On 02/17/2015 04:35 PM, Ilia Mirkin wrote: > On Tue, Feb 17, 2015 at 7:12 PM, Jordan Justen > <[email protected]> wrote: >> +++ >> b/tests/spec/arb_shader_atomic_counters/execution/vs-simple-inc-dec-read.shader_test >> @@ -0,0 +1,70 @@ >> +# Simple test of atomicCounterIncrement, atomicCounterDecrement and >> +# atomicCounter being used in the VS. >> + >> +[require] >> +GL CORE >= 3.1 >> +GLSL >= 1.40 > > Why require core? Mesa exposes this in compat contexts too... (Also > you could just require GLSL >= 1.40 afaik...) > >> +probe all rgba 0.0 1.0 0.0 1.0 >> \ No newline at end of file > > Please fix :) > > May be good to also make a fs version of this. Lastly, an > implementation is only required to support this in fs... for the rest, > MAX_*_ATOMIC_COUNTERS may be 0. I guess there's no good way to > represent that in piglit... thoughts?
Some MAX_* constants, like GL_MAX_VERTEX_UNIFORM_COMPONENTS, are available to the [require] section. Since MAX_*_ATOMIC_COUNTERS can be zero, we should probably support for it. Looking at the way the GL_MAX_* checks are structured in shader_runner.c, it could probably use a refactor to a table-based approach. > -ilia > _______________________________________________ > mesa-dev mailing list > [email protected] > http://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ Piglit mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/piglit
