#pragma clang diagnostic ignored "-Wmissing-prototypes"

#include <metal_stdlib>
#include <simd/simd.h>

using namespace metal;

template<typename T, uint N>
inline void spvArrayCopyFromConstantToStack(thread T (&dst)[N], constant T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromConstantToThreadGroup(threadgroup T (&dst)[N], constant T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromStackToStack(thread T (&dst)[N], thread const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromStackToThreadGroup(threadgroup T (&dst)[N], thread const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromThreadGroupToStack(thread T (&dst)[N], threadgroup const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromThreadGroupToThreadGroup(threadgroup T (&dst)[N], threadgroup const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromDeviceToDevice(device T (&dst)[N], device const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromConstantToDevice(device T (&dst)[N], constant T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromStackToDevice(device T (&dst)[N], thread const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromThreadGroupToDevice(device T (&dst)[N], threadgroup const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromDeviceToStack(thread T (&dst)[N], device const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

template<typename T, uint N>
inline void spvArrayCopyFromDeviceToThreadGroup(threadgroup T (&dst)[N], device const T (&src)[N])
{
    for (uint i = 0; i < N; i++)
    {
        dst[i] = src[i];
    }
}

struct Data
{
    float a;
    float b;
};

constant float X_tmp [[function_constant(0)]];
constant float X = is_function_constant_defined(X_tmp) ? X_tmp : 4.0;

struct Data_1
{
    float a;
    float b;
};

struct SSBO
{
    Data_1 outdata[1];
};

constant uint3 gl_WorkGroupSize [[maybe_unused]] = uint3(2u, 1u, 1u);

kernel void main0(device SSBO& _53 [[buffer(0)]], uint3 gl_WorkGroupID [[threadgroup_position_in_grid]], uint3 gl_LocalInvocationID [[thread_position_in_threadgroup]])
{
    Data _25[2] = { Data{ 1.0, 2.0 }, Data{ 3.0, 4.0 } };
    
    Data _31[2] = { Data{ X, 2.0 }, Data{ 3.0, 5.0 } };
    Data data2[2];
    spvArrayCopyFromStackToStack(data2, _31);
    _53.outdata[gl_WorkGroupID.x].a = _25[gl_LocalInvocationID.x].a + data2[gl_LocalInvocationID.x].a;
    _53.outdata[gl_WorkGroupID.x].b = _25[gl_LocalInvocationID.x].b + data2[gl_LocalInvocationID.x].b;
}