I'm having problems with applications crashing on RHEL5 because my home directory is shared with a Fedora 11 machine. The problem is somewhere in the cache management as if I remove ~/.fontconfig things work fine. Any time the cache is updated on the F11 machine things start crashing in RHEL again though. Sounds like this is the same issue: http://lists.freedesktop.org/archives/fontconfig/2009-June/003168.html Backtrace from vmware: in function Util_BacktraceWithFunc in object /usr/lib/vmware/lib/libvmwarebase.so.0/libvmwarebase.so.0 loaded at 00002ba2fc9fa000 in function (null) in object /usr/lib/vmware/lib/libvmwarebase.so.0/libvmwarebase.so.0 loaded at 00002ba2fc9fa000 in function (null) in object /lib64/libpthread.so.0 loaded at 0000003663400000 in function FcStrListCreate in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000 in function FcLangSetHasLang in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000 in function (null) in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000 in function (null) in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000 in function (null) in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000 in function FcFontSetSort in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000 in function FcFontSort in object /usr/lib64/libfontconfig.so.1 loaded at 0000003667400000
Another traceback, this time from PyGTK: Program received signal SIGSEGV, Segmentation fault. 0x005ea997 in FcStrListCreate () from /usr/lib/libfontconfig.so.1 (gdb) bt #0 0x005ea997 in FcStrListCreate () from /usr/lib/libfontconfig.so.1 #1 0x005e396d in FcLangSetHasLang () from /usr/lib/libfontconfig.so.1 #2 0x005e5b63 in ?? () from /usr/lib/libfontconfig.so.1 #3 0x005e565c in ?? () from /usr/lib/libfontconfig.so.1 #4 0x005e5960 in ?? () from /usr/lib/libfontconfig.so.1 #5 0x005e5d36 in FcFontSetSort () from /usr/lib/libfontconfig.so.1 #6 0x005e6268 in FcFontSort () from /usr/lib/libfontconfig.so.1 #7 0x00da4bee in ?? () from /usr/lib/libpangoft2-1.0.so.0 #8 0x00d65d30 in pango_font_map_load_fontset () from /usr/lib/libpango-1.0.so.0 #9 0x00da4fe5 in ?? () from /usr/lib/libpangoft2-1.0.so.0 #10 0x00d65db9 in pango_font_map_load_font () from /usr/lib/libpango-1.0.so.0 #11 0x00d642b9 in pango_context_load_font () from /usr/lib/libpango-1.0.so.0 #12 0x00d69f66 in pango_layout_line_get_extents () from /usr/lib/libpango-1.0.so.0 #13 0x00d6a26f in ?? () from /usr/lib/libpango-1.0.so.0 #14 0x00d6cdc5 in ?? () from /usr/lib/libpango-1.0.so.0 #15 0x012b3520 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #16 0x0073e7d9 in g_cclosure_marshal_VOID__BOXED () from /lib/libgobject-2.0.so.0 #17 0x00730639 in ?? () from /lib/libgobject-2.0.so.0 #18 0x00731f81 in g_closure_invoke () from /lib/libgobject-2.0.so.0 #19 0x00743242 in ?? () from /lib/libgobject-2.0.so.0 #20 0x007442e7 in g_signal_emit_valist () from /lib/libgobject-2.0.so.0 #21 0x00745d4e in g_signal_emit_by_name () from /lib/libgobject-2.0.so.0 #22 0x0132b916 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #23 0x0132bb67 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #24 0x013e0b2c in gtk_widget_size_request () from /usr/lib/libgtk-x11-2.0.so.0 #25 0x013d7430 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #26 0x0073e7d9 in g_cclosure_marshal_VOID__BOXED () from /lib/libgobject-2.0.so.0 #27 0x00730639 in ?? () from /lib/libgobject-2.0.so.0 #28 0x00731f81 in g_closure_invoke () from /lib/libgobject-2.0.so.0 #29 0x00743242 in ?? () from /lib/libgobject-2.0.so.0 #30 0x007442e7 in g_signal_emit_valist () from /lib/libgobject-2.0.so.0 #31 0x00745d4e in g_signal_emit_by_name () from /lib/libgobject-2.0.so.0 #32 0x0132b916 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #33 0x0132bb67 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #34 0x013e0b2c in gtk_widget_size_request () from /usr/lib/libgtk-x11-2.0.so.0 #35 0x013e89a0 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #36 0x0073e7d9 in g_cclosure_marshal_VOID__BOXED () from /lib/libgobject-2.0.so.0 #37 0x00730639 in ?? () from /lib/libgobject-2.0.so.0 #38 0x00731e98 in g_closure_invoke () from /lib/libgobject-2.0.so.0 #39 0x00743242 in ?? () from /lib/libgobject-2.0.so.0 #40 0x007442e7 in g_signal_emit_valist () from /lib/libgobject-2.0.so.0 #41 0x00745d4e in g_signal_emit_by_name () from /lib/libgobject-2.0.so.0 #42 0x0132b916 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #43 0x0132bb67 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #44 0x013e0b2c in gtk_widget_size_request () from /usr/lib/libgtk-x11-2.0.so.0 #45 0x013e8d20 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #46 0x013f1fd1 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #47 0x0073f169 in g_cclosure_marshal_VOID__VOID () from /lib/libgobject-2.0.so.0 #48 0x00730639 in ?? () from /lib/libgobject-2.0.so.0 #49 0x00731e98 in g_closure_invoke () from /lib/libgobject-2.0.so.0 #50 0x00743242 in ?? () from /lib/libgobject-2.0.so.0 #51 0x007442e7 in g_signal_emit_valist () from /lib/libgobject-2.0.so.0 #52 0x007444a9 in g_signal_emit () from /lib/libgobject-2.0.so.0 #53 0x013e1b18 in gtk_widget_show () from /usr/lib/libgtk-x11-2.0.so.0 #54 0x010852a1 in ?? () from /usr/lib/python2.4/site-packages/gtk-2.0/gtk/_gtk.so #55 0x02cb9f5a in PyEval_EvalFrame () from /usr/lib/libpython2.4.so.1.0 #56 0x02cbb486 in PyEval_EvalCodeEx () from /usr/lib/libpython2.4.so.1.0 #57 0x02cbb513 in PyEval_EvalCode () from /usr/lib/libpython2.4.so.1.0 #58 0x02cd8348 in ?? () from /usr/lib/libpython2.4.so.1.0 #59 0x02cd9a58 in PyRun_SimpleFileExFlags () from /usr/lib/libpython2.4.so.1.0 #60 0x02cda13a in PyRun_AnyFileExFlags () from /usr/lib/libpython2.4.so.1.0 #61 0x02ce0b45 in Py_Main () from /usr/lib/libpython2.4.so.1.0 #62 0x080485b2 in main ()
I think we have fixed these issues in more recent versions of fontconfig. But for RHEL5, my answer simply is: "Don't do that then". Can't fix it in RHEL without causing more headache for normal users.
I'm sorry, but I don't think that's good enough. Not being able to share home directories is a serious deficiency in RHEL. Having a performance penalty for doing it is one thing, but crashes is completely unacceptable. Fontconfig should have better input validation than that.
Ok, some more debugging and I understand the crash now. Now have to see how I can fix it. Lemme see...
Ok, I've pushed the changes to master. Will make a 2.8.0 release and then push in f11 and f12. Should take a few more days. Note that I had to bump the cache version to fix this as the 2.6 and 2.7 versions where incompatible (unintended) and there was no way to fix that without a bump at this point.
A version bump is completely acceptable from my point of view. Am I interpreting you correctly that you haven't changed anything in RHEL's fontconfig, only made sure that Fedora's (and upstream's) no longer creates files that makes it crash?
(a solution that's fine with me, in case that isn't clear :))
fontconfig-2.8.0-1.fc11 has been submitted as an update for Fedora 11. http://admin.fedoraproject.org/updates/fontconfig-2.8.0-1.fc11
Yes, no change in RHEL. Pushed testing updates for F11 and F12 now. Will hit stable in a few days.
fontconfig-2.8.0-1.fc11 has been pushed to the Fedora 11 stable repository. If problems still persist, please make note of it in this bug report.
I suppose this bug can be closed now?
Hi Pierre, you are right, this can be closed now. I'm closing this with resolution NOTABUG since it was fixed in Fedora 11 and report is for Red Hat Enterprise Linux 5. Regards Marek