From a3d859fb4f3c55eccdd82b5969e8aabd48fa1ae6 Mon Sep 17 00:00:00 2001
From: Michael Fabian 'Xaymar' Dirks <info@xaymar.com>
Date: Mon, 8 Jan 2018 15:23:25 +0100
Subject: [PATCH] util-math: Implement aligned vec3 and vec4 versions

A bug in Visual C++ 2013 32-bit & 2015 32-bit causes the C++ compiler to incorrectly align the vec3 and vec4 structs to 8-byte instead of 16-byte, resulting in a crash if the target PC supports SSE. Visual Studio 2017 and 64-bit builds are not affected.

Related: #9
---
 source/util-math.cpp | 32 ++++++++++++++++++++++++++++++++
 source/util-math.h   | 23 +++++++++++++++++++++++
 2 files changed, 55 insertions(+)

diff --git a/source/util-math.cpp b/source/util-math.cpp
index 4e5d3c92..8a420274 100644
--- a/source/util-math.cpp
+++ b/source/util-math.cpp
@@ -18,4 +18,36 @@
 */
 
 #include "util-math.h"
+#include "util-memory.h"
 
+void* util::vec3a::operator new(size_t count){
+	return malloc_aligned(16, count);
+}
+
+void* util::vec3a::operator new[](size_t count) {
+	return malloc_aligned(16, count);
+}
+
+void util::vec3a::operator delete(void* p) {
+	free_aligned(p);
+}
+
+void util::vec3a::operator delete[](void* p) {
+	free_aligned(p);
+}
+
+void* util::vec4a::operator new(size_t count) {
+	return malloc_aligned(16, count);
+}
+
+void* util::vec4a::operator new[](size_t count) {
+	return malloc_aligned(16, count);
+}
+
+void util::vec4a::operator delete(void* p) {
+	free_aligned(p);
+}
+
+void util::vec4a::operator delete[](void* p) {
+	free_aligned(p);
+}
diff --git a/source/util-math.h b/source/util-math.h
index 3fcfd50b..35bd3a43 100644
--- a/source/util-math.h
+++ b/source/util-math.h
@@ -21,6 +21,12 @@
 #include <math.h>
 #include <inttypes.h>
 
+// OBS
+#include <graphics/vec2.h>
+#include <graphics/vec3.h>
+#include <graphics/vec4.h>
+
+// Constants
 #define PI		3.1415926535897932384626433832795
 #define PI2		6.283185307179586476925286766559
 #define PI2_SQROOT	2.506628274631000502415765284811		
@@ -39,3 +45,20 @@ inline size_t GetNearestPowerOfTwoAbove(size_t v) {
 inline size_t GetNearestPowerOfTwoBelow(size_t v) {
 	return 1ull << size_t(floor(log10(double(v)) / log10(2.0)));
 }
+
+
+namespace util {
+	__declspec(align(16)) struct vec3a : public vec3 {
+		static void* vec3a::operator new(size_t count);
+		static void* vec3a::operator new[](size_t count);
+		static void vec3a::operator delete(void* p);
+		static void vec3a::operator delete[](void* p);
+	};
+
+	__declspec(align(16)) struct vec4a : public vec4 {
+		static void* vec4a::operator new(size_t count);
+		static void* vec4a::operator new[](size_t count);
+		static void vec4a::operator delete(void* p);
+		static void vec4a::operator delete[](void* p);
+	};
+}
\ No newline at end of file