Bug 106124

Summary: segfault in gtk_im_context_xim_new on x86_64 in CJK locales
Product: Red Hat Enterprise Linux 3 Reporter: Mike McLean <mikem>
Component: gtk2Assignee: Owen Taylor <otaylor>
Status: CLOSED RAWHIDE QA Contact: David Lawrence <dkl>
Severity: high Docs Contact:
Priority: high    
Version: 3.0CC: bfox
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2003-10-15 19:21:43 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 104704    
Attachments:
Description Flags
full backtrace
none
full backtrace (gnome-terminal) none

Description Mike McLean 2003-10-02 23:03:48 UTC 
* Taroon-re0929.RC1.0 ES
* LANG=ja_JP.eucJP
gtk2-2.2.4-3.0.i386
gtk2-2.2.4-3.0.x86_64
glibc-2.3.2-95.2.i686
glibc-2.3.2-95.2.x86_64
kernel-2.4.21-3.EL.x86_64
python-2.2.3-5.x86_64
firstboot-1.1.19-1.noarch

After a japanese language install, firstboot dies as soon as it hits the
date/time configuration screen.  The debugger suggests gtk is to blame.

Will attach full backtrace.

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 182894217376 (LWP 3648)]
0x0000002a9e64fbde in gtk_im_context_xim_new () from
/usr/lib64/gtk-2.0/2.2.0/immodules/im-xim.so
(gdb) bt
#0  0x0000002a9e64fbde in gtk_im_context_xim_new () from
/usr/lib64/gtk-2.0/2.2.0/immodules/im-xim.so
#1  0x0000002a96e04d30 in gtk_im_context_set_cursor_location () from
/usr/lib64/libgtk-x11-2.0.so.0
#2  0x0000002a96e06bf8 in gtk_im_multicontext_new () from
/usr/lib64/libgtk-x11-2.0.so.0
#3  0x0000002a96e04d30 in gtk_im_context_set_cursor_location () from
/usr/lib64/libgtk-x11-2.0.so.0
#4  0x0000002a96de9281 in gtk_entry_get_type () from
Comment 1 Mike McLean 2003-10-02 23:04:25 UTC 
Created attachment 94901 [details]
full backtrace
Comment 2 Mike McLean 2003-10-03 00:28:22 UTC 
Also happens in ko_KR, zh_TW, zh_TW.Big5, and zh_CN.GB18030

For some reason it does not seem to happen with LANG=zh_CN
Comment 3 Mike McLean 2003-10-03 01:36:25 UTC 
Getting something similar from with gnome-terminal:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 182894217280 (LWP 4541)]
gtk_im_context_xim_focus_in (context=0x30) at gtkimcontextxim.c:745
745           context_xim->has_focus = TRUE;
(gdb) bt
#0  gtk_im_context_xim_focus_in (context=0x30) at gtkimcontextxim.c:745
#1  0x0000002a9705db45 in gtk_im_context_focus_in (context=0x89d1a0)
    at gtkimcontext.c:337
#2  0x0000002a9705fb59 in gtk_im_multicontext_focus_in (context=0x0)
    at gtkimmulticontext.c:339
#3  0x0000002a9705db45 in gtk_im_context_focus_in (context=0x72dca0)
    at gtkimcontext.c:337
#4  0x0000002a9569abbe in vte_terminal_focus_in (widget=0x6810a0,
    event=0x8bb480) at vte.c:10002
Comment 4 Mike McLean 2003-10-03 01:37:05 UTC 
Created attachment 94906 [details]
full backtrace (gnome-terminal)
Comment 5 Bill Nottingham 2003-10-03 14:34:31 UTC 
On a lark, what happens if you install the XFree86-libs & other XFree86 bits
from dist-3.0E-HEAD?
Comment 6 Mike McLean 2003-10-03 16:04:38 UTC 
hmm, gnome-terminal runs fine if you run it within a gnome-session, but blows up
if you try it under X+twm.
Comment 7 Mike McLean 2003-10-03 16:12:16 UTC 
hmm, the firstboot bug does not appear if you run it under a gnome-session...
Comment 8 Mike McLean 2003-10-03 16:30:44 UTC 
*** Bug 106133 has been marked as a duplicate of this bug. ***
Comment 9 Mike McLean 2003-10-03 16:33:12 UTC 
export XMODIFIERS="@im=kinput2" gets gdm, firstboot, and gnome-terminal to run
in the ja_JP.eucJP locale.
Comment 10 Bill Nottingham 2003-10-03 19:10:47 UTC 
I cannot reproduce this failure on ia64.
Comment 11 Mike McLean 2003-10-03 20:18:07 UTC 
does not happen s390x
Comment 12 Mike McLean 2003-10-06 16:54:45 UTC 
seems fixed in Taroon-re1004.RC1.0 (gtk2-2.2.4-4.0)
Comment 13 Owen Taylor 2003-10-06 17:02:46 UTC 
Yeah, I stilil need to rebuild in Cambridge, but other than that, it should
be fixed.
Comment 14 Owen Taylor 2003-10-15 19:21:43 UTC 
* Fri Oct  3 2003 Owen Taylor <otaylor> 2.2.4-4.0
- Fix 64-bit problem in gtkimcontextxim.c (#106124)