Version-Release number of selected component: wireshark-1.6.10-1.fc17 Additional info: libreport version: 2.0.15 abrt_version: 2.0.14 backtrace_rating: 4 cmdline: tshark -l -V port 1935 crash_function: addMacro kernel: 3.5.5-1.fc17.i686.PAE truncated backtrace: :Thread no. 1 (10 frames) : #0 addMacro at data.c:3869 : #1 rpmPlatform at rpmrc.c:703 : #2 defaultMachine at rpmrc.c:1078 : #3 rpmSetMachine at rpmrc.c:1460 : #4 rpmRebuildTargetVars at rpmrc.c:1557 : #5 rpmReadConfigFiles at rpmrc.c:1776 : #6 initModule at rpmmodule.c:280 : #7 init_rpm at rpmmodule.c:264 : #8 _PyImport_LoadDynamicModule at /usr/src/debug/Python-2.7.3/Python/importdl.c:53 : #9 load_module at /usr/src/debug/Python-2.7.3/Python/import.c:1831
Created attachment 622769 [details] File: core_backtrace
Created attachment 622770 [details] File: environ
Created attachment 622771 [details] File: backtrace
Created attachment 622772 [details] File: limits
Created attachment 622773 [details] File: cgroup
Created attachment 622774 [details] File: maps
Created attachment 622775 [details] File: dso_list
Created attachment 622776 [details] File: proc_pid_status
Created attachment 622777 [details] File: var_log_messages
Created attachment 622778 [details] File: open_fds
This used to work. Some package update from updates-testing must have caused a regression - but I haven't been able to find a downgrade combination that works. I thought it could have been caused by being on USB wireless ... but I see the same on wired. A plain wireshark also crashes.
I also have this on x86_64.
Wireshark has just stopped working for me as well, so it seems whatever broke it has hit the updates repo in the last day or so (I definitely used it successfully on 22nd after doing a yum update that day).
The bad package seems to be abrt-addon-python. If I delete that (and the dependent abrt-desktop package) then wireshark works again.
(In reply to comment #14) > The bad package seems to be abrt-addon-python. If I delete that (and the > dependent abrt-desktop package) then wireshark works again. Confirmed. Reassigning. abrt-addon-python-2.0.16-1.fc17.i686 I would expect that abrt 1. Didn't try to do anything for embedded Python interpreters. But of course, if you in general can handle it in a useful way then it is great. 2. Didn't try to do anything even slightly complex in the python pth hooks. Especially not loading other modules such as rpm. Modules should only be loaded on demand, and as much as possible should be done in another process. Somewhat related: time python -c 'print' with abrt-addon-python: real 0m0.157s without: real 0m0.077s and something I care more about: time hg version with: real 0m0.220s without: real 0m0.132s It seems like this is some secret work from the git cabal to make Mercurial seem sluggish ;-)
*** This bug has been marked as a duplicate of bug 864324 ***