Thanks. I'm not sure what all the segfaults you list at the top are
about. But the difference in WindowPopup between now and in January that
might be causing a problem is (I think) that some single buffer drawing
routines were corrected and some drawing turning off/on was done for long
operations. But the problem you observe seems to be a chicken and egg
type problem, whereby some drawing request is happening before the window
is ready, and I'm not quite why those changes would cause that kind of
problem.
Is anyone else having this problem in GL mode?
Wayne
On Wed, 9 Apr 2008, Tolga Helmbrecht wrote:
> Hello,
>
> GL display causes some problems in analysis:
> Linking against 64 bit libs on fc8 always segfaults when trying to open a GL
> project, it runs with the 32 bit libs.
> Linking against 64 bit libs on ubuntu 7.10 works.
> It was not possible to get analysis to work with mesa.
> Compiling the old mesa6 on ubuntu, the included (and also the system) glxgears
> runs in the LD_LIBRARY_PATH environment, but analysis always just segfaults.
> Also tried with intel i810 chipset graphics which use mesa on fc6 and fc8.
> Analysis just segfaults.
>
> In the two cases where GL display works, the latest update of
> ccpnmr1.0/python/ccpnmr/analysis/WindowPopup.py
> causes analysis to crash with this rather generic message:
>
> GL_VENDOR = NVIDIA Corporation
> GL_RENDERER = GeForce 8800 GT/PCI/SSE2
> GL_VERSION = 2.1.2 NVIDIA 169.12
> X Error of failed request: GLXBadContext
> Major opcode of failed request: 143 (GLX)
> Minor opcode of failed request: 5 (X_GLXMakeCurrent)
> Serial number of failed request: 20024
> Current serial number in output stream: 20024
>
> It is possible to temporarlily work around this crash by scrolling in the
> windows while the project is loading. GL display works then, but after some
> minutes without scrolling the crash occours.
>
> Reverting back the current ccpnmr1.0/python/ccpnmr/analysis/WindowPopup.py to
> a version from january fixes this for now.
>
> Tolga
>
|