On 20 June 2012 18:41, Anuj Phogat <<a href="mailto:anuj.phogat@gmail.com" target="_blank">anuj.phogat@gmail.com</a>> wrote: <div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> <div class="im">On Sat, Jun 16, 2012 at 8:09 AM, Paul Berry <<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a>> wrote: > On 8 June 2012 14:46, <<a href="mailto:anuj.phogat@gmail.com">anuj.phogat@gmail.com</a>> wrote: >> >> From: Anuj Phogat <<a href="mailto:anuj.phogat@gmail.com">anuj.phogat@gmail.com</a>> >> >> This test varifies that with GL_SAMPLE_ALPHA_TO_ONE enabled, each sample's >> alpha value is replaced by maximum alpha value i.e. 1.0. > > > This test is: > > Reviewed-by: Paul Berry <<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a>> > > I would recommend adding two additional tests, preferably in separate files: > > 1. We should make sure that when both GL_SAMPLE_ALPHA_TO_ONE and > GL_SAMPLE_ALPHA_TO_COVERAGE are enabled, GL_SAMPLE_ALPHA_TO_COVERAGE takes > effect first (that is, the coverage value needs to be computed from the > alpha value *before* the alpha value gets changed to 1). > </div>To avoid lot of code duplication, I included this as a sub test in sample-alpha-to-coverage test case: [PATCH 3/4] Add test to verify sample alpha to coverage with multisample fbo. Do you want me to split it into a separate test case? </blockquote><div> No, you're right. Combining the tests makes good sense. </div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> <div class="im"> > 2. We should make sure that if GL_SAMPLE_ALPHA_TO_ONE is enabled and > GL_MULTISAMPLE is disabled, alpha values are not modified. Also, we should > make sure that if GL_SAMPLE_ALPHA_TO_ONE is enabled and the buffer is > single-sampled, alpha values are not modified. > </div>I will work on this test case. Thanks Anuj </blockquote></div>