Broken document created.
Hey guys,
I was doing a teaching session on Inkscape for the Ubuntu User Days a few days ago. As with almost every show-and-tell, something went wrong. This was Inkscape 0.47 for Ubuntu 10.10 and the document that I created from scratch caused inkscape to crash.
What's interesting is that the image was saved by the crash protection save feature. And although the image can be opened by gnome and seen in firefox, opening it in Inkscape 0.47 causes it to crash again.
I'd like to know if the crash still exists in Inkscape 0.48, which unfortunately isn't available for Ubuntu yet. If it still exists, I'll report a bug and we'll get it nailed.
If you're interested in what I did to cause the problem, it's all recorded in the live session: http://livestre.am/AWfN Crash is at about 00:37:10
Best Regards, Martin Owens
Wasn't 0.48 in 10.10? On Jan 30, 2011 2:28 PM, "Martin Owens" <doctormo@...400...> wrote:
Hey guys,
I was doing a teaching session on Inkscape for the Ubuntu User Days a few days ago. As with almost every show-and-tell, something went wrong. This was Inkscape 0.47 for Ubuntu 10.10 and the document that I created from scratch caused inkscape to crash.
What's interesting is that the image was saved by the crash protection save feature. And although the image can be opened by gnome and seen in firefox, opening it in Inkscape 0.47 causes it to crash again.
I'd like to know if the crash still exists in Inkscape 0.48, which unfortunately isn't available for Ubuntu yet. If it still exists, I'll report a bug and we'll get it nailed.
If you're interested in what I did to cause the problem, it's all recorded in the live session: http://livestre.am/AWfN Crash is at about 00:37:10
Best Regards, Martin Owens
------------------------------------------------------------------------------
Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)! Finally, a world-class log management solution at an even better
price-free!
Download using promo code Free_Logger_4_Dev2Dev. Offer expires February 28th, so secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsight-sfd2d _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On 30/1/11 23:27, Martin Owens wrote:
If you're interested in what I did to cause the problem, it's all recorded in the live session: http://livestre.am/AWfN Crash is at about 00:37:10
Do you still have the file available? Could you attach it? I watched the related part of the recorded session 3 times but I can't tell what exactly happened. A backtrace and any console messages from Inkscape when it crashes trying to reopen the broken file might be helpful.
Also - since the crash happened when you started to draw a path with the calligraphy tool - did you use a tablet?
~suv
On Mon, 2011-01-31 at 01:45 +0100, ~suv wrote:
Do you still have the file available? Could you attach it? I watched the related part of the recorded session 3 times but I can't tell what exactly happened. A backtrace and any console messages from Inkscape when it crashes trying to reopen the broken file might be helpful.
Also - since the crash happened when you started to draw a path with the calligraphy tool - did you use a tablet?
I must be dense:
http://divajutta.com/doctormo/foo/Broken%20document% 201.2011_01_29_12_30_38.0.svg
I uploaded it but forgot to include the link... there must be a better way of doing that sort of thing.
I also meant 0.48 and 0.48.1 respectively.
Sorry for the confusion.
Martin,
On 31/1/11 06:18, Martin Owens wrote:
http://divajutta.com/doctormo/foo/Broken%20document%201.2011_01_29_12_30_38....
I also meant 0.48 and 0.48.1 respectively.
AFAICT fixed for 0.48.1:
Bug #511577 segfault in sp_shape_update_marker_view() https://bugs.launchpad.net/inkscape/+bug/511577
Crash reproduced with Inkscape 0.48.0 on OS X 10.5.8 (same console message and bt as in bug #511577)
Not reproduced with current trunk (0.48+devel r10019) and with a build from 0.48.x (r9752)
hth, ~suv
On 31/1/11 06:42, ~suv wrote:
On 31/1/11 06:18, Martin Owens wrote:
http://divajutta.com/doctormo/foo/Broken%20document%201.2011_01_29_12_30_38....
I also meant 0.48 and 0.48.1 respectively.
AFAICT fixed for 0.48.1:
Bug #511577 segfault in sp_shape_update_marker_view() https://bugs.launchpad.net/inkscape/+bug/511577
Crash reproduced with Inkscape 0.48.0 on OS X 10.5.8 (same console message and bt as in bug #511577)
Not reproduced with current trunk (0.48+devel r10019) and with a build from 0.48.x (r9752)
This path - last added to the document - causes the crash when loading the file:
<path style="stroke:#ffffff;fill:#ffffff;fill-opacity:1;stroke-opacity:1;stroke-width:10;stroke-miterlimit:4;stroke-dasharray:none;stroke-linejoin:round;stroke-linecap:round;opacity:0.58577406;color:#000000;fill-rule:nonzero;stroke-dashoffset:0;marker:none;visibility:visible;display:inline;overflow:visible;enable-background:accumulate;marker-end:url(#Arrow2Mend)" id="path6420" />
Removing the marker from the style attribute allows the file to open without crash even in 0.48.0.
In current trunk (r10019) and 0.48.x (r9752), I could trigger a different crash though, when trying to node-edit the path (which has no path data), and <Tab> through the nodes:
Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: KERN_INVALID_ADDRESS at address: 0x47746b7d 0x005f57df in Inkscape::UI::SelectableControlPoint::updateState () at selectable-control-point.h:82 82 x->updateState(); (gdb) bt #0 0x005f57df in Inkscape::UI::SelectableControlPoint::updateState () at selectable-control-point.h:82 #1 0x005f57df in Inkscape::UI::SelectableControlPoint::updateState () at selectable-control-point.h:82 #2 0x005f57df in Inkscape::UI::ControlPointSelection::insert (this=0x79c1b80, x=@0xbfffde0c) at ui/tool/control-point-selection.cpp:82 #3 0x005fc7d8 in Inkscape::UI::MultiPathManipulator::shiftSelection (this=0x79ab5b0, dir=1) at ui/tool/control-point-selection.cpp:82 #4 0x000fed2c in Inkscape::SelectionHelper::selectNext (dt=0x476bd20) at ui/tool/control-point-selection.cpp:82 #5 0x00333040 in sp_action_perform (action=0x4787700, data=0x0) at ui/tool/control-point-selection.cpp:82 #6 0x00111a53 in sp_shortcut_invoke (shortcut=65289, view=0x476bd20) at ui/tool/control-point-selection.cpp:82 #7 0x00071b11 in sp_event_context_private_root_handler (event_context=0x6d35430, event=0x6d99f10) at ui/tool/control-point-selection.cpp:82 #8 0x0060ca11 in (anonymous namespace)::ink_node_tool_root_handler (event_context=0x6d35430, event=0x6d99f10) at ui/tool/control-point-selection.cpp:82 #9 0x0006d77c in sp_event_context_virtual_root_handler (event_context=0xb49f10, event=0x6d99f10) at ui/tool/control-point-selection.cpp:82 #10 0x000396f7 in sp_desktop_root_handler (event=0x6d99f10, desktop=0x476bd20) at ui/tool/control-point-selection.cpp:82 #11 0x0033ed55 in sp_marshal_BOOLEAN__POINTER (closure=0x53e5b60, return_value=0xbfffe6ac, n_param_values=2, param_values=0x55a0530, invocation_hint=0xbfffe560, marshal_data=0x392c0) at ui/tool/control-point-selection.cpp:82 #12 0x032ec0a9 in g_closure_invoke () #13 0x032fe163 in signal_emit_unlocked_R () #14 0x032ff537 in g_signal_emit_valist () #15 0x01918741 in gtk_signal_emit () #16 0x00263f03 in emit_event (canvas=<value temporarily unavailable, due to optimizations>, event=0x6d9ba20) at ui/tool/control-point-selection.cpp:82 #17 0x0026a427 in sp_canvas_key (widget=0x41ff380, event=0x6d9ba20) at ui/tool/control-point-selection.cpp:82 #18 0x0178052b in _gtk_marshal_BOOLEAN__BOXED () #19 0x032ec0a9 in g_closure_invoke () #20 0x032fe2e8 in signal_emit_unlocked_R () #21 0x032ff537 in g_signal_emit_valist () #22 0x032ffaf9 in g_signal_emit () #23 0x018af8a6 in gtk_widget_event_internal () #24 0x018c5c99 in gtk_window_propagate_key_event () #25 0x018caefc in gtk_window_key_press_event () #26 0x0178052b in _gtk_marshal_BOOLEAN__BOXED () #27 0x032ec0a9 in g_closure_invoke () #28 0x032fe2e8 in signal_emit_unlocked_R () #29 0x032ff537 in g_signal_emit_valist () #30 0x032ffaf9 in g_signal_emit () #31 0x018af8a6 in gtk_widget_event_internal () #32 0x0177e765 in gtk_propagate_event () #33 0x0177ec9d in gtk_main_do_event () #34 0x02771b15 in gdk_event_dispatch () #35 0x033a6a9d in g_main_context_dispatch () #36 0x033aa59b in g_main_context_iterate () #37 0x033aa877 in g_main_loop_run () #38 0x0177dc71 in gtk_main () #39 0x01192d4b in Gtk::Main::run () #40 0x0000555c in Inkscape::UI::SelectableControlPoint::updateState () at selectable-control-point.h:986 #41 sp_main_gui (argc=1, argv=0xbffff394) at ui/tool/control-point-selection.cpp:82 #42 0x00004156 in start () (gdb)
~suv
On Mon, 2011-01-31 at 07:03 +0100, ~suv wrote:
Removing the marker from the style attribute allows the file to open without crash even in 0.48.0.
Thanks suv, as promised:
https://bugs.launchpad.net/inkscape/+bug/710637
p.s. I was using a pen.
Martin,
On 31/1/11 14:03, Martin Owens wrote:
On Mon, 2011-01-31 at 07:03 +0100, ~suv wrote:
Removing the marker from the style attribute allows the file to open without crash even in 0.48.0.
Thanks suv, as promised:
thx :)
p.s. I was using a pen.
I tried to reproduce the initial crash when drawing the calligraphic path, but didn't succeed so far (no tablet/pen available here either).
Possibly it was related to the earlier attempts editing the style attributes to get rid of the marker on the rectangle (or from the 'current style' setting) - you certainly tripped over this bug first:
Bug #420202 Rectangle start marker - cannot turn of https://bugs.launchpad.net/inkscape/+bug/420202
though I'm not aware that the combination of markers (from the current style setting) and a calligraphic path can trigger a crash which then results in a corrupt path (without path data) in the emergency-saved file.
~suv
participants (4)
-
Chris Mohler -
Josh Andler -
Martin Owens -
~suv