Re: [PATCH 0/1] MR3263: d3dcompiler/tests: Add some tests for matrix packing order compile flags.

[Matteo Bruni (＠Mystral)]

Matteo Bruni (@Mystral) commented about dlls/d3dcompiler_43/tests/hlsl_d3d9.c:
> - if (ps_code)
> - {
> -
>    hr = pD3DXGetShaderConstantTable(ID3D10Blob_GetBufferPointer(ps_code), &constants);
> -
>    ok(hr == D3D_OK, "Got unexpected hr %#lx.\n", hr);
> -
>    hr = ID3DXConstantTable_SetMatrix(constants, device, "r", &matrix);
> -
>    ok(hr == D3D_OK, "Got unexpected hr %#lx.\n", hr);
> -
>    hr = ID3DXConstantTable_SetMatrix(constants, device, "c", &matrix);
> -
>    ok(hr == D3D_OK, "Got unexpected hr %#lx.\n", hr);
> -
>    ID3DXConstantTable_Release(constants);
> -
>        h = ID3DXConstantTable_GetConstantByName(constants, NULL, "m1");
> -
>        ok(!!h, "Failed to find a constant.\n");
> -
>        count = 1;
> -
>        hr = ID3DXConstantTable_GetConstantDesc(constants, h, &desc, &count);
> -
>        ok(hr == D3D_OK, "Got unexpected hr %#lx.\n", hr);
> -
>        hr = IDirect3DDevice9_SetPixelShaderConstantF(test_context.device, desc.RegisterIndex,
> -
>                data, desc.RegisterCount);

It might make sense to add a comment somewhere explaining why we're using SetPixelShaderCosntantF() instead of SetMatrix() and discouraging attempts to revert the change at some point in the future...