On Wed, Jul 15, 2009 at 10:05:29AM +0100, Keith Whitwell wrote: | Basically I think the problem is with the spec,...
The functionality is pretty useful. Certainly you can object to the way it's implemented (and people have done so for a long time). | ... and glean unfortunately | has decided to test a part of the spec that is pretty problematic to | implement. Glean has to check rendered images for correctness, so it needs to know whether it's valid to write a value into a color channel, read it back, and perform a comparison. If the spec didn't have language describing precisely the conditions under which that's supposed to work, I guess I would have written a test that attempted to discover which values behaved correctly, and then all other tests would use only those values. I've never understood why people obssess about consistent test failures. If you've analyzed the test, and you understand why it failed, and you're comfortable with the behavior, then just forget about it. Generally you should be comparing subsequent runs with "glean -c" and concentrating on the differences, instead of trying to reanalyze each run in toto from scratch. There are too many test results for that to be viable. Allen ------------------------------------------------------------------------------ Enter the BlackBerry Developer Challenge This is your chance to win up to $100,000 in prizes! For a limited time, vendors submitting new applications to BlackBerry App World(TM) will have the opportunity to enter the BlackBerry Developer Challenge. See full prize details at: http://p.sf.net/sfu/Challenge _______________________________________________ Mesa3d-dev mailing list Mesa3d-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
