On Mon, Aug 26, 2013 at 10:26 PM, Ian Romanick <[email protected]> wrote: > Have you tried this on other drivers? I've seen many cases where things > that seem like the should be errors get optimized out before the error gets > generated. While this is a known "bug" in those implementations, it's > generally better to test the thing you're trying to test instead of testing > something else with your test when you're testing a test. Sorry, I got > carried away there. :)
I haven't tested on other drivers. (Someone buy me something!) Are you saying that you don't think this is necessarily in spec either (like the bad_constant_folding() trick), or just that it may not work on other drivers? If it's in spec, I'm not sure I can care too much about proprietary drivers. Within spec and confirms a piece of functionality in our driver seem to be the two most important criteria, and maybe not even in that order. _______________________________________________ Piglit mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/piglit
