All subtests in the scissor.c conformance test are failing. Since this test used to pass, I spent some quality with with git-bisect. I tracked the failure down to the following commit. commit cb433d91c6e198b7c77f747f1a38803532bc9be9 Author: Eric Anholt <eric@anholt.net> Date: Sat Dec 6 14:21:12 2008 -0800 intel: Fix glBitmap blit acceleration for FBO destinations. Bug #18914. Fixes fbo_firecube hang due to drawing outside the FBO bounds. Thanks to Pierre Willenbrock for debugging the issue.
This same commit is also responsible for the failures in windowpos.c and bitmap-pos.c. I looked at the code a bit. The only thing that looked like it could leak into non-bitmap rendering is the call to intel_get_cliprects. That function can set some context state in some cases. The comment says "FBO or DRI2 rendering, which can just use the fb's size," but I wonder if this accidentally leaking into other cases?
Initial patch series that doesn't fix it quite yet posted to mesa3d-dev.
commit 1db63713046efd49071aae448105df8a80e1e83d Author: Eric Anholt <eric@anholt.net> Date: Thu Dec 18 16:00:09 2008 -0800 intel: Fix glBitmap top/bottom clipping. Bug #19139. (Actually needs the previous patches in the series as well)
Mass version move, cvs -> git
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.