Description: regression, Version: 5.4.0.0.alpha0+ Build ID: b29ae3dace6a6a2e24e531bb648cd3cb97ea3225 CPU Threads: 4; OS Version: Linux 4.4; UI Render: default; VCL: gtk2; TinderBox: Linux-rpm_deb-x86_64@70-TDF, Branch:master, Time: 2016-12-14_10:19:59 Steps to Reproduce: 1.write text to cell 2.copy 3.right click -> paste only -> text Actual Results: crash Expected Results: text pasted Reproducible: Always User Profile Reset: No Additional Info: User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:50.0) Gecko/20100101 Firefox/50.0
Could be a dup of tdf#104481 put in see also. Anyway, I could reproduce this on pc Debian x86-64 with master sources updated today. I put the max priority because: - it's a crash - it's a regression - it's a basic use case
it looks more like a duplicate of bug 104572
Regression introduced by: author Kohei Yoshida <kohei.yoshida@collabora.com> 2016-11-22 03:56:46 (GMT) committer Kohei Yoshida <libreoffice@kohei.us> 2016-11-22 13:18:00 (GMT) commit 87e040fd0f04307534920d0765af6d5878794a98 (patch) tree 8faa3f12d311d9c1fc205d9c333e2b15da934894 parent a0ef7474521413c8967559a635e6fdc0d88f1df6 (diff) tdf#71409: Stop listening to objects going out-of-focus. This helps eliminate the number of AtkListener objects growing over time. Adding Cc: to Kohei Yoshida
*** Bug 104572 has been marked as a duplicate of this bug. ***
On pc Debian x86-64 with master sources updated today, I confirm than just reverting 87e040fd0f04307534920d0765af6d5878794a98 (quoted by Xisco) makes the bug not reproduceable.
I tried to understand why it doesn't work on gtk2 but I'm a bit lost. I noticed this part on gtk2 #31 0x00002aaac8dc7dc0 in GtkData::Yield (this=0x5555558ae320, bWait=true, bHandleAllCurrentEvents=false) at /home/julien/lo/libreoffice/vcl/unx/gtk/gtkdata.cxx:499 #32 0x00002aaac8dcc1b9 in GtkInstance::DoYield (this=0x5555558b9cc0, bWait=true, bHandleAllCurrentEvents=false, nReleased=0) at /home/julien/lo/libreoffice/vcl/unx/gtk/gtkinst.cxx:427 #33 0x00002aaab5b599ad in ImplYield (i_bWait=true, i_bAllEvents=false, nReleased=0) at /home/julien/lo/libreoffice/vcl/source/app/svapp.cxx:506 #34 0x00002aaab5b55784 in Application::Yield () at /home/julien/lo/libreoffice/vcl/source/app/svapp.cxx:551 on gtk3, it goes to vcl/unx/gtk/gtkinst.cxx:427 not in vcl/unx/gtk/gtkdata.cxx:499 but I suppose it's expected. Caolan: since it concerns vcl and a difference of treatment between gtk2/gtk3, thought you might be interested in this one. BTW, will LO abandon gtk2 soon or it isn't expected at all in near future (some months)?
Any chance of a back-trace with symbols ? what does the this object look like inside stopListening ? in particular is it a simple NULL mpWrapper de-reference ?
Michael: I had put a part of bt here: https://bugs.documentfoundation.org/show_bug.cgi?id=104481#c3 I got it by typing Ctrl-C when there was the empty popup error. I didn't keep the complete one of https://bugs.documentfoundation.org/show_bug.cgi?id=104687#c6. Given the fact that we don't get a crash but an abort, where should I put a break in gdb to retrieve something interesting?
Could be that Michael Stahl fixed this in bug#104488 ? =)
(In reply to Michael Meeks from comment #9) > Could be that Michael Stahl fixed this in bug#104488 ? =) it's still reproducible in Version: 5.4.0.0.alpha0+ Build ID: 5a20df55ff829978c880b22e0a1f32c35d0ba30f CPU Threads: 4; OS Version: Linux 4.8; UI Render: default; VCL: gtk2; Locale: ca-ES (ca_ES.UTF-8); Calc: group
Created attachment 129852 [details] 3 bts got with __cxa_throw Michael: I followed your advice from https://bugs.documentfoundation.org/show_bug.cgi?id=104481#c4 Hope these 3 bt help.
Reverting https://cgit.freedesktop.org/libreoffice/core/commit/?id=a0ef7474521413c8967559a635e6fdc0d88f1df6 make disappear the crash here but not for tdf#104481
@julien FWIW gtk2 isn't going away in any short to medium term
Caolán McNamara committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=88dad8fe1a94055dcbb05a6e6f0df2d007914a45 Resolves: tdf#104687 gtk2 paste cell text under a11y is fatal It will be available in 5.4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
brute force fix, following bt3 and makes it work here again. Might want to check the other gtk2 things after using a popup menu to see if there are also fixed
Caolán McNamara committed a patch related to this issue. It has been pushed to "libreoffice-5-3": http://cgit.freedesktop.org/libreoffice/core/commit/?id=08e6fe3d120ad3e64c42022686579982a8af11c1&h=libreoffice-5-3 Resolves: tdf#104687 gtk2 paste cell text under a11y is fatal It will be available in 5.3.0.1. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Thank you Caolán for the fix. I've retrieved it and gonna test it once local build will be finished. Hope it'll help for tdf#104481 too. Concerning gtk2, thank you for your feedback, I wasn't sure about it.
With master sources updated yesterday (which includes the fix for this bugtracker), I don't reproduce the crash. The weird thing is I don't see any log corresponding to this part: catch (const uno::Exception& e) { SAL_INFO("vcl.a11y", "exception: " << e.Message); }
After changing locally SAL_INFO to SAL_WARN for the test, I got this: warn:vcl.a11y:17405:1:vcl/unx/gtk/a11y/atkutil.cxx:106: exception: so no exception message here. (idem for tdf#104481)