Afflicted packages: mail-client/sylpheed-claws mail-client/evolution media-gfx/gimp As first noted in bug #135886, mail-client/sylpheed-claws suffers a bug where, after switching to a thread, control is never properly returned to the parent process. The common factor between these packages seems to be x11-libs/gtk+ or one of its dependencies, notably dev-libs/glib.
Created attachment 92696 [details] Strace output for media-gfx/gimp-2.2.12 # strace -f -o gimp.strace gimp
At over 1.6 megabytes, the strace output from mail-client/evolution-2.6.2-r1: http://dev.gentoo.org/~jer/evolution.strace
jeroen@elmer ~ $ gdb evolution GNU gdb 6.4 Copyright 2005 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "hppa2.0-unknown-linux-gnu"...Using host libthread_db library "/lib/libthread_db.so.1". (gdb) run Starting program: /usr/bin/evolution [Thread debugging using libthread_db enabled] [New Thread 16384 (LWP 15048)] (evolution-2.6:15048): GnomeUI-WARNING **: While connecting to session manager: Could not open network socket. /home/jeroen/.kde/share/config/gtkrc-2.0:22: error: unexpected identifier `nan', expected number (float) CalDAV Eplugin starting up ... [New Thread 32769 (LWP 15052)] [New Thread 16386 (LWP 15053)] [Thread 16386 (LWP 15053) exited] [New Thread 32770 (LWP 15054)] [New Thread 49155 (LWP 15055)] [New Thread 65540 (LWP 15056)] [Thread 65540 (LWP 15056) exited] [New Thread 81924 (LWP 15058)] [New Thread 98309 (LWP 15059)] [New Thread 114694 (LWP 15060)] Unknown RESP-CODE encountered: MYRIGHTS Program received signal SIGINT, Interrupt. [Switching to Thread 114694 (LWP 15060)] 0x41da7df4 in nanosleep () from /lib/libpthread.so.0 (gdb) backtrace full #0 0x41da7df4 in nanosleep () from /lib/libpthread.so.0 No symbol table info available. #1 0x41da7de0 in nanosleep () from /lib/libpthread.so.0 No symbol table info available. Previous frame identical to this frame (corrupt stack?)
Unable to reproduce on x86, glib/gtk however is notoriously bad at threading. I think this should go upstream but with a more complete backtrace of a lockup if possible and preferrably a reproducable testcase.
Created attachment 93193 [details] gdb backtrace of sylpheed-claws
I have now tested sylpheed-claws-2.0.0 in a chroot with sys-libs/glibc-2.3.4.20040808-r1 and there the problem does NOT occur. This lays the blame squarely on glibc/pthreads. I'll upgrade to sylpheed-claws-2.4.0 and see if that works as well so I can at least resolve bug #135886.
Created attachment 94049 [details] sylpheed-claws.gdb5-WORKSFORME This one didn't experience any problems.
Block mail-client/claws-mail and mail-client/sylpheed stabilisation bugs: they won't properly work until we have glibc-2.5. Relieve gnome@gentoo.org of bug duty.
these versions of glibc are no longer supported ... if glibc-2.5 shows issues, open a new bug report
