I'm having curious problems with XtVaGetValues (and also by extension XtVaCreatePopupShell) on x86_64, it keeps core dumping. Here's a stacktrace: (gdb) where#0 0x00000037c3f48c43 in _XtCountVaList () from /usr/X11R6/lib64/libXt.so.6#1 0x00000037c3f48746 in XtVaGetValues () from /usr/X11R6/lib64/libXt.so.6#2 0x0000000000400dd4 in main (argc=1, argv=0x7fbffff768) at motifbug.cxx:26 (gdb) packages come from Fedora Core2: [mwsadm@fims64 hdfloader]$ rpm -qa "xorg*" xorg-x11-tools-6.7.0-2 xorg-x11-xauth-6.7.0-2 xorg-x11-75dpi-fonts-6.7.0-2 xorg-x11-Mesa-libGL-6.7.0-2 xorg-x11-font-utils-6.7.0-2 xorg-x11-base-fonts-6.7.0-2 xorg-x11-libs-6.7.0-2 xorg-x11-Mesa-libGL-6.7.0-2 xorg-x11-xfs-6.7.0-2 xorg-x11-6.7.0-2 xorg-x11-libs-data-6.7.0-2 xorg-x11-libs-6.7.0-2 xorg-x11-100dpi-fonts-6.7.0-2 xorg-x11-Mesa-libGLU-6.7.0-2 xorg-x11-twm-6.7.0-2 xorg-x11-devel-6.7.0-2 xorg-x11-Mesa-libGLU-6.7.0-2 Attached is the code to reproduce the bug in a simple pure-X/Motif test code, compiled via: g++ -g -o motifbug motifbug.cxx -L /usr/X11R6/lib64 -lXm -lXt This generates an X warning. Whatever. Add -DDUMP to the cmd line and it will coredump on the XtVaGetValues function. What gives? #ifdef'ing a function that comes *after* the location of the core matters? I know X is asynchronous but that seems to indicate a compiler problem??? I don't rule out user error, like improper compiler flags, but the example is really simple, so I don't see how.
Created attachment 1206 [details] code to reproduce the core dump compile via: g++ -g -o motifbug motifbug.cxx -L /usr/X11R6/lib64 -lXm -lXt for core: g++ -DDUMP -g -o motifbug motifbug.cxx -L /usr/X11R6/lib64 -lXm -lXt
Lest one thinks this is related to g++, I C-ified the example and the core trace was identical. I've tried getting better debug, but the Fedora x86_64 debug package list doesn't include any of the xorg packages.
Egbert: Can you reproduce the crash on SuSE 9.2/x86_64 ?
Any progress on this bug?
Carl Nygard wrote: > Any progress on this bug? Well, we need someone with Linux x86/64 who can reproduce the problem...
Is there somewhere I can get debuginfo rpms for x86_64? For FC3? (or even FC2)
The code used 0 to terminate the varargs list, which doesn't fly well in x86_64 land. switching to NULL or (void*)0 works fine.
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.