// Each #kernel tells which function to compile; you can have many kernels
#pragma kernel TextureToBuffer

Texture2D<float4> source;
RWStructuredBuffer<uint> result;
RWStructuredBuffer<uint> forceWaitResultBuffer;
uint width, height;
SamplerState MyPointRepeatSampler;
uint startIdx;
uint sentinelIdx;
uint forceWaitValue;
float gamma;

[numthreads(32,32,1)] // Must match threadsX, threadsY in CapturePanorama.cs
void TextureToBuffer (uint3 id : SV_DispatchThreadID)
{
	if (id.x >= width || id.y >= height) // In case width/height not multiple of numthreads
	    return;

	float4 color = source.SampleLevel(MyPointRepeatSampler, float2(((float)id.x + 0.5)/ width, ((float)id.y + 0.5)/ height), 0);
	color = pow(color, gamma);
	color *= 255.0;
	result[startIdx + (id.y * width) + id.x] = ((int)color.r << 16u) | ((int)color.g << 8u) | (int)color.b;

	if (id.x == width - 1u && id.y == height - 1u && id.z == 0u)
	{
        forceWaitResultBuffer[0] = forceWaitValue; // Used on CPU side to force a wait for this operation to complete
		result[sentinelIdx] = 1419455993u; // Sentinel value - must match BufferSentinelValue in CapturePanorama.cs
	}
}