Description of problem: I started ipe with "ipe ./file.pdf" Version-Release number of selected component: ipe-7.1.2-1.fc17 Additional info: libreport version: 2.0.14 abrt_version: 2.0.13 backtrace_rating: 4 cmdline: ipe ./oath_for_security_vpn.pdf crash_function: QMetaObject::cast kernel: 3.5.2-3.fc17.x86_64 truncated backtrace: :Thread no. 1 (10 frames) : #0 QMetaObject::cast at kernel/qmetaobject.cpp:274 : #1 qobject_cast<QToolButton*> at ../../src/corelib/kernel/qobject.h:380 : #2 update_toolbar_style at styles/qgtkstyle_p.cpp:232 : #8 g_object_notify_by_spec_internal at gobject.c:1133 : #9 g_object_notify at gobject.c:1175 : #10 g_cclosure_marshal_VOID__ENUMv at gmarshal.c:706 : #11 _g_closure_invoke_va at gclosure.c:840 : #14 gtk_toolbar_unset_style at gtktoolbar.c:2944 : #15 style_change_notify at gtktoolbar.c:2008 : #16 gtk_toolbar_screen_changed at gtktoolbar.c:2089
Created attachment 620634 [details] File: core_backtrace
Created attachment 620635 [details] File: limits
Created attachment 620636 [details] File: backtrace
Created attachment 620637 [details] File: maps
Created attachment 620638 [details] File: dso_list
Created attachment 620639 [details] File: build_ids
Created attachment 620640 [details] File: open_fds
Hi Paul, is the crash reproducable? Does it depend on the file you open with ipe?
Yes, it happens on any non-empty file, regardless of content. I'm also getting this error dialog box if I start ipe first and try to open a pdf file: Document '<full path>/file.pdf' could not be opened Error opening file: Resource temporarily unavailable. I was searching for an answer to that problem. I didn't find much, so I tried 'ipe ./file' and go this crash.
I cannot reproduce. Is not it a bug of the Qt-Gtk style? The backtrace seems to indicate that. Can you test another Qt program using the Gtk style?
I'm not sure. I didn't notice. I would be happy to test another program, but I'm not sure what. Can you suggest a program?
The bottom of the backtrace shows things about QGtkStyle, which is an orphaned package of CGAL for a long time. Have you compiled QGtkStyle yourself? What does `rpm -q qgtkstyle` display?
(In reply to comment #12) > The bottom of the backtrace shows things about QGtkStyle, which is an > orphaned package of CGAL for a long time. > > Have you compiled QGtkStyle yourself? What does `rpm -q qgtkstyle` display? Forget that! QGtkStyle is now in Qt itself since Qt-4.5. Anyway, I think that the bug is related to the interaction of Qt and you Gtk style. Probably not an ipe bug.
(In reply to comment #13) > Anyway, I think that the bug is related to the interaction of Qt and you Gtk > style. Probably not an ipe bug. My suggestion, to quickly check/reject that hypothesis: 1/ Launch qtconfig-qt4. 2/ Choose another mainstream style, like Oxygen. 3/ Launch ipe and verify if it still crashes using that other style (please use Abrt to post the backtrace with that other style). 4/ In qtconfig-qt4, switch back to your prefered QtGtk style.
Hi Laurent, Thanks, I appreciate your help. I found qtconfig-qt4 in package qt-config, installed it, and switched to the GTK+ style. The result is still a crash, but this time abrt didn't notice and doesn't show any new crash event for it. > ipe ./oath_for_security_vpn.pdf /usr/share/ipe/7.1.2/lua/actions.lua:929: attempt to index field 'doc' (a nil value) stack traceback: /usr/share/ipe/7.1.2/lua/actions.lua:929: in function 'action_fit_width' /usr/share/ipe/7.1.2/lua/actions.lua:938: in function 'action_fit_top' /usr/share/ipe/7.1.2/lua/model.lua:134: in function </usr/share/ipe/7.1.2/lua/model.lua:132> [C]: in function 'messageBox' /usr/share/ipe/7.1.2/lua/model.lua:243: in function 'warning' /usr/share/ipe/7.1.2/lua/model.lua:517: in function 'loadDocument' /usr/share/ipe/7.1.2/lua/model.lua:95: in function 'init' /usr/share/ipe/7.1.2/lua/model.lua:37: in function 'new' /usr/share/ipe/7.1.2/lua/main.lua:383: in main chunk [C]: in function 'require' [string "require "main""]:1: in main chunk Segmentation fault (core dumped) Also, the program still cannot read files I open via File > Open. I suppose I can get you a backtrace with gdb, but it will not be as nice as the ones ABRT makes. Do you know how to get ABRT to analyze a crash event it didn't notice?
In qtconfig-qt4, please choose native Qt style, like Oxygen, Cleanlooks, or Plastik, and not the special Qt-gtk styles. As for abrt, I must say I do not know how to launch it manually.
Oh, sorry, I didn't see Oxygen so I just picked one of the others. It looks like the same error as last time, but possibly different from the first. Fortunately, ABRT has detected this crash, though. So, do you want me to submit a new ABRT report or can somehow get ABRT to analyze the new crash and post the new backtrace in this bug? > ipe ./e-mail.txt /usr/share/ipe/7.1.2/lua/actions.lua:929: attempt to index field 'doc' (a nil value) stack traceback: /usr/share/ipe/7.1.2/lua/actions.lua:929: in function 'action_fit_width' /usr/share/ipe/7.1.2/lua/actions.lua:938: in function 'action_fit_top' /usr/share/ipe/7.1.2/lua/model.lua:134: in function </usr/share/ipe/7.1.2/lua/model.lua:132> [C]: in function 'messageBox' /usr/share/ipe/7.1.2/lua/model.lua:243: in function 'warning' /usr/share/ipe/7.1.2/lua/model.lua:517: in function 'loadDocument' /usr/share/ipe/7.1.2/lua/model.lua:95: in function 'init' /usr/share/ipe/7.1.2/lua/model.lua:37: in function 'new' /usr/share/ipe/7.1.2/lua/main.lua:383: in main chunk [C]: in function 'require' [string "require "main""]:1: in main chunk Segmentation fault (core dumped)
Yes, I would prefer to see the new backtrace without the Gtk stuff. I think you can signal an existing bug to abrt, to that it does not create a new bug. If you cannot, just create a new bug.
Okay, sorry, I guess I submitted a new bug. See bug 863604. *** This bug has been marked as a duplicate of bug 863604 ***