str32 str32_push_buffer(mem_arena* arena, u64 len, u32* buffer); +str32 str32_push_copy(mem_arena* arena, str32 s); +str32 str32_push_slice(mem_arena* arena, str32 s, u64 start, u64 end); + +//---------------------------------------------------------------------------------- +// Paths helpers +//---------------------------------------------------------------------------------- +str8 mp_path_directory(str8 fullPath); +str8 mp_path_base_name(str8 fullPath); + +#ifdef __cplusplus +} // extern "C" +#endif + + +#endif //__STRINGS_H_ diff --git a/src/util/typedefs.h b/src/util/typedefs.h new file mode 100644 index 0000000..b0cc630 --- /dev/null +++ b/src/util/typedefs.h @@ -0,0 +1,70 @@ +//***************************************************************** +// +// $file: typedefs.h $ +// $author: Martin Fouilleul $ +// $date: 23/36/2015 $ +// $revision: $ +// $note: (C) 2015 by Martin Fouilleul - all rights reserved $ +// +//***************************************************************** +#ifndef __TYPEDEFS_H_ +#define __TYPEDEFS_H_ + +#include +#include //FLT_MAX/MIN etc... + +#ifndef __cplusplus +#include +#endif //__cplusplus + +typedef uint8_t byte; +typedef uint8_t u8; +typedef uint16_t u16; +typedef uint32_t u32; +typedef uint64_t u64; + +typedef int8_t i8; +typedef int16_t i16; +typedef int32_t i32; +typedef int64_t i64; + +typedef float f32; +typedef double f64; + +typedef union +{ + struct + { + f32 x; + f32 y; + }; + f32 c[2]; +} vec2; + +typedef union +{ + struct + { + f32 x; + f32 y; + f32 z; + f32 w; + }; + f32 c[4]; +} vec4; + +#define vec4_expand_xyz(v) (v).x, (v).y, (v).z + +typedef union +{ + struct + { + f32 x; + f32 y; + f32 w; + f32 h; + }; + f32 c[4]; +} mp_rect; + +#endif //__TYPEDEFS_H_ diff --git a/src/util/utf8.c b/src/util/utf8.c new file mode 100644 index 0000000..f0136df --- /dev/null +++ b/src/util/utf8.c @@ -0,0 +1,283 @@ +//***************************************************************** +// +// $file: utf8.c $ +// $author: Martin Fouilleul $ +// $date: 05/11/2016 $ +// $revision: $ +// $note: (C) 2016 by Martin Fouilleul - all rights reserved $ +// +//***************************************************************** +#include"utf8.h" +#include + +//----------------------------------------------------------------- +// utf-8 gore +//----------------------------------------------------------------- +const u_int32_t offsetsFromUTF8[6] = { + 0x00000000UL, 0x00003080UL, 0x000E2080UL, + 0x03C82080UL, 0xFA082080UL, 0x82082080UL }; + +const char trailingBytesForUTF8[256] = { + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, + 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 +}; + +#define utf8_is_start_byte(c) (((c)&0xc0)!=0x80) + +//----------------------------------------------------------------- +//NOTE: getting sizes / offsets / indices +//----------------------------------------------------------------- + +u32 utf8_size_from_leading_char(char leadingChar) +{ + return(trailingBytesForUTF8[(unsigned int)(unsigned char)leadingChar] + 1); +} + +u32 utf8_codepoint_size(utf32 codePoint) +{ + if(codePoint < 0x80) + { + return(1); + } + if(codePoint < 0x800) + { + return(2); + } + if(codePoint < 0x10000) + { + return(3); + } + if(codePoint < 0x110000) + { + return(4); + } + return(0); +} + +u64 utf8_codepoint_count_for_string(str8 string) +{ + u64 byteOffset = 0; + u64 codePointIndex = 0; + for(; + (byteOffset < string.len) && (string.ptr[byteOffset] != 0); + codePointIndex++) + { + utf8_dec decode = utf8_decode_at(string, byteOffset); + byteOffset += decode.size; + } + return(codePointIndex); +} + +u64 utf8_byte_count_for_codepoints(str32 codePoints) +{ + //NOTE(martin): return the exact number of bytes taken by the encoded + // version of codePoints. (ie do not attempt to provision + // for a zero terminator). + u64 byteCount = 0; + for(u64 i=0; i= string.len) + { + res = string.len; + } + else + { + u64 nextOffset = byteOffset + utf8_size_from_leading_char(string.ptr[byteOffset]); + res = minimum(nextOffset, string.len); + } + return(res); +} + +u64 utf8_prev_offset(str8 string, u64 byteOffset) +{ + u64 res = 0; + if(byteOffset > string.len) + { + res = string.len; + } + else if(byteOffset) + { + byteOffset--; + while(byteOffset > 0 && !utf8_is_start_byte(string.ptr[byteOffset])) + { + byteOffset--; + } + res = byteOffset; + } + return(res); +} + +//----------------------------------------------------------------- +//NOTE: encoding / decoding +//----------------------------------------------------------------- + +utf8_dec utf8_decode_at(str8 string, u64 offset) +{ + //NOTE(martin): get the first codepoint in str, and advance index to the + // next utf8 character + //TODO(martin): check for utf-16 surrogate pairs + utf32 cp = 0; + u64 sz = 0; + + if(offset >= string.len || !string.ptr[offset]) + { + cp = 0; + sz = 1; + } + else if( !utf8_is_start_byte(string.ptr[offset])) + { + //NOTE(martin): unexpected continuation or invalid character. + cp = 0xfffd; + sz = 1; + } + else + { + int expectedSize = utf8_size_from_leading_char(string.ptr[offset]); + do + { + /*NOTE(martin): + we shift 6 bits and add the next byte at each round. + at the end we have our utf8 codepoint, added to the shifted versions + of the utf8 leading bits for each encoded byte. These values are + precomputed in offsetsFromUTF8. + */ + unsigned char b = string.ptr[offset]; + cp <<= 6; + cp += b; + offset += 1; + sz++; + + if(b == 0xc0 || b == 0xc1 || b >= 0xc5) + { + //NOTE(martin): invalid byte encountered + break; + } + + } while( offset < string.len + && string.ptr[offset] + && !utf8_is_start_byte(string.ptr[offset]) + && sz < expectedSize); + + if(sz != expectedSize) + { + //NOTE(martin): if we encountered an error, we return the replacement codepoint U+FFFD + cp = 0xfffd; + } + else + { + cp -= offsetsFromUTF8[sz-1]; + + //NOTE(martin): check for invalid codepoints + if(cp > 0x10ffff || (cp >= 0xd800 && cp <= 0xdfff)) + { + cp = 0xfffd; + } + } + } + utf8_dec res = {.codepoint = cp, .size = sz}; + return(res); +} + +utf8_dec utf8_decode(str8 string) +{ + return(utf8_decode_at(string, 0)); +} + +str8 utf8_encode(char* dest, utf32 codePoint) +{ + u64 sz = 0; + if (codePoint < 0x80) + { + dest[0] = (char)codePoint; + sz = 1; + } + else if (codePoint < 0x800) + { + dest[0] = (codePoint>>6) | 0xC0; + dest[1] = (codePoint & 0x3F) | 0x80; + sz = 2; + } + else if (codePoint < 0x10000) + { + dest[0] = (codePoint>>12) | 0xE0; + dest[1] = ((codePoint>>6) & 0x3F) | 0x80; + dest[2] = (codePoint & 0x3F) | 0x80; + sz = 3; + } + else if (codePoint < 0x110000) + { + dest[0] = (codePoint>>18) | 0xF0; + dest[1] = ((codePoint>>12) & 0x3F) | 0x80; + dest[2] = ((codePoint>>6) & 0x3F) | 0x80; + dest[3] = (codePoint & 0x3F) | 0x80; + sz = 4; + } + str8 res = {.len = sz, .ptr = dest}; + return(res); +} + +str32 utf8_to_codepoints(u64 maxCount, utf32* backing, str8 string) +{ + u64 codePointIndex = 0; + u64 byteOffset = 0; + for(; codePointIndex < maxCount && byteOffset < string.len; codePointIndex++) + { + utf8_dec decode = utf8_decode_at(string, byteOffset); + backing[codePointIndex] = decode.codepoint; + byteOffset += decode.size; + } + str32 res = {.len = codePointIndex, .ptr = backing}; + return(res); +} + +str8 utf8_from_codepoints(u64 maxBytes, char* backing, str32 codePoints) +{ + u64 byteOffset = 0; + for(u64 codePointIndex = 0; (codePointIndex < codePoints.len); codePointIndex++) + { + utf32 codePoint = codePoints.ptr[codePointIndex]; + u32 byteCount = utf8_codepoint_size(codePoint); + if(byteOffset + byteCount > maxBytes) + { + break; + } + utf8_encode(backing+byteOffset, codePoint); + byteOffset += byteCount; + } + str8 res = {.len = byteOffset, .ptr = backing}; + return(res); +} + +str32 utf8_push_to_codepoints(mem_arena* arena, str8 string) +{ + u64 count = utf8_codepoint_count_for_string(string); + utf32* backing = mem_arena_alloc_array(arena, utf32, count); + str32 res = utf8_to_codepoints(count, backing, string); + return(res); +} + +str8 utf8_push_from_codepoints(mem_arena* arena, str32 codePoints) +{ + u64 count = utf8_byte_count_for_codepoints(codePoints); + char* backing = mem_arena_alloc_array(arena, char, count); + str8 res = utf8_from_codepoints(count, backing, codePoints); + return(res); +} + +#define UNICODE_RANGE(start, cnt, name) const unicode_range _cat2_(UNICODE_RANGE_, name) = { .firstCodePoint = start, .count = cnt }; +UNICODE_RANGES +#undef UNICODE_RANGE diff --git a/src/util/utf8.h b/src/util/utf8.h new file mode 100644 index 0000000..06fd1c8 --- /dev/null +++ b/src/util/utf8.h @@ -0,0 +1,201 @@ +//***************************************************************** +// +// $file: utf8.h $ +// $author: Martin Fouilleul $ +// $date: 05/11/2016 $ +// $revision: $ +// $note: (C) 2016 by Martin Fouilleul - all rights reserved $ +// +//***************************************************************** +#ifndef __UTF8_H_ +#define __UTF8_H_ + +#include"typedefs.h" +#include"macro_helpers.h" + +#ifdef __cplusplus +extern "C" { +#endif + +typedef u32 utf32; + +//----------------------------------------------------------------- +//NOTE: getting sizes / offsets / indices +//----------------------------------------------------------------- +u32 utf8_size_from_leading_char(char leadingChar); +u32 utf8_codepoint_size(utf32 codePoint); + +u64 utf8_codepoint_count_for_string(str8 string); +u64 utf8_byte_count_for_codepoints(str32 codePoints); + +u64 utf8_next_offset(str8 string, u64 byteOffset); +u64 utf8_prev_offset(str8 string, u64 byteOffset); + +//----------------------------------------------------------------- +//NOTE: encoding / decoding +//----------------------------------------------------------------- +typedef struct utf8_dec +{ + utf32 codepoint; //NOTE: decoded codepoint + u32 size; //NOTE: size of corresponding utf8 sequence +} utf8_dec; + +utf8_dec utf8_decode(str8 string); //NOTE: decode a single utf8 sequence at start of string +utf8_dec utf8_decode_at(str8 string, u64 offset); //NOTE: decode a single utf8 sequence starting at byte offset +str8 utf8_encode(char* dst, utf32 codePoint); //NOTE: encode codepoint into backing buffer dst + +str32 utf8_to_codepoints(u64 maxCount, utf32* backing, str8 string); +str8 utf8_from_codepoints(u64 maxBytes, char* backing, str32 codePoints); + +str32 utf8_push_to_codepoints(mem_arena* arena, str8 string); +str8 utf8_push_from_codepoints(mem_arena* arena, str32 codePoints); + +//----------------------------------------------------------------- +// utf8 range struct and X-macros for defining utf8 ranges +//----------------------------------------------------------------- + +typedef struct unicode_range +{ + utf32 firstCodePoint; + u32 count; +} unicode_range; + +//NOTE(martin): range declared here are defined in utf8.cpp +// they can be used by prefixing them with UTF8_RANGE_, as in 'UTF8_RANGE_BASIC_LATIN' +#define UNICODE_RANGES \ +UNICODE_RANGE(0x0000, 127, BASIC_LATIN) \ +UNICODE_RANGE(0x0080, 127, C1_CONTROLS_AND_LATIN_1_SUPPLEMENT) \ +UNICODE_RANGE(0x0100, 127, LATIN_EXTENDED_A) \ +UNICODE_RANGE(0x0180, 207, LATIN_EXTENDED_B) \ +UNICODE_RANGE(0x0250, 95, IPA_EXTENSIONS) \ +UNICODE_RANGE(0x02b0, 79, SPACING_MODIFIER_LETTERS) \ +UNICODE_RANGE(0x0300, 111, COMBINING_DIACRITICAL_MARKS) \ +UNICODE_RANGE(0x0370, 143, GREEK_COPTIC) \ +UNICODE_RANGE(0x0400, 255, CYRILLIC) \ +UNICODE_RANGE(0x0500, 47, CYRILLIC_SUPPLEMENT) \ +UNICODE_RANGE(0x0530, 95, ARMENIAN) \ +UNICODE_RANGE(0x0590, 111, HEBREW) \ +UNICODE_RANGE(0x0600, 255, ARABIC) \ +UNICODE_RANGE(0x0700, 79, SYRIAC) \ +UNICODE_RANGE(0x0780, 63, THAANA) \ +UNICODE_RANGE(0x0900, 127, DEVANAGARI) \ +UNICODE_RANGE(0x0980, 127, BENGALI_ASSAMESE) \ +UNICODE_RANGE(0x0a00, 127, GURMUKHI) \ +UNICODE_RANGE(0x0a80, 127, GUJARATI) \ +UNICODE_RANGE(0x0b00, 127, ORIYA) \ +UNICODE_RANGE(0x0b80, 127, TAMIL) \ +UNICODE_RANGE(0x0c00, 127, TELUGU) \ +UNICODE_RANGE(0x0c80, 127, KANNADA) \ +UNICODE_RANGE(0x0d00, 255, MALAYALAM) \ +UNICODE_RANGE(0x0d80, 127, SINHALA) \ +UNICODE_RANGE(0x0e00, 127, THAI) \ +UNICODE_RANGE(0x0e80, 127, LAO) \ +UNICODE_RANGE(0x0f00, 255, TIBETAN) \ +UNICODE_RANGE(0x1000, 159, MYANMAR) \ +UNICODE_RANGE(0x10a0, 95, GEORGIAN) \ +UNICODE_RANGE(0x1100, 255, HANGUL_JAMO) \ +UNICODE_RANGE(0x1200, 383, ETHIOPIC) \ +UNICODE_RANGE(0x13a0, 95, CHEROKEE) \ +UNICODE_RANGE(0x1400, 639, UNIFIED_CANADIAN_ABORIGINAL_SYLLABICS) \ +UNICODE_RANGE(0x1680, 31, OGHAM) \ +UNICODE_RANGE(0x16a0, 95, RUNIC) \ +UNICODE_RANGE(0x1700, 31, TAGALOG) \ +UNICODE_RANGE(0x1720, 31, HANUNOO) \ +UNICODE_RANGE(0x1740, 31, BUHID) \ +UNICODE_RANGE(0x1760, 31, TAGBANWA) \ +UNICODE_RANGE(0x1780, 127, KHMER) \ +UNICODE_RANGE(0x1800, 175, MONGOLIAN) \ +UNICODE_RANGE(0x1900, 79, LIMBU) \ +UNICODE_RANGE(0x1950, 47, TAI_LE) \ +UNICODE_RANGE(0x19e0, 31, KHMER_SYMBOLS) \ +UNICODE_RANGE(0x1d00, 127, PHONETIC_EXTENSIONS) \ +UNICODE_RANGE(0x1e00, 255, LATIN_EXTENDED_ADDITIONAL) \ +UNICODE_RANGE(0x1f00, 255, GREEK_EXTENDED) \ +UNICODE_RANGE(0x2000, 111, GENERAL_PUNCTUATION) \ +UNICODE_RANGE(0x2070, 47, SUPERSCRIPTS_AND_SUBSCRIPTS) \ +UNICODE_RANGE(0x20a0, 47, CURRENCY_SYMBOLS) \ +UNICODE_RANGE(0x20d0, 47, COMBINING_DIACRITICAL_MARKS_FOR_SYMBOLS) \ +UNICODE_RANGE(0x2100, 79, LETTERLIKE_SYMBOLS) \ +UNICODE_RANGE(0x2150, 63, NUMBER_FORMS) \ +UNICODE_RANGE(0x2190, 111, ARROWS) \ +UNICODE_RANGE(0x2200, 255, MATHEMATICAL_OPERATORS) \ +UNICODE_RANGE(0x2300, 255, MISCELLANEOUS_TECHNICAL) \ +UNICODE_RANGE(0x2400, 63, CONTROL_PICTURES) \ +UNICODE_RANGE(0x2440, 31, OPTICAL_CHARACTER_RECOGNITION) \ +UNICODE_RANGE(0x2460, 159, ENCLOSED_ALPHANUMERICS) \ +UNICODE_RANGE(0x2500, 127, BOX_DRAWING) \ +UNICODE_RANGE(0x2580, 31, BLOCK_ELEMENTS) \ +UNICODE_RANGE(0x25a0, 95, GEOMETRIC_SHAPES) \ +UNICODE_RANGE(0x2600, 255, MISCELLANEOUS_SYMBOLS) \ +UNICODE_RANGE(0x2700, 191, DINGBATS) \ +UNICODE_RANGE(0x27c0, 47, MISCELLANEOUS_MATHEMATICAL_SYMBOLS_A) \ +UNICODE_RANGE(0x27f0, 15, SUPPLEMENTAL_ARROWS_A) \ +UNICODE_RANGE(0x2800, 255, BRAILLE_PATTERNS) \ +UNICODE_RANGE(0x2900, 127, SUPPLEMENTAL_ARROWS_B) \ +UNICODE_RANGE(0x2980, 127, MISCELLANEOUS_MATHEMATICAL_SYMBOLS_B) \ +UNICODE_RANGE(0x2a00, 255, SUPPLEMENTAL_MATHEMATICAL_OPERATORS) \ +UNICODE_RANGE(0x2b00, 255, MISCELLANEOUS_SYMBOLS_AND_ARROWS) \ +UNICODE_RANGE(0x2e80, 127, CJK_RADICALS_SUPPLEMENT) \ +UNICODE_RANGE(0x2f00, 223, KANGXI_RADICALS) \ +UNICODE_RANGE(0x2ff0, 15, IDEOGRAPHIC_DESCRIPTION_CHARACTERS) \ +UNICODE_RANGE(0x3000, 63, CJK_SYMBOLS_AND_PUNCTUATION) \ +UNICODE_RANGE(0x3040, 95, HIRAGANA) \ +UNICODE_RANGE(0x30a0, 95, KATAKANA) \ +UNICODE_RANGE(0x3100, 47, BOPOMOFO) \ +UNICODE_RANGE(0x3130, 95, HANGUL_COMPATIBILITY_JAMO) \ +UNICODE_RANGE(0x3190, 15, KANBUN_KUNTEN) \ +UNICODE_RANGE(0x31a0, 31, BOPOMOFO_EXTENDED) \ +UNICODE_RANGE(0x31f0, 15, KATAKANA_PHONETIC_EXTENSIONS) \ +UNICODE_RANGE(0x3200, 255, ENCLOSED_CJK_LETTERS_AND_MONTHS) \ +UNICODE_RANGE(0x3300, 255, CJK_COMPATIBILITY) \ +UNICODE_RANGE(0x3400, 6591, CJK_UNIFIED_IDEOGRAPHS_EXTENSION_A) \ +UNICODE_RANGE(0x4dc0, 63, YIJING_HEXAGRAM_SYMBOLS) \ +UNICODE_RANGE(0x4e00, 20911, CJK_UNIFIED_IDEOGRAPHS) \ +UNICODE_RANGE(0xa000, 1167, YI_SYLLABLES) \ +UNICODE_RANGE(0xa490, 63, YI_RADICALS) \ +UNICODE_RANGE(0xac00, 11183, HANGUL_SYLLABLES) \ +UNICODE_RANGE(0xd800, 1023, HIGH_SURROGATE_AREA) \ +UNICODE_RANGE(0xdc00, 1023, LOW_SURROGATE_AREA) \ +UNICODE_RANGE(0xe000, 6399, PRIVATE_USE_AREA) \ +UNICODE_RANGE(0xf900, 511, CJK_COMPATIBILITY_IDEOGRAPHS) \ +UNICODE_RANGE(0xfb00, 79, ALPHABETIC_PRESENTATION_FORMS) \ +UNICODE_RANGE(0xfb50, 687, ARABIC_PRESENTATION_FORMS_A) \ +UNICODE_RANGE(0xfe00, 15, VARIATION_SELECTORS) \ +UNICODE_RANGE(0xfe20, 15, COMBINING_HALF_MARKS) \ +UNICODE_RANGE(0xfe30, 31, CJK_COMPATIBILITY_FORMS) \ +UNICODE_RANGE(0xfe50, 31, SMALL_FORM_VARIANTS) \ +UNICODE_RANGE(0xfe70, 143, ARABIC_PRESENTATION_FORMS_B) \ +UNICODE_RANGE(0xff00, 239, HALFWIDTH_AND_FULLWIDTH_FORMS) \ +UNICODE_RANGE(0xfff0, 15, SPECIALS) \ +UNICODE_RANGE(0x10000, 127, LINEAR_B_SYLLABARY) \ +UNICODE_RANGE(0x10080, 127, LINEAR_B_IDEOGRAMS) \ +UNICODE_RANGE(0x10100, 63, AEGEAN_NUMBERS) \ +UNICODE_RANGE(0x10300, 47, OLD_ITALIC) \ +UNICODE_RANGE(0x10330, 31, GOTHIC) \ +UNICODE_RANGE(0x10380, 31, UGARITIC) \ +UNICODE_RANGE(0x10400, 79, DESERET) \ +UNICODE_RANGE(0x10450, 47, SHAVIAN) \ +UNICODE_RANGE(0x10480, 47, OSMANYA) \ +UNICODE_RANGE(0x10800, 63, CYPRIOT_SYLLABARY) \ +UNICODE_RANGE(0x1d000, 255, BYZANTINE_MUSICAL_SYMBOLS) \ +UNICODE_RANGE(0x1d100, 255, MUSICAL_SYMBOLS) \ +UNICODE_RANGE(0x1d300, 95, TAI_XUAN_JING_SYMBOLS) \ +UNICODE_RANGE(0x1d400, 1023, MATHEMATICAL_ALPHANUMERIC_SYMBOLS) \ +UNICODE_RANGE(0x20000, 42719, CJK_UNIFIED_IDEOGRAPHS_EXTENSION_B) \ +UNICODE_RANGE(0x2f800, 543, CJK_COMPATIBILITY_IDEOGRAPHS_SUPPLEMENT) \ +UNICODE_RANGE(0xe0000, 127, TAGS) \ +UNICODE_RANGE(0xe0100, 239, VARIATION_SELECTORS_SUPPLEMENT) \ +UNICODE_RANGE(0xf0000, 65533, SUPPLEMENTARY_PRIVATE_USE_AREA_A) \ +UNICODE_RANGE(0x100000, 65533, SUPPLEMENTARY_PRIVATE_USE_AREA_B) + + +#define UNICODE_RANGE(start, count, name) \ + extern const unicode_range _cat2_(UNICODE_RANGE_, name); +UNICODE_RANGES +#undef UNICODE_RANGE + +#ifdef __cplusplus +} // extern "C" +#endif + +#endif //__UTF8_H_ diff --git a/todo.txt b/todo.txt new file mode 100644 index 0000000..af71165 --- /dev/null +++ b/todo.txt @@ -0,0 +1,115 @@ + + +Shortlist +--------- +[x] let pass flat args in ui_size_push() and ui_box_set_size() +[>>] separate style stacks +[>] margins? as part of size, or different styling stack? +[>] animation time stack +[ ] Let build code set target style directly, and animate from current to target +[ ] filter styles stack by tag +[ ] image backgrounds/gradients? +[ ] animating open/close widgets? + + +Canvas Drawing +-------------- +[.] Correctly handle resizing / viewport + [x] associate surfaces with mp_windows + [x] window resize resizes surface. Surface always renders to whole mp_window + [x] Add ability to create sub-views, and create surfaces for these + - window comes with a main view, so we can get a surface for the whole window + [>] Clean native mp_window_data struct (don't need to cache a lot of stuff here) + +[.] Add images bliting + [x] Clean, rename uv vs texUV stuff + [>] Destroy stuff + [>] More unified handle system + [.] Rounded images (sortof) + [ ] path clipped images + +[ ] Add color gradients? +[ ] Make canvas implicit? +[/] Handle based error signaling +[/] Allow polling events in main thread, and updating/rendering in background thread. + +[x] font metrics shouldn't depend on surface, & font + shouldn't really depend on canvas either??? + > meaning we should be able to pass fonts without canvas in ui + > and only pass canvas when drawing at the end... + +UI +-- +[x] Make gui context implicit? +[x] Uniform ui_box struct + cache widgets +[x] Prune unused boxes +[.] Layout boxes + [x] basic two pass layout + [x] Layout from start or end + [>] Add overflow flags to layout, & solve conflicts + +[x] Temporarily push transform and text flip when rendering UI, so that the coord system is y down, origin at top left +[x] Canvas render the same size on a high-dpi surface + > it works with abstract 'pixel' units, which are transformed to pixels in the shader, according to backing store scaling + +[.] Style struct and style stack + [ ] Maybe use individual stack for different style attributes +[x] Pass initial style in ui_begin_frame() + +[.] Draw boxes + [x] use flags to enable/disable drawing each feature + [ ] active/hovered transitions + +[x] Change input state handling a move it to app layer +[x] Compute signals for ui_box +[x] Use ui_size_push() and pass axis instead of ui_width/height_push() +[>] Use value is ui_size as margin when kind == text or == children? + +[ ] Allow animating sizes according to hot/active? + +[ ] Basic helpers + [.] button + [.] slider (or rather, scroll bar) + [.] simple spacers + [ ] have a flag for non-cached stuff + [.] scrolling panel + [ ] Allow/disallow scrolling in x/y + [ ] Scroll with mousewheel + [/] add margins to scrollbars (disallow scrollbars crossing) + +[?] Maybe let builder code handle "active"/"hot" state, since it + depends on the widgets + [ ] On the other hand, this state must be set before layouting, in + particular font/fontSize -> maybe do a pass for static layout, + instead of doing it in box creation... + [ ] this way we can compute styling after user has set active/hot, but before layout + > Maybe just let user set style selector, and provide persistent state bits that can + > be used in any way? (to replace eg active/hot?) or perhaps not needed if we have just + > 'dragging' state + + +[x] Mask mouse outside of parent rects -> maintain clip stack and clip mouse against it +[x] Mask mouse below panels/other widgets +[x] popups and tooltips + [x] allow pushing/popping boxes irrespective of parent/child relation + [x] ui_begin_frame() prepares two containers, user ui goes in the first one + [x] tooltips and menus go to the second one + [x] Add menus +[ ] line editing widget + +Misc +---- +[x] Split metal surface and metal painter (but put them in the same compilation unit?) +[x] Have only one rect struct +[x] Shorten mp_string to str8 + +[ ] Better/Simpler time API + +[/] Frame throttling + [x] For now, we always wait on vblank during mg_surface_present(), regardless of target fps + [ ] Then actually get the correct display interval from the surface's current monitor + [ ] Allow waiting for more display interval than one? (ie allow throttling at 30fps for a 60fps display) + +[/] split osx_app and move all platform independant stuff outside + +[ ] Sort out mg_matrix_push/pop() -> transform vs. set...