Re: [PATCH 5/5] d3dcompiler_43/tests: Added trigonometric function tests to HLSL test suite

[Henri Verbeet]

On 12 November 2010 21:08, Travis Athougies <iammisc(a)gmail.com> wrote:
> +    static struct hlsl_probe_info sin_probes[] =
> This should be const. As a general rule, try to pay attention to at
> least other patches to the same module. I fixed a couple of these
> about two weeks ago. Same goes for calculating the probe count.
> +    {
> +        {0, 0, {0.5000f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {1, 0, {0.5975f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {2, 0, {0.6913f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {3, 0, {0.7778f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {4, 0, {0.8536f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {5, 0, {0.9157f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {6, 0, {0.9620f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {7, 0, {0.9904f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {8, 0, {1.0000f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {9, 0, {0.9904f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {10, 0, {0.9619f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {11, 0, {0.9157f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {12, 0, {0.8536f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {13, 0, {0.7778f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {14, 0, {0.6913f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {15, 0, {0.5975f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {16, 0, {0.5000f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {17, 0, {0.4025f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {18, 0, {0.3087f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {19, 0, {0.2222f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {20, 0, {0.1464f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {21, 0, {0.0843f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {22, 0, {0.0381f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {23, 0, {0.0096f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {24, 0, {0.0000f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {25, 0, {0.0096f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {26, 0, {0.0381f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {27, 0, {0.0843f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {28, 0, {0.1464f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {29, 0, {0.2222f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {30, 0, {0.3087f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +        {31, 0, {0.4025f, 0.0f, 0.0f, 0.0f}, 0.001f, "sin test failed"},
> +    };
> +
> 32 probes is probably a bit excessive, but I guess it's not a major
> problem. Note that you can easily return the cosine results in the .y
> component though, instead of duplicating the complete test.