GNOME Bugzilla – Bug 161843
gnome-session-save crashes gnome-terminal
Last modified: 2008-05-29 19:55:44 UTC
Distribution: Fedora Core release 3 (Heidelberg) Package: gnome-terminal Severity: normal Version: GNOME2.8.0 unspecified Gnome-Distributor: Red Hat, Inc Synopsis: gnome-session-save crashes gnome-terminal Bugzilla-Product: gnome-terminal Bugzilla-Component: general Bugzilla-Version: unspecified BugBuddy-GnomeVersion: 2.0 (2.8.0) Description: Description of the crash: I type "gnome-session-save" when I have 4 terminal windows, from 3 different profiles, open. Gnome terminal crashes. Same effect when I log out and click "save session". Steps to reproduce the crash: 1. Open 4 gnome-terminal windows in 3 different profiles. (Not sure how necessary the profiles are, but it's the way I do it). I usually have 2 open & 2 iconified when I do this. 2. Run "gnome-session-save". 3. Expected Results: gnome-terminal crashes. How often does this happen? I'll try a bunch more, in last version it was every time, but that may have changed. Additional Information: This is similar to a bug I had in fedora core 2. Debugging Information: Backtrace was generated from '/usr/bin/gnome-terminal' (no debugging symbols found)...Using host libthread_db library "/lib/tls/libthread_db.so.1". (no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...[Thread debugging using libthread_db enabled] [New Thread -151127840 (LWP 4170)] [New Thread 96275376 (LWP 4204)] [Thread debugging using libthread_db enabled] [New Thread -151127840 (LWP 4170)] [New Thread 96275376 (LWP 4204)] [Thread debugging using libthread_db enabled] [New Thread -151127840 (LWP 4170)] [New Thread 96275376 (LWP 4204)] (no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...0x005b77a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2
+ Trace 53728
Thread 1 (Thread -151127840 (LWP 4170))
------- Bug moved to this database by unknown@bugzilla.gnome.org 2004-12-20 19:17 ------- Unknown platform unknown. Setting to default platform "Other". Unknown milestone "unknown" in product "gnome-terminal". Setting to default milestone for this product, '---' The original reporter of this bug does not have an account here. Reassigning to the person who moved it here, unknown@bugzilla.gnome.org. Previous reporter was wms@igoweb.org. Setting to default status "UNCONFIRMED". Setting qa contact to the default for this product. This bug either had no qa contact or an invalid one.
*** Bug 146214 has been marked as a duplicate of this bug. ***
I'm the original submitter of this. If anybody has any questions or wants me to run any tests, just let me know. Has anybody been able to duplicate it yet? I'm worried that it's something screwy about my gnome preferences or system configuration.
Looks like something strange is going on here at least. terminal_app_edit_encodings() never calls g_print() in the current sources at least. Can you reproduce this using later releases? I can't reproduce this here with 2.11.x at least.
*** Bug 313732 has been marked as a duplicate of this bug. ***
Bug 313732 seems to show this bug more easily, and also on a newer version of gnome (2.10.0 instead of 2.8.0). I need to use my system for work, and it is risky enough to combine 2.11 gnome with older fedora-supplied packages that it just isn't practical for me to try to reproduce now...but maybe with the instructions on bug 313732 it would be easy for somebody else to test on 2.11?
For the record, I've just tried a couple of times to reproduce #313732 on a FC4 box, and wasn't able to.
I've tried without luck also. Guillaume, have you had a chance to try a later release of vte?
*** Bug 336897 has been marked as a duplicate of this bug. ***
Thanks Fabio for your comment in bug 336897 - I never knew before how to install debugging symbols! I installed debugging symbols for gnome-terminal, gtk2, pango, glibc-common, glibc, and gnome-libs. If there is another library that would be useful let me know, but I think I got the ones that are relevant in this bug. Then I crashed it again, and here is the stack trace it gave (if I do this again, are stack traces better at attachments or comments?): Backtrace was generated from '/usr/bin/gnome-terminal' Using host libthread_db library "/lib/libthread_db.so.1". `shared object read from target memory' has disappeared; keeping its symbols. [Thread debugging using libthread_db enabled] [New Thread -1208510800 (LWP 18212)] [New Thread -1219920992 (LWP 18215)] 0x00782402 in __kernel_vsyscall ()
+ Trace 67410
Thread 1 (Thread -1208510800 (LWP 18212))
You are welcome William! Stacktraces within comments are just fine, furthermore this stacktrace should be useful to determine the crash's cause, thanks.
William, can you reproduce this with a recent version of GNOME?
I am still running FC5 (with latest upgrades) and it happens all the time (every logout). That uses gnome 1.4.1.2. It has been happening for years. Several times I've been told to try the newest version, it is never fixed, I hope you understand if I'm not willing to spend a few hours upgrading my gnome (always tricky without conflicting with other RPMs!) just to try again. :-(
I have always found that gnome-terminal crashes when saving the session if there are too many terminal windows open. Here is a trace from gnome 2.16 from saving the session with 13 windows open: Using host libthread_db library "/lib64/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread 46912496407008 (LWP 4824)] [New Thread 1084229952 (LWP 4829)] 0x000000382700d96f in __libc_waitpid (pid=5337, stat_loc=0x7fff95fffadc, options=0) at gmarshal.c:614 614 {
+ Trace 84268
Thread 1 (Thread 46912496407008 (LWP 4824))
Created attachment 76196 [details] valgrind log of crash This is a valgrind log of gnome-terminal showing what happens when I try and save a sessions with lots of windows open - note the writes in save_yourself_callback which are beyond the end of of buffer which has been allocated on the stack with calloc. The stack is probably getting trashed as a result which will limit the information available in a normal backtrace.
*** Bug 378983 has been marked as a duplicate of this bug. ***
*** Bug 430243 has been marked as a duplicate of this bug. ***
*** Bug 415050 has been marked as a duplicate of this bug. ***
*** Bug 462743 has been marked as a duplicate of this bug. ***
*** This bug has been marked as a duplicate of 501635 ***