First, note that until last summer, I could open as many pictures as I wanted. Since a few months, trying to open a second pic when one is showinf on screen makes it segfault. I get segfault on the 2nd try. But, from time to time, the fist window may also crash (this is hard to reproduce).
Created attachment 137924 [details] full strace when trying to open a second picture, called /tmp/PowerPc 002.jpg, when an other window is already showing /tmp/PowerPc 001.jpg
Created attachment 137926 [details] /tmp/emerge--info [ebuild R ] media-gfx/eog-2.20.2 USE="dbus jpeg lcms python -debug" 0 kB
Happens with media-gfx/eog-2.20 media-gfx/eog-2.21 media-gfx/eog-2.22 . media-gfx/eog-2.18.2 works fine for me.
No such problem here. Could you attach a backtrace ? (btw nosplit is redundant when you have splitdebug)
What do you call backtrace ?
read this very carefully : http://www.gentoo.org/proj/en/qa/backtraces.xml Thanks
same with 2.20.3 I have readed this several times in the past, and never could do it. do you _really_ need BT for this bug ?
Yes we need backtraces. I can't reproduce this bug on my system, so we're going to need something, anything, to get a clue where this comes from. Thanks
Created attachment 139309 [details] /tmp/emerge--info See my actual emerge --info ... I remind that the bug happends when I run two simult eog; so, I need to BT both of them. I course, I run first the 1, and then the 2, in seperate consoles ... *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** dhp@moon_gen_2:/tmp$ gdb /usr/bin/eog GNU gdb 6.7.1 Copyright (C) 2007 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "i686-pc-linux-gnu"... Using host libthread_db library "/lib/libthread_db.so.1". (gdb) set args /tmp/holiday07_5.gif (gdb) set logging file backtrace_1.log (gdb) set logging on Copying output to backtrace_1.log. (gdb) run Starting program: /usr/bin/eog /tmp/holiday07_5.gif [Thread debugging using libthread_db enabled] [New Thread 0xb6d016c0 (LWP 5632)] [New Thread 0xb6c7eb90 (LWP 5636)] [New Thread 0xb5919b90 (LWP 5637)] Program received signal SIGABRT, Aborted. [Switching to Thread 0xb6d016c0 (LWP 5632)] 0xffffe410 in __kernel_vsyscall () (gdb) kill Kill the program being debugged? (y or n) y (gdb) quit dhp@moon_gen_2:/tmp$ *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** dhp@moon_gen_2:/tmp$ gdb /usr/bin/eog GNU gdb 6.7.1 Copyright (C) 2007 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "i686-pc-linux-gnu"... Using host libthread_db library "/lib/libthread_db.so.1". (gdb) set args /tmp/holiday07_5.gif (gdb) set logging file backtrace_2.log (gdb) set logging on Copying output to backtrace_2.log. (gdb) run Starting program: /usr/bin/eog /tmp/holiday07_5.gif [Thread debugging using libthread_db enabled] [New Thread 0xb6d2d6c0 (LWP 5638)] Program exited normally. (gdb) quit dhp@moon_gen_2:/tmp$ *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
Created attachment 139310 [details] /tmp/backtrace_1.log
Created attachment 139311 [details] /tmp/backtrace_2.log I cant see how those short BT can be more helpfull that straces. I did this after adding -ggdb ( nostrip splitdebug were already in FEATURES ), and unmasked media-gfx/eog-2.20.3 and updated it.
fixed in the following configuration: media-gfx/eog-2.20.4 USE="dbus jpeg lcms python -debug"
