--- /dev/null
+#pragma clang diagnostic ignored "-Wmissing-prototypes"
+#pragma clang diagnostic ignored "-Wmissing-braces"
+
+#include <metal_stdlib>
+#include <simd/simd.h>
+
+using namespace metal;
+
+template<typename T, size_t Num>
+struct spvUnsafeArray
+{
+ T elements[Num ? Num : 1];
+
+ thread T& operator [] (size_t pos) thread
+ {
+ return elements[pos];
+ }
+ constexpr const thread T& operator [] (size_t pos) const thread
+ {
+ return elements[pos];
+ }
+
+ device T& operator [] (size_t pos) device
+ {
+ return elements[pos];
+ }
+ constexpr const device T& operator [] (size_t pos) const device
+ {
+ return elements[pos];
+ }
+
+ constexpr const constant T& operator [] (size_t pos) const constant
+ {
+ return elements[pos];
+ }
+
+ threadgroup T& operator [] (size_t pos) threadgroup
+ {
+ return elements[pos];
+ }
+ constexpr const threadgroup T& operator [] (size_t pos) const threadgroup
+ {
+ return elements[pos];
+ }
+};
+
+struct Meep
+{
+ float a;
+ float b;
+};
+
+struct Block
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+};
+
+struct Block_1
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+};
+
+struct main0_out
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+ spvUnsafeArray<float, 2> Block_a;
+ float Block_b;
+ float2x2 Block_m;
+ Meep Block_meep;
+ spvUnsafeArray<Meep, 2> Block_meeps;
+ float4 gl_Position;
+};
+
+struct main0_in
+{
+ spvUnsafeArray<float, 2> in_a;
+ float in_b;
+ float2x2 in_m;
+ Meep in_meep;
+ spvUnsafeArray<Meep, 2> in_meeps;
+ spvUnsafeArray<float, 2> Block_a;
+ float Block_b;
+ float2x2 Block_m;
+ Meep Block_meep;
+ spvUnsafeArray<Meep, 2> Block_meeps;
+};
+
+kernel void main0(uint3 gl_GlobalInvocationID [[thread_position_in_grid]], device main0_out* spvOut [[buffer(28)]], constant uint* spvIndirectParams [[buffer(29)]], device MTLQuadTessellationFactorsHalf* spvTessLevel [[buffer(26)]], device main0_in* spvIn [[buffer(22)]])
+{
+ device main0_out* gl_out = &spvOut[gl_GlobalInvocationID.x - gl_GlobalInvocationID.x % 4];
+ device main0_in* gl_in = &spvIn[min(gl_GlobalInvocationID.x / 4, spvIndirectParams[1] - 1) * spvIndirectParams[0]];
+ uint gl_InvocationID = gl_GlobalInvocationID.x % 4;
+ uint gl_PrimitiveID = min(gl_GlobalInvocationID.x / 4, spvIndirectParams[1]);
+ gl_out[gl_InvocationID].gl_Position = float4(1.0);
+ gl_out[gl_InvocationID].a[0] = gl_in[gl_InvocationID].in_a[0];
+ gl_out[gl_InvocationID].a[1] = gl_in[gl_InvocationID].in_a[1];
+ gl_out[gl_InvocationID].b = gl_in[gl_InvocationID].in_b;
+ gl_out[gl_InvocationID].m = gl_in[gl_InvocationID].in_m;
+ gl_out[gl_InvocationID].meep.a = gl_in[gl_InvocationID].in_meep.a;
+ gl_out[gl_InvocationID].meep.b = gl_in[gl_InvocationID].in_meep.b;
+ gl_out[gl_InvocationID].meeps[0].a = gl_in[gl_InvocationID].in_meeps[0].a;
+ gl_out[gl_InvocationID].meeps[0].b = gl_in[gl_InvocationID].in_meeps[0].b;
+ gl_out[gl_InvocationID].meeps[1].a = gl_in[gl_InvocationID].in_meeps[1].a;
+ gl_out[gl_InvocationID].meeps[1].b = gl_in[gl_InvocationID].in_meeps[1].b;
+ gl_out[gl_InvocationID].Block_a[0] = gl_in[gl_InvocationID].Block_a[0];
+ gl_out[gl_InvocationID].Block_a[1] = gl_in[gl_InvocationID].Block_a[1];
+ gl_out[gl_InvocationID].Block_b = gl_in[gl_InvocationID].Block_b;
+ gl_out[gl_InvocationID].Block_m = gl_in[gl_InvocationID].Block_m;
+ gl_out[gl_InvocationID].Block_meep.a = gl_in[gl_InvocationID].Block_meep.a;
+ gl_out[gl_InvocationID].Block_meep.b = gl_in[gl_InvocationID].Block_meep.b;
+ gl_out[gl_InvocationID].Block_meeps[0].a = gl_in[gl_InvocationID].Block_meeps[0].a;
+ gl_out[gl_InvocationID].Block_meeps[0].b = gl_in[gl_InvocationID].Block_meeps[0].b;
+ gl_out[gl_InvocationID].Block_meeps[1].a = gl_in[gl_InvocationID].Block_meeps[1].a;
+ gl_out[gl_InvocationID].Block_meeps[1].b = gl_in[gl_InvocationID].Block_meeps[1].b;
+}
+
--- /dev/null
+#pragma clang diagnostic ignored "-Wmissing-prototypes"
+#pragma clang diagnostic ignored "-Wmissing-braces"
+
+#include <metal_stdlib>
+#include <simd/simd.h>
+
+using namespace metal;
+
+template<typename T, size_t Num>
+struct spvUnsafeArray
+{
+ T elements[Num ? Num : 1];
+
+ thread T& operator [] (size_t pos) thread
+ {
+ return elements[pos];
+ }
+ constexpr const thread T& operator [] (size_t pos) const thread
+ {
+ return elements[pos];
+ }
+
+ device T& operator [] (size_t pos) device
+ {
+ return elements[pos];
+ }
+ constexpr const device T& operator [] (size_t pos) const device
+ {
+ return elements[pos];
+ }
+
+ constexpr const constant T& operator [] (size_t pos) const constant
+ {
+ return elements[pos];
+ }
+
+ threadgroup T& operator [] (size_t pos) threadgroup
+ {
+ return elements[pos];
+ }
+ constexpr const threadgroup T& operator [] (size_t pos) const threadgroup
+ {
+ return elements[pos];
+ }
+};
+
+struct Meep
+{
+ float a;
+ float b;
+};
+
+struct Block
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+};
+
+struct Block_1
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+};
+
+struct main0_out
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+ spvUnsafeArray<float, 2> Block_a;
+ float Block_b;
+ float2x2 Block_m;
+ Meep Block_meep;
+ spvUnsafeArray<Meep, 2> Block_meeps;
+ float4 gl_Position;
+};
+
+struct main0_in
+{
+ float in_a_0 [[attribute(0)]];
+ float in_a_1 [[attribute(1)]];
+ float in_b [[attribute(2)]];
+ float2 in_m_0 [[attribute(3)]];
+ float2 in_m_1 [[attribute(4)]];
+ float Meep_a [[attribute(5)]];
+ float Meep_b [[attribute(6)]];
+ float Block_a_0 [[attribute(11)]];
+ float Block_a_1 [[attribute(12)]];
+ float Block_b [[attribute(13)]];
+ float2 Block_m_0 [[attribute(14)]];
+ float2 Block_m_1 [[attribute(15)]];
+};
+
+kernel void main0(main0_in in [[stage_in]], uint gl_InvocationID [[thread_index_in_threadgroup]], uint gl_PrimitiveID [[threadgroup_position_in_grid]], device main0_out* spvOut [[buffer(28)]], constant uint* spvIndirectParams [[buffer(29)]], device MTLQuadTessellationFactorsHalf* spvTessLevel [[buffer(26)]], threadgroup main0_in* gl_in [[threadgroup(0)]])
+{
+ device main0_out* gl_out = &spvOut[gl_PrimitiveID * 4];
+ if (gl_InvocationID < spvIndirectParams[0])
+ gl_in[gl_InvocationID] = in;
+ threadgroup_barrier(mem_flags::mem_threadgroup);
+ if (gl_InvocationID >= 4)
+ return;
+ gl_out[gl_InvocationID].gl_Position = float4(1.0);
+ gl_out[gl_InvocationID].a[0] = gl_in[gl_InvocationID].in_a_0;
+ gl_out[gl_InvocationID].a[1] = gl_in[gl_InvocationID].in_a_1;
+ gl_out[gl_InvocationID].b = gl_in[gl_InvocationID].in_b;
+ float2x2 _178 = float2x2(gl_in[gl_InvocationID].in_m_0, gl_in[gl_InvocationID].in_m_1);
+ gl_out[gl_InvocationID].m = _178;
+ gl_out[gl_InvocationID].meep.a = gl_in[gl_InvocationID].Meep_a;
+ gl_out[gl_InvocationID].meep.b = gl_in[gl_InvocationID].Meep_b;
+ gl_out[gl_InvocationID].meeps[0].a = 1.0;
+ gl_out[gl_InvocationID].meeps[0].b = 2.0;
+ gl_out[gl_InvocationID].meeps[1].a = 3.0;
+ gl_out[gl_InvocationID].meeps[1].b = 4.0;
+ gl_out[gl_InvocationID].Block_a[0] = gl_in[gl_InvocationID].Block_a_0;
+ gl_out[gl_InvocationID].Block_a[1] = gl_in[gl_InvocationID].Block_a_1;
+ gl_out[gl_InvocationID].Block_b = gl_in[gl_InvocationID].Block_b;
+ float2x2 _216 = float2x2(gl_in[gl_InvocationID].Block_m_0, gl_in[gl_InvocationID].Block_m_1);
+ gl_out[gl_InvocationID].Block_m = _216;
+ gl_out[gl_InvocationID].Block_meep.a = 10.0;
+ gl_out[gl_InvocationID].Block_meep.b = 20.0;
+ gl_out[gl_InvocationID].Block_meeps[0].a = 5.0;
+ gl_out[gl_InvocationID].Block_meeps[0].b = 6.0;
+ gl_out[gl_InvocationID].Block_meeps[1].a = 7.0;
+ gl_out[gl_InvocationID].Block_meeps[1].b = 8.0;
+}
+
--- /dev/null
+#pragma clang diagnostic ignored "-Wmissing-prototypes"
+#pragma clang diagnostic ignored "-Wmissing-braces"
+
+#include <metal_stdlib>
+#include <simd/simd.h>
+
+using namespace metal;
+
+template<typename T, size_t Num>
+struct spvUnsafeArray
+{
+ T elements[Num ? Num : 1];
+
+ thread T& operator [] (size_t pos) thread
+ {
+ return elements[pos];
+ }
+ constexpr const thread T& operator [] (size_t pos) const thread
+ {
+ return elements[pos];
+ }
+
+ device T& operator [] (size_t pos) device
+ {
+ return elements[pos];
+ }
+ constexpr const device T& operator [] (size_t pos) const device
+ {
+ return elements[pos];
+ }
+
+ constexpr const constant T& operator [] (size_t pos) const constant
+ {
+ return elements[pos];
+ }
+
+ threadgroup T& operator [] (size_t pos) threadgroup
+ {
+ return elements[pos];
+ }
+ constexpr const threadgroup T& operator [] (size_t pos) const threadgroup
+ {
+ return elements[pos];
+ }
+};
+
+struct Meep
+{
+ float a;
+ float b;
+};
+
+struct Block
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+};
+
+struct Block_1
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+};
+
+struct main0_out
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+ spvUnsafeArray<float, 2> Block_a;
+ float Block_b;
+ float2x2 Block_m;
+ Meep Block_meep;
+ spvUnsafeArray<Meep, 2> Block_meeps;
+ float4 gl_Position;
+};
+
+struct main0_in
+{
+ spvUnsafeArray<float, 2> in_a;
+ float in_b;
+ float2x2 in_m;
+ Meep in_meep;
+ spvUnsafeArray<Meep, 2> in_meeps;
+ spvUnsafeArray<float, 2> Block_a;
+ float Block_b;
+ float2x2 Block_m;
+ Meep Block_meep;
+ spvUnsafeArray<Meep, 2> Block_meeps;
+};
+
+static inline __attribute__((always_inline))
+void write_in_func(device main0_out* thread & gl_out, thread uint& gl_InvocationID, device main0_in* thread & gl_in)
+{
+ gl_out[gl_InvocationID].gl_Position = float4(1.0);
+ gl_out[gl_InvocationID].a[0] = gl_in[gl_InvocationID].in_a[0];
+ gl_out[gl_InvocationID].a[1] = gl_in[gl_InvocationID].in_a[1];
+ gl_out[gl_InvocationID].b = gl_in[gl_InvocationID].in_b;
+ gl_out[gl_InvocationID].m = gl_in[gl_InvocationID].in_m;
+ gl_out[gl_InvocationID].meep.a = gl_in[gl_InvocationID].in_meep.a;
+ gl_out[gl_InvocationID].meep.b = gl_in[gl_InvocationID].in_meep.b;
+ gl_out[gl_InvocationID].meeps[0].a = gl_in[gl_InvocationID].in_meeps[0].a;
+ gl_out[gl_InvocationID].meeps[0].b = gl_in[gl_InvocationID].in_meeps[0].b;
+ gl_out[gl_InvocationID].meeps[1].a = gl_in[gl_InvocationID].in_meeps[1].a;
+ gl_out[gl_InvocationID].meeps[1].b = gl_in[gl_InvocationID].in_meeps[1].b;
+ gl_out[gl_InvocationID].Block_a[0] = gl_in[gl_InvocationID].Block_a[0];
+ gl_out[gl_InvocationID].Block_a[1] = gl_in[gl_InvocationID].Block_a[1];
+ gl_out[gl_InvocationID].Block_b = gl_in[gl_InvocationID].Block_b;
+ gl_out[gl_InvocationID].Block_m = gl_in[gl_InvocationID].Block_m;
+ gl_out[gl_InvocationID].Block_meep.a = gl_in[gl_InvocationID].Block_meep.a;
+ gl_out[gl_InvocationID].Block_meep.b = gl_in[gl_InvocationID].Block_meep.b;
+ gl_out[gl_InvocationID].Block_meeps[0].a = gl_in[gl_InvocationID].Block_meeps[0].a;
+ gl_out[gl_InvocationID].Block_meeps[0].b = gl_in[gl_InvocationID].Block_meeps[0].b;
+ gl_out[gl_InvocationID].Block_meeps[1].a = gl_in[gl_InvocationID].Block_meeps[1].a;
+ gl_out[gl_InvocationID].Block_meeps[1].b = gl_in[gl_InvocationID].Block_meeps[1].b;
+}
+
+kernel void main0(uint3 gl_GlobalInvocationID [[thread_position_in_grid]], device main0_out* spvOut [[buffer(28)]], constant uint* spvIndirectParams [[buffer(29)]], device MTLQuadTessellationFactorsHalf* spvTessLevel [[buffer(26)]], device main0_in* spvIn [[buffer(22)]])
+{
+ device main0_out* gl_out = &spvOut[gl_GlobalInvocationID.x - gl_GlobalInvocationID.x % 4];
+ device main0_in* gl_in = &spvIn[min(gl_GlobalInvocationID.x / 4, spvIndirectParams[1] - 1) * spvIndirectParams[0]];
+ uint gl_InvocationID = gl_GlobalInvocationID.x % 4;
+ uint gl_PrimitiveID = min(gl_GlobalInvocationID.x / 4, spvIndirectParams[1]);
+ write_in_func(gl_out, gl_InvocationID, gl_in);
+}
+
--- /dev/null
+#pragma clang diagnostic ignored "-Wmissing-prototypes"
+#pragma clang diagnostic ignored "-Wmissing-braces"
+
+#include <metal_stdlib>
+#include <simd/simd.h>
+
+using namespace metal;
+
+template<typename T, size_t Num>
+struct spvUnsafeArray
+{
+ T elements[Num ? Num : 1];
+
+ thread T& operator [] (size_t pos) thread
+ {
+ return elements[pos];
+ }
+ constexpr const thread T& operator [] (size_t pos) const thread
+ {
+ return elements[pos];
+ }
+
+ device T& operator [] (size_t pos) device
+ {
+ return elements[pos];
+ }
+ constexpr const device T& operator [] (size_t pos) const device
+ {
+ return elements[pos];
+ }
+
+ constexpr const constant T& operator [] (size_t pos) const constant
+ {
+ return elements[pos];
+ }
+
+ threadgroup T& operator [] (size_t pos) threadgroup
+ {
+ return elements[pos];
+ }
+ constexpr const threadgroup T& operator [] (size_t pos) const threadgroup
+ {
+ return elements[pos];
+ }
+};
+
+struct Meep
+{
+ float a;
+ float b;
+};
+
+struct Block
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+};
+
+struct Block_1
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+};
+
+struct main0_out
+{
+ spvUnsafeArray<float, 2> a;
+ float b;
+ float2x2 m;
+ Meep meep;
+ spvUnsafeArray<Meep, 2> meeps;
+ spvUnsafeArray<float, 2> Block_a;
+ float Block_b;
+ float2x2 Block_m;
+ Meep Block_meep;
+ spvUnsafeArray<Meep, 2> Block_meeps;
+ float4 gl_Position;
+};
+
+struct main0_in
+{
+ float in_a_0 [[attribute(0)]];
+ float in_a_1 [[attribute(1)]];
+ float in_b [[attribute(2)]];
+ float2 in_m_0 [[attribute(3)]];
+ float2 in_m_1 [[attribute(4)]];
+ float Meep_a [[attribute(5)]];
+ float Meep_b [[attribute(6)]];
+ float Block_a_0 [[attribute(11)]];
+ float Block_a_1 [[attribute(12)]];
+ float Block_b [[attribute(13)]];
+ float2 Block_m_0 [[attribute(14)]];
+ float2 Block_m_1 [[attribute(15)]];
+};
+
+static inline __attribute__((always_inline))
+void write_in_func(device main0_out* thread & gl_out, thread uint& gl_InvocationID, threadgroup main0_in* thread & gl_in)
+{
+ gl_out[gl_InvocationID].gl_Position = float4(1.0);
+ gl_out[gl_InvocationID].a[0] = gl_in[gl_InvocationID].in_a_0;
+ gl_out[gl_InvocationID].a[1] = gl_in[gl_InvocationID].in_a_1;
+ gl_out[gl_InvocationID].b = gl_in[gl_InvocationID].in_b;
+ float2x2 _72 = float2x2(gl_in[gl_InvocationID].in_m_0, gl_in[gl_InvocationID].in_m_1);
+ gl_out[gl_InvocationID].m = _72;
+ gl_out[gl_InvocationID].meep.a = gl_in[gl_InvocationID].Meep_a;
+ gl_out[gl_InvocationID].meep.b = gl_in[gl_InvocationID].Meep_b;
+ gl_out[gl_InvocationID].meeps[0].a = 1.0;
+ gl_out[gl_InvocationID].meeps[0].b = 2.0;
+ gl_out[gl_InvocationID].meeps[1].a = 3.0;
+ gl_out[gl_InvocationID].meeps[1].b = 4.0;
+ gl_out[gl_InvocationID].Block_a[0] = gl_in[gl_InvocationID].Block_a_0;
+ gl_out[gl_InvocationID].Block_a[1] = gl_in[gl_InvocationID].Block_a_1;
+ gl_out[gl_InvocationID].Block_b = gl_in[gl_InvocationID].Block_b;
+ float2x2 _134 = float2x2(gl_in[gl_InvocationID].Block_m_0, gl_in[gl_InvocationID].Block_m_1);
+ gl_out[gl_InvocationID].Block_m = _134;
+ gl_out[gl_InvocationID].Block_meep.a = 10.0;
+ gl_out[gl_InvocationID].Block_meep.b = 20.0;
+ gl_out[gl_InvocationID].Block_meeps[0].a = 5.0;
+ gl_out[gl_InvocationID].Block_meeps[0].b = 6.0;
+ gl_out[gl_InvocationID].Block_meeps[1].a = 7.0;
+ gl_out[gl_InvocationID].Block_meeps[1].b = 8.0;
+}
+
+kernel void main0(main0_in in [[stage_in]], uint gl_InvocationID [[thread_index_in_threadgroup]], uint gl_PrimitiveID [[threadgroup_position_in_grid]], device main0_out* spvOut [[buffer(28)]], constant uint* spvIndirectParams [[buffer(29)]], device MTLQuadTessellationFactorsHalf* spvTessLevel [[buffer(26)]], threadgroup main0_in* gl_in [[threadgroup(0)]])
+{
+ device main0_out* gl_out = &spvOut[gl_PrimitiveID * 4];
+ if (gl_InvocationID < spvIndirectParams[0])
+ gl_in[gl_InvocationID] = in;
+ threadgroup_barrier(mem_flags::mem_threadgroup);
+ if (gl_InvocationID >= 4)
+ return;
+ write_in_func(gl_out, gl_InvocationID, gl_in);
+}
+
--- /dev/null
+#version 450
+layout(vertices = 4) out;
+
+struct Meep
+{
+ float a;
+ float b;
+};
+
+layout(location = 0) out float a[][2];
+layout(location = 2) out float b[];
+layout(location = 3) out mat2 m[];
+layout(location = 5) out Meep meep[];
+layout(location = 7) out Meep meeps[][2];
+
+layout(location = 11) out Block
+{
+ float a[2];
+ float b;
+ mat2 m;
+ Meep meep;
+ Meep meeps[2];
+} B[];
+
+layout(location = 0) in float in_a[][2];
+layout(location = 2) in float in_b[];
+layout(location = 3) in mat2 in_m[];
+layout(location = 5) in Meep in_meep[];
+layout(location = 7) in Meep in_meeps[][2];
+
+layout(location = 11) in Block
+{
+ float a[2];
+ float b;
+ mat2 m;
+ Meep meep;
+ Meep meeps[2];
+} in_B[];
+
+void write_in_func()
+{
+ gl_out[gl_InvocationID].gl_Position = vec4(1.0);
+
+ a[gl_InvocationID][0] = in_a[gl_InvocationID][0];
+ a[gl_InvocationID][1] = in_a[gl_InvocationID][1];
+ b[gl_InvocationID] = in_b[gl_InvocationID];
+ m[gl_InvocationID] = in_m[gl_InvocationID];
+ meep[gl_InvocationID].a = in_meep[gl_InvocationID].a;
+ meep[gl_InvocationID].b = in_meep[gl_InvocationID].b;
+ meeps[gl_InvocationID][0].a = in_meeps[gl_InvocationID][0].a;
+ meeps[gl_InvocationID][0].b = in_meeps[gl_InvocationID][0].b;
+ meeps[gl_InvocationID][1].a = in_meeps[gl_InvocationID][1].a;
+ meeps[gl_InvocationID][1].b = in_meeps[gl_InvocationID][1].b;
+
+ B[gl_InvocationID].a[0] = in_B[gl_InvocationID].a[0];
+ B[gl_InvocationID].a[1] = in_B[gl_InvocationID].a[1];
+ B[gl_InvocationID].b = in_B[gl_InvocationID].b;
+ B[gl_InvocationID].m = in_B[gl_InvocationID].m;
+ B[gl_InvocationID].meep.a = in_B[gl_InvocationID].meep.a;
+ B[gl_InvocationID].meep.b = in_B[gl_InvocationID].meep.b;
+ B[gl_InvocationID].meeps[0].a = in_B[gl_InvocationID].meeps[0].a;
+ B[gl_InvocationID].meeps[0].b = in_B[gl_InvocationID].meeps[0].b;
+ B[gl_InvocationID].meeps[1].a = in_B[gl_InvocationID].meeps[1].a;
+ B[gl_InvocationID].meeps[1].b = in_B[gl_InvocationID].meeps[1].b;
+}
+
+void main()
+{
+ write_in_func();
+}
--- /dev/null
+#version 450
+layout(vertices = 4) out;
+
+struct Meep
+{
+ float a;
+ float b;
+};
+
+layout(location = 0) out float a[][2];
+layout(location = 2) out float b[];
+layout(location = 3) out mat2 m[];
+layout(location = 5) out Meep meep[];
+layout(location = 7) out Meep meeps[][2];
+
+layout(location = 11) out Block
+{
+ float a[2];
+ float b;
+ mat2 m;
+ Meep meep;
+ Meep meeps[2];
+} B[];
+
+layout(location = 0) in float in_a[][2];
+layout(location = 2) in float in_b[];
+layout(location = 3) in mat2 in_m[];
+layout(location = 5) in Meep in_meep[];
+
+layout(location = 11) in Block
+{
+ float a[2];
+ float b;
+ mat2 m;
+ // Non-multi-patch path cannot support structs inside structs.
+} in_B[];
+
+void write_in_func()
+{
+ gl_out[gl_InvocationID].gl_Position = vec4(1.0);
+
+ a[gl_InvocationID][0] = in_a[gl_InvocationID][0];
+ a[gl_InvocationID][1] = in_a[gl_InvocationID][1];
+ b[gl_InvocationID] = in_b[gl_InvocationID];
+ m[gl_InvocationID] = in_m[gl_InvocationID];
+ meep[gl_InvocationID].a = in_meep[gl_InvocationID].a;
+ meep[gl_InvocationID].b = in_meep[gl_InvocationID].b;
+ meeps[gl_InvocationID][0].a = 1.0;
+ meeps[gl_InvocationID][0].b = 2.0;
+ meeps[gl_InvocationID][1].a = 3.0;
+ meeps[gl_InvocationID][1].b = 4.0;
+
+ B[gl_InvocationID].a[0] = in_B[gl_InvocationID].a[0];
+ B[gl_InvocationID].a[1] = in_B[gl_InvocationID].a[1];
+ B[gl_InvocationID].b = in_B[gl_InvocationID].b;
+ B[gl_InvocationID].m = in_B[gl_InvocationID].m;
+ B[gl_InvocationID].meep.a = 10.0;
+ B[gl_InvocationID].meep.b = 20.0;
+ B[gl_InvocationID].meeps[0].a = 5.0;
+ B[gl_InvocationID].meeps[0].b = 6.0;
+ B[gl_InvocationID].meeps[1].a = 7.0;
+ B[gl_InvocationID].meeps[1].b = 8.0;
+}
+
+void main()
+{
+ write_in_func();
+}
uint32_t locn = get_decoration(var.self, DecorationLocation);
if (storage == StorageClassInput)
{
- type_id = ensure_correct_input_type(var.basetype, locn, location_meta ? location_meta->num_components : 0);
+ type_id = ensure_correct_input_type(var.basetype, locn,
+ location_meta ? location_meta->num_components : 0,
+ meta.strip_array);
if (!location_meta)
var.basetype = type_id;
uint32_t locn = get_decoration(var.self, DecorationLocation) + i;
if (storage == StorageClassInput)
{
- var.basetype = ensure_correct_input_type(var.basetype, locn);
- uint32_t mbr_type_id = ensure_correct_input_type(usable_type->self, locn);
+ var.basetype = ensure_correct_input_type(var.basetype, locn, 0, meta.strip_array);
+ uint32_t mbr_type_id = ensure_correct_input_type(usable_type->self, locn, 0, meta.strip_array);
if (storage == StorageClassInput && pull_model_inputs.count(var.self))
ib_type.member_types[ib_mbr_idx] = build_msl_interpolant_type(mbr_type_id, is_noperspective);
else
uint32_t locn = get_member_decoration(var_type.self, mbr_idx, DecorationLocation);
if (storage == StorageClassInput)
{
- mbr_type_id = ensure_correct_input_type(mbr_type_id, locn);
+ mbr_type_id = ensure_correct_input_type(mbr_type_id, locn, 0, meta.strip_array);
var_type.member_types[mbr_idx] = mbr_type_id;
if (storage == StorageClassInput && pull_model_inputs.count(var.self))
ib_type.member_types[ib_mbr_idx] = build_msl_interpolant_type(mbr_type_id, is_noperspective);
uint32_t locn = get_accumulated_member_location(var, mbr_idx, meta.strip_array);
if (storage == StorageClassInput)
{
- mbr_type_id = ensure_correct_input_type(mbr_type_id, locn);
+ mbr_type_id = ensure_correct_input_type(mbr_type_id, locn, 0, meta.strip_array);
var_type.member_types[mbr_idx] = mbr_type_id;
if (storage == StorageClassInput && pull_model_inputs.count(var.self))
ib_type.member_types[ib_mbr_idx] = build_msl_interpolant_type(mbr_type_id, is_noperspective);
// Ensure that the type is compatible with the shader input.
// If it is, simply return the given type ID.
// Otherwise, create a new type, and return its ID.
-uint32_t CompilerMSL::ensure_correct_input_type(uint32_t type_id, uint32_t location, uint32_t num_components)
+uint32_t CompilerMSL::ensure_correct_input_type(uint32_t type_id, uint32_t location, uint32_t num_components, bool strip_array)
{
auto &type = get<SPIRType>(type_id);
- // Struct types must match exactly.
- if (type.basetype == SPIRType::Struct)
+ uint32_t max_array_dimensions = strip_array ? 1 : 0;
+
+ // Struct and array types must match exactly.
+ if (type.basetype == SPIRType::Struct || type.array.size() > max_array_dimensions)
return type_id;
auto p_va = inputs_by_location.find(location);
void mark_location_as_used_by_shader(uint32_t location, const SPIRType &type, spv::StorageClass storage);
uint32_t ensure_correct_builtin_type(uint32_t type_id, spv::BuiltIn builtin);
- uint32_t ensure_correct_input_type(uint32_t type_id, uint32_t location, uint32_t num_components = 0);
+ uint32_t ensure_correct_input_type(uint32_t type_id, uint32_t location,
+ uint32_t num_components, bool strip_array);
void emit_custom_templates();
void emit_custom_functions();