test/tint/buffer/uniform/std140/unnested/mat4x3_f32/to_private.wgsl.expected.msl

#include <metal_stdlib>
using namespace metal;

template<typename T, size_t N>
struct tint_array {
  const constant T& operator[](size_t i) const constant { return elements[i]; }
  device T& operator[](size_t i) device { return elements[i]; }
  const device T& operator[](size_t i) const device { return elements[i]; }
  thread T& operator[](size_t i) thread { return elements[i]; }
  const thread T& operator[](size_t i) const thread { return elements[i]; }
  threadgroup T& operator[](size_t i) threadgroup { return elements[i]; }
  const threadgroup T& operator[](size_t i) const threadgroup { return elements[i]; }
  T elements[N];
};

struct tint_packed_vec3_f32_array_element {
  /* 0x0000 */ packed_float3 packed;
  /* 0x000c */ tint_array<int8_t, 4> tint_pad;
};

struct tint_module_vars_struct {
  const constant tint_array<tint_packed_vec3_f32_array_element, 4>* u;
  thread float4x3* p;
};

kernel void f(const constant tint_array<tint_packed_vec3_f32_array_element, 4>* u [[buffer(0)]]) {
  thread float4x3 p = float4x3(0.0f);
  tint_module_vars_struct const tint_module_vars = tint_module_vars_struct{.u=u, .p=(&p)};
  tint_array<tint_packed_vec3_f32_array_element, 4> const v = (*tint_module_vars.u);
  (*tint_module_vars.p) = float4x3(float3(v[0u].packed), float3(v[1u].packed), float3(v[2u].packed), float3(v[3u].packed));
  (*tint_module_vars.p)[1u] = float3((*tint_module_vars.u)[0u].packed);
  (*tint_module_vars.p)[1u] = float3((*tint_module_vars.u)[0u].packed).zxy;
  (*tint_module_vars.p)[0u].y = (*tint_module_vars.u)[1u].packed.x;
}