const_shader_layout 0.4.0

Compile-time checks to ensure structs conform to WGSL's memory layout rules
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

#[cfg(feature = "glam")]
use glam::{
    I16Vec2, I16Vec3, I16Vec4, IVec2, IVec3, IVec4, Mat2, Mat3A, Mat4, Quat, U16Vec2, U16Vec3,
    U16Vec4, UVec2, UVec3, UVec4, Vec2, Vec3, Vec3A, Vec4,
};

use crate::{
    impl_shader_layout_array_element, impl_shader_layout_compat,
    impl_shader_layout_compat_array_element,
};

// Vector
impl_shader_layout_compat!(4, I16Vec2, U16Vec2);
impl_shader_layout_compat!(8, I16Vec3, U16Vec3);
impl_shader_layout_compat!(8, I16Vec4, U16Vec4);
impl_shader_layout_compat!(8, IVec2, UVec2, Vec2);
impl_shader_layout_compat!(16, IVec3, UVec3, Vec3);
impl_shader_layout_compat!(16, IVec4, UVec4, Vec4, Quat);

// Matrix
impl_shader_layout_compat!(8, Mat2);
// Use `Mat3A` as `Mat3`'s column vectors are not properly aligned.
impl_shader_layout_compat!(16, Mat3A);
impl_shader_layout_compat!(16, Mat4);

// Array
// `*Vec3` shouldn't be implemented, because the size of `[Vec3; N]` != `N * roundUp(AlignOf(E), SizeOf(E))`
impl_shader_layout_array_element!(I16Vec2, U16Vec2, I16Vec4, U16Vec4, IVec2, UVec2, Vec2);
// Bypass `ShaderLayoutCompat` check and implement `ShaderLayoutCompatArrayElement` for `Vec3A`
impl crate::ShaderLayoutArrayElement for Vec3A {}
impl crate::ShaderLayoutCompatArrayElement for Vec3A {}

// Only 16-byte aligned types should implement `ShaderLayoutCompatArrayElement`
impl_shader_layout_compat_array_element!(IVec4, UVec4, Vec4, Quat);
impl_shader_layout_compat_array_element!(Mat2, Mat3A, Mat4);