From ffb900d872a456cc5e7cc1762b1846c0c4df08cd Mon Sep 17 00:00:00 2001
From: Martin Fouilleul <martinfouilleul@gmail.com>
Date: Mon, 26 Jun 2023 16:06:26 +0200
Subject: [PATCH] String functions that return str8/16/32 allocated on an arena
 implicitly zero-terminate strings, so we can avoid one copy when passing to C
 stdlib APIs

---
 src/platform/osx_path.m      |  3 ++-
 src/platform/platform_path.h |  6 ++++++
 src/platform/win32_path.c    |  6 ++++--
 src/util/strings.c           | 35 ++++++++++++++++++-----------------
 src/util/strings.h           | 12 +++++++++++-
 5 files changed, 41 insertions(+), 21 deletions(-)

diff --git a/src/platform/osx_path.m b/src/platform/osx_path.m
index 575f6c8..28c7eb4 100644
--- a/src/platform/osx_path.m
+++ b/src/platform/osx_path.m
@@ -23,8 +23,9 @@ str8 path_executable(mem_arena* arena)
 	u32 size = 0;
 	_NSGetExecutablePath(0, &size);
 	result.len = size;
-	result.ptr = mem_arena_alloc_array(arena, char, result.len);
+	result.ptr = mem_arena_alloc_array(arena, char, result.len+1);
 	_NSGetExecutablePath(result.ptr, &size);
+	result.ptr[result.len] = '\0';
 	return(result);
 }}
 
diff --git a/src/platform/platform_path.h b/src/platform/platform_path.h
index 7665204..da1c23f 100644
--- a/src/platform/platform_path.h
+++ b/src/platform/platform_path.h
@@ -10,6 +10,12 @@
 
 #include"util/strings.h"
 
+/*NOTE:
+	by convention, functions that take an arena and return a path
+	allocated on that arena allocate 1 more character and null-terminate
+	the string.
+*/
+
 MP_API str8 path_slice_directory(str8 path);
 MP_API str8 path_slice_filename(str8 path);
 
diff --git a/src/platform/win32_path.c b/src/platform/win32_path.c
index a3d8bad..9f39f12 100644
--- a/src/platform/win32_path.c
+++ b/src/platform/win32_path.c
@@ -22,8 +22,11 @@ bool path_is_absolute(str8 path)
 
 str8 path_executable(mem_arena* arena)
 {
+	///////////////////////////////////////////////////////////////////
 	//TODO use wide chars
-	char* buffer = mem_arena_alloc_array(arena, char, MAX_PATH+1);
+	///////////////////////////////////////////////////////////////////
+
+	char* buffer = mem_arena_alloc_array(arena, char, MAX_PATH+2);
 	int size = GetModuleFileName(NULL, buffer, MAX_PATH+1);
 	//TODO: check for errors...
 	for(int i=0; i<size; i++)
@@ -33,7 +36,6 @@ str8 path_executable(mem_arena* arena)
 			buffer[i] = '/';
 		}
 	}
-
 	return(str8_from_buffer(size, buffer));
 }
 
diff --git a/src/util/strings.c b/src/util/strings.c
index 33e3a3f..d63c6c5 100644
--- a/src/util/strings.c
+++ b/src/util/strings.c
@@ -28,21 +28,20 @@ str8 str8_push_buffer(mem_arena* arena, u64 len, char* buffer)
 {
 	str8 str = {0};
 	str.len = len;
-	str.ptr = mem_arena_alloc_array(arena, char, len);
+	str.ptr = mem_arena_alloc_array(arena, char, len+1);
 	memcpy(str.ptr, buffer, len);
+	str.ptr[str.len] = '\0';
 	return(str);
 }
 
 str8 str8_push_cstring(mem_arena* arena, const char* str)
 {
-	if(!str)
+	int len = 0;
+	if(str)
 	{
-		return((str8){0});
-	}
-	else
-	{
-		return(str8_push_buffer(arena, strlen(str), (char*)str));
+		len = strlen(str);
 	}
+	return(str8_push_buffer(arena, strlen(str), (char*)str));
 }
 
 str8 str8_push_copy(mem_arena* arena, str8 s)
@@ -60,11 +59,8 @@ 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.
+	//  then we allocate len+1 (since vsnprint always terminates with a '\0').
+
 	char dummy;
 	str8 str = {0};
 	va_list argCopy;
@@ -137,7 +133,7 @@ str8 str8_list_collate(mem_arena* arena, str8_list list, str8 prefix, str8 separ
 {
 	str8 str = {0};
 	str.len = prefix.len + list.len + list.eltCount*separator.len + postfix.len;
-	str.ptr = mem_arena_alloc_array(arena, char, str.len);
+	str.ptr = mem_arena_alloc_array(arena, char, str.len + 1);
 	char* dst = str.ptr;
 	memcpy(dst, prefix.ptr, prefix.len);
 	dst += prefix.len;
@@ -158,6 +154,7 @@ str8 str8_list_collate(mem_arena* arena, str8_list list, str8 prefix, str8 separ
 		dst += elt->string.len;
 	}
 	memcpy(dst, postfix.ptr, postfix.len);
+	str.ptr[str.len] = '\0';
 	return(str);
 }
 
@@ -239,8 +236,9 @@ str16 str16_push_buffer(mem_arena* arena, u64 len, u16* buffer)
 {
 	str16 str = {0};
 	str.len = len;
-	str.ptr = mem_arena_alloc_array(arena, u16, len);
+	str.ptr = mem_arena_alloc_array(arena, u16, len+1);
 	memcpy(str.ptr, buffer, len*sizeof(u16));
+	str.ptr[str.len] = (u16)0;
 	return(str);
 }
 
@@ -275,7 +273,7 @@ str16 str16_list_collate(mem_arena* arena, str16_list list, str16 prefix, str16
 {
 	str16 str = {0};
 	str.len = prefix.len + list.len + list.eltCount*separator.len + postfix.len;
-	str.ptr = mem_arena_alloc_array(arena, u16, str.len);
+	str.ptr = mem_arena_alloc_array(arena, u16, str.len + 1);
 	char* dst = (char*)str.ptr;
 	memcpy(dst, prefix.ptr, prefix.len*sizeof(u16));
 	dst += prefix.len*sizeof(u16);
@@ -296,6 +294,7 @@ str16 str16_list_collate(mem_arena* arena, str16_list list, str16 prefix, str16
 		dst += elt->string.len*sizeof(u16);
 	}
 	memcpy(dst, postfix.ptr, postfix.len*sizeof(u16));
+	str.ptr[str.len] = (u16)0;
 	return(str);
 }
 
@@ -323,8 +322,9 @@ str32 str32_push_buffer(mem_arena* arena, u64 len, u32* buffer)
 {
 	str32 str = {0};
 	str.len = len;
-	str.ptr = mem_arena_alloc_array(arena, u32, len);
+	str.ptr = mem_arena_alloc_array(arena, u32, len+1);
 	memcpy(str.ptr, buffer, len*sizeof(u32));
+	str.ptr[str.len] = 0;
 	return(str);
 }
 
@@ -359,7 +359,7 @@ str32 str32_list_collate(mem_arena* arena, str32_list list, str32 prefix, str32
 {
 	str32 str = {0};
 	str.len = prefix.len + list.len + list.eltCount*separator.len + postfix.len;
-	str.ptr = mem_arena_alloc_array(arena, u32, str.len);
+	str.ptr = mem_arena_alloc_array(arena, u32, str.len+1);
 	char* dst = (char*)str.ptr;
 	memcpy(dst, prefix.ptr, prefix.len*sizeof(u32));
 	dst += prefix.len*sizeof(u32);
@@ -380,6 +380,7 @@ str32 str32_list_collate(mem_arena* arena, str32_list list, str32 prefix, str32
 		dst += elt->string.len*sizeof(u32);
 	}
 	memcpy(dst, postfix.ptr, postfix.len*sizeof(u32));
+	str.ptr[str.len] = 0;
 	return(str);
 }
 
diff --git a/src/util/strings.h b/src/util/strings.h
index fdc0876..6507ed0 100644
--- a/src/util/strings.h
+++ b/src/util/strings.h
@@ -19,8 +19,18 @@
 extern "C" {
 #endif
 
+/*NOTE:
+	By convention, functions that take an arena and return a string slice allocated on
+	this arena, always allocate one more element and null-terminate the string. This is
+	done so we can pass those strings directly to C APIs that requires C strings without
+	having to do a copy with str8_to_cstring().
+
+	This does _not_ applies to the string returned by str8_split(). Those are slices
+	into the original string. Only the _list nodes_ are allocated on the arena.
+*/
+
 //----------------------------------------------------------------------------------
-// string slices as values
+// u8 strings
 //----------------------------------------------------------------------------------
 typedef struct str8
 {