#include "BLI_assert.h"
#include "BLI_sys_types.h"
#include <math.h>
#include "BLI_math_matrix_types.hh"
#include "BLI_math_vector_types.hh"

Macros
#define	GPU_THREAD error_not_in_a_shader_question_mark

Typedefs
using	bool32_t = int32_t
using	packed_float3 = blender::float3
using	packed_int3 = blender::int3
using	packed_uint3 = blender::uint3
using	float2
using	float4
using	int2
using	int4
using	uint2
using	uint4
using	float3
using	int3
using	uint3
using	float2x2
using	float4x2
using	float2x4
using	float3x4
using	float4x4

Detailed Description

Glue definition to make shared declaration of struct & functions work in both C / C++ and GLSL. We use the same vector and matrix types as Blender C++. Some math functions are defined to use the float version to match the GLSL syntax. This file can be used for C & C++ code and the syntax used should follow the same rules. Some preprocessing is done by the GPU back-end to make it GLSL compatible.

IMPORTANT:

Always use u suffix for enum values. GLSL do not support implicit cast.
Define all values. This is in order to simplify custom pre-processor code.
(C++ only) Always use uint32_t as underlying type (enum MyEnum : uint32_t).
(C only) do NOT use the enum type inside UBO/SSBO structs and use uint instead.
Use float suffix by default for float literals to avoid double promotion in C++.
Pack one float or int after a vec3/ivec3 to fulfill alignment rules.

NOTE: Due to alignment restriction and buggy drivers, do not try to use mat3 inside structs. NOTE: (UBO only) Do not use arrays of float. They are padded to arrays of vec4 and are not worth it. This does not apply to SSBO.

IMPORTANT: Do not forget to align mat4, vec3 and vec4 to 16 bytes, and vec2 to 8 bytes.

NOTE: You can use bool type using bool32_t a int boolean type matching the GLSL type.

Definition in file GPU_shader_shared_utils.hh.