Bug 1381877

Summary: [abrt] trojita: QXcbConnection::processXcbEvents(): trojita killed by SIGSEGV
Product: [Fedora] Fedora Reporter: Karel Volný <kvolny>
Component: trojitaAssignee: Raphael Groner <projects.rg>
Status: CLOSED CANTFIX QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 25CC: kde-sig, projects.rg
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Unspecified   
URL: https://retrace.fedoraproject.org/faf/reports/bthash/cfebf4d709e8914da7ef5c25b5b52311dcae9c32
Whiteboard: abrt_hash:cf1a054f4c225c9d4e659fcdcb3017ef39b0cc42;
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-11-05 22:49:17 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
File: backtrace
none
File: cgroup
none
File: core_backtrace
none
File: dso_list
none
File: environ
none
File: exploitable
none
File: limits
none
File: maps
none
File: mountinfo
none
File: namespaces
none
File: open_fds
none
File: proc_pid_status
none
File: var_log_messages none

Description Karel Volný 2016-10-05 09:05:38 UTC 
Description of problem:
It crashed after X crash ... I guess the disconnect should be handled more gracefully.

Version-Release number of selected component:
trojita-0.7-3.fc25

Additional info:
reporter:       libreport-2.8.0
backtrace_rating: 4
cmdline:        /usr/bin/trojita
crash_function: QXcbConnection::processXcbEvents
executable:     /usr/bin/trojita
global_pid:     32010
kernel:         4.8.0-0.rc7.git0.1.fc25.x86_64
pkg_fingerprint: 4089 D8F2 FDB1 9C98
pkg_vendor:     Fedora Project
runlevel:       N 5
type:           CCpp
uid:            1000

Truncated backtrace:
Thread no. 1 (10 frames)
 #15 QXcbConnection::processXcbEvents at qxcbconnection.cpp:1699
 #16 QObject::event at kernel/qobject.cpp:1263
 #17 QApplicationPrivate::notify_helper at kernel/qapplication.cpp:3799
 #18 QApplication::notify at kernel/qapplication.cpp:3556
 #19 QCoreApplication::notifyInternal2 at kernel/qcoreapplication.cpp:988
 #20 QCoreApplication::sendEvent at kernel/qcoreapplication.h:231
 #21 QCoreApplicationPrivate::sendPostedEvents at kernel/qcoreapplication.cpp:1649
 #22 QCoreApplication::sendPostedEvents at kernel/qcoreapplication.cpp:1503
 #23 postEventSourceDispatch at kernel/qeventdispatcher_glib.cpp:276
 #27 g_main_context_iteration at gmain.c:3988
Comment 1 Karel Volný 2016-10-05 09:05:44 UTC 
Created attachment 1207501 [details]
File: backtrace
Comment 2 Karel Volný 2016-10-05 09:05:45 UTC 
Created attachment 1207502 [details]
File: cgroup
Comment 3 Karel Volný 2016-10-05 09:05:47 UTC 
Created attachment 1207503 [details]
File: core_backtrace
Comment 4 Karel Volný 2016-10-05 09:05:49 UTC 
Created attachment 1207504 [details]
File: dso_list
Comment 5 Karel Volný 2016-10-05 09:05:50 UTC 
Created attachment 1207505 [details]
File: environ
Comment 6 Karel Volný 2016-10-05 09:05:51 UTC 
Created attachment 1207506 [details]
File: exploitable
Comment 7 Karel Volný 2016-10-05 09:05:52 UTC 
Created attachment 1207507 [details]
File: limits
Comment 8 Karel Volný 2016-10-05 09:05:54 UTC 
Created attachment 1207508 [details]
File: maps
Comment 9 Karel Volný 2016-10-05 09:05:56 UTC 
Created attachment 1207509 [details]
File: mountinfo
Comment 10 Karel Volný 2016-10-05 09:05:57 UTC 
Created attachment 1207510 [details]
File: namespaces
Comment 11 Karel Volný 2016-10-05 09:05:58 UTC 
Created attachment 1207511 [details]
File: open_fds
Comment 12 Karel Volný 2016-10-05 09:05:59 UTC 
Created attachment 1207512 [details]
File: proc_pid_status
Comment 13 Karel Volný 2016-10-05 09:06:01 UTC 
Created attachment 1207513 [details]
File: var_log_messages
Comment 14 Raphael Groner 2016-10-05 19:05:22 UTC 
Well, X is the parent process for all forked application processes. I doubt we can do anything better for X crashers.
Comment 15 Karel Volný 2016-12-14 12:41:57 UTC 
(In reply to Raphael Groner from comment #14)
> Well, X is the parent process for all forked application processes.

ahem, it doesn't seem to be this way on my system:

[kvolny@kvolny ~]$ ps -o ppid -C trojita
 PPID
    1

... and pid eins is systemd

using tree view like `ps xfa`, Xorg doesn't own anything; there is lxqt-session under sddm => sddm-helper and a few processes under that, but most X applications are detached (thus collected by pid 1)

> I doubt we can do anything better for X crashers.

I believe there should be a way, as after X crash, I don't get fifty reports for all the applications, I get only a few for some of them (that are buggy) ... I also often run kmail via xpra, which tends to lock up with 'wrong' versions of python, and cutting its branch by forcefully killing xvfb doesn't lead to kmail crash (nor crash of any dependent akonadi and various other processes)

I know it can be hard to debug as the Trojitá crash does not happen on each X crash, and I understand there is not enough manpower[*] to deal wih each and every corner case, but closing for capacity reasons is quite different from saying it can't be fixed because of the nature of things ...

[*] look at me, I'm dealing with my bugzilla queue after a few months as part of "Christmas cleanup", when I had to ask my manager for a pause from the usual daily work :-(