| Summary: | No sane visuals are reported, "couldn't find RGB GLX visual" | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Guillaume Melquiond <guillaume.melquiond> |
| Component: | GLX | Assignee: | mesa-dev |
| Status: | RESOLVED DUPLICATE | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | git | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Guillaume Melquiond
2008-07-15 01:46:25 UTC
Actually, it's a bit worse than what I thought. The very first run of the old glxinfo works fine, it is only the later runs that are going wrong. So it seems there is some persistent data somewhere that are broken by mesa git. Note that applications using the shared libraries of old mesa are still running fine, even after the failure of mesa git. It is the applications that are using the libraries of mesa git that work only once (except for the strange case of glxinfo from git). To summarize, when run in the following order, I get 1. glxinfo "old" with mesa "old" OK 2. glxinfo "old" with mesa "git" OK 3. glxinfo "old" with mesa "git" FAIL 4. glxinfo "git" with mesa "git" OK 5. glxgear "git" with mesa "git" FAIL 6. glxgear "git" with mesa "old" OK Sorry for the spam, I'm not even sure it is related. But I was trying to understand why the glxinfo from git was still working properly, so I looked through the recent patches to the tools. As a consequence, I noticed some strange code at line 433:
if (!visinfo)
configs = glXChooseFBConfig(dpy, scrnum, fbAttribSingle, &nConfigs);
if (!visinfo)
configs = glXChooseFBConfig(dpy, scrnum, fbAttribDouble, &nConfigs);
It seems to me that the first assignment is some dead code. Moreover, since it is followed by Xfree(configs), I suppose this is also a memory leak.
(In reply to comment #1) > The very first run of the old glxinfo works fine, it is only the later runs > that are going wrong. Sounds like bug 13358 (not sure why it was resolved as fixed). Is this on x86_64 as well? Yes, it is on x86-64 as well. And it does indeed sound like the same bug, especially since the reporter said that "if I install mesa 7.0 I can get opengl working again without restarting the X server", which I'm experiencing as well. It may be worth reopening the first report and marking this one as a duplicate then. *** This bug has been marked as a duplicate of bug 13358 *** 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.