GNOME Bugzilla – Bug 112548
frequent crashes & lockups
Last modified: 2004-12-22 21:47:04 UTC
Frequent crashes and lockups at unpredictable times. Not sure if this is related to libgnet 2, which obviously is not supported. I'm using glibc 2.3.1. $ gcc -v Reading specs from /usr/lib/gcc-lib/i386-linux/3.2.3/specs Configured with: ../src/configure -v --enable-languages=c,c++,java,f77,objc,ada --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-gxx-include-dir=/usr/include/c++/3.2 --enable-shared --with-system-zlib --enable-nls --without-included-gettext --enable-__cxa_atexit --enable-clocale=gnu --enable-java-gc=boehm --enable-objc-gc i386-linux Thread model: posix gcc version 3.2.3 20030415 (Debian prerelease) Backtrace from one crash: (gdb) run Starting program: /usr/bin/pan [New Thread 16384 (LWP 14505)] [New Thread 32769 (LWP 14506)] [New Thread 16386 (LWP 14507)] [New Thread 32771 (LWP 14508)] [New Thread 49156 (LWP 14509)] [New Thread 65541 (LWP 14510)] [New Thread 81926 (LWP 14511)] [New Thread 98311 (LWP 14512)] [New Thread 114696 (LWP 14513)] [New Thread 131081 (LWP 14514)] [New Thread 147466 (LWP 14515)] [New Thread 163851 (LWP 14516)] [New Thread 180236 (LWP 14517)] [New Thread 196621 (LWP 14518)] [New Thread 213006 (LWP 14519)] [New Thread 229391 (LWP 14520)] Program received signal SIGSEGV, Segmentation fault. 0x40587d56 in mallopt () from /lib/libc.so.6 (gdb) bt
+ Trace 36676
Thread 16384 (LWP 14540)
additionally... the "unpredictable times" are generally (but not always) when pan has been sitting idle for a few minutes. This is rather unusual. I had to make three source changes to get pan to compile with libgnet 2.0.2. gnet_tcp_socket_get_iochannel becomes gnet_tcp_socket_get_io_channel in sockets.c line 142, and gnet_tcp_socket_gethostname becomes gnet_tcp_socket_get_host_name in smtp.c line 282 and dialog-newuser.c line 100. I also had to fake gnet-config. --- additionally additionally, it seems all these problems are indeed caused by an incompatibility with gnet 2. Recompiling with gnet 1.1.9 (like it's supposed to be done) seems to resolve the problem. Canceling bug report... BTW, you guys are awesome.
For future reference if anyone runs across the bug, the stack trace looks similar to the ones in bug 98308, bug 98630, and bug 115666. Each of these bug reports was on a Debian testing/unstable system.