STR: (1) Download template 06_minutes.ott from https://bugs.freedesktop.org/attachment.cgi?id=59830 and open it in Writer. Program displays dislog "LibreOfficeDev 4.3" with a section headed "Minutes Type". (2) Drag that dialog box around. (3) In the LibreOffice dialog, click the <OK> button. Program issues the message soffice.bin: /home/terry/lo_hacking/git/libo4/vcl/source/app/dbggui.cxx: 1624: void ImplDbgTestSolarMutex(): Assertion `ImplGetSVData()->mpDefInst->CheckYieldMutex()' failed. Application Error Fatal exception: Signal 6 and exits. My LibreOffice is master commit 806f4d8, fetched 2014-03-04, configured: --enable-option-checking=fatal --enable-dbgutil --enable-crashdump --without-system-postgresql --without-myspell-dicts --with-extra-buildid --without-doxygen --with-external-tar=/home/terry/lo_hacking/git/src built and running on debian-wheezy. BTW, this started as I was looking at Thomas Hackert's message on the qa list "Crash, when undoing numbering of headers in a Writer template" <http://nabble.documentfoundation.org/Libreoffice-qa-Crash-when-undoing-numbering-of-headers-in-a-Writer-template-tp4100662.html> and the bug report arising, bug 75996 "Crash, when undoing steps in attachment from bug #48582".
Created attachment 95807 [details] backtrace with symbols Note frame 17: #17 0x00007f77b9148cc3 in ImplDbgTestSolarMutex () at /home/terry/lo_hacking/git/libo4/vcl/source/app/dbggui.cxx:1624 The referenced function is: 1622: void ImplDbgTestSolarMutex() 1623: { 1624: assert(ImplGetSVData()->mpDefInst->CheckYieldMutex()); 1625: }
I am leaving the bug status UNCONFIRMED because I suspect that the problem may be not widespread. Well, obviously it can happen only in a debug build, but even less widespread than that. OTOH, developers may be interested just because solar mutex is involved. Advice welcome.
FWIW, I see something similar when testing Firebird in my debug build : warn:legacy.osl:22546:1:connectivity/source/drivers/firebird/DatabaseMetaData.cxx:1447: Not implemented yet! warn:tools.debug:22546:1:tools/source/debug/debug.cxx:1121: no DbgTestSolarMutex function set
(In reply to comment #3) > FWIW, I see something similar when testing Firebird in my debug build : > > warn:legacy.osl:22546:1:connectivity/source/drivers/firebird/ > DatabaseMetaData.cxx:1447: Not implemented yet! > warn:tools.debug:22546:1:tools/source/debug/debug.cxx:1121: no > DbgTestSolarMutex function set That's unrelated to Terrence's issue. Here, the warning is because (presumably in an additional process besides soffice.bin?) the functionality has not been set up to detect potential violations of SolarMutex locking. If you have a recipe to reproduce it, I can have a look to eliminate the warning.
(In reply to comment #4) Hi Stephan, > > That's unrelated to Terrence's issue. Here, the warning is because > (presumably in an additional process besides soffice.bin?) the functionality > has not been set up to detect potential violations of SolarMutex locking. > If you have a recipe to reproduce it, I can have a look to eliminate the > warning. OK, will open another bug report for that. Easy to reproduce on my system. Create an empty Firebird embedded db. Add a table and/or data, e.g. by dragging and dropping from a Calc sheet. Save, and close the Firebird ODB file. But, like I said, I will open a separate issue for this. Alex
@Terrence : I don't get a crash with my debug master 440 build on OSX, with the document you indicated. Do you still get this ?
Created attachment 108117 [details] backtrace from master 2014-10-14 Thank you, Alex, for the reminder. The daily dbgutil bisect version 2014-10-20 and my dbgutil of master commit 9d24f6b, fetched 2014-10-14, fail with a different assertion: soffice.bin: /home/terry/lo_hacking/git/libo4/vcl/source/app/dbggui.cxx:793: void ImplDbgTestSolarMutex(): Assertion `ImplGetSVData()->mpDefInst->CheckYieldMutex()' failed. Attached backtrace is from my build. I am *assuming* that this information belongs here, but if it deserves a new bug report, I can do that. The daily dbgutil bisect version 2014-10-20 raises the same assertion. I am setting bug status UNCONFIRMED. My build is master commit 9d24f6b, fetched 2014-10-14 18:56 UTC, configured: CXXFLAGS=-std=c++11 --enable-option-checking=fatal --enable-dbgutil --enable-crashdump --without-system-postgresql --without-myspell-dicts --with-extra-buildid --without-doxygen --with-external-tar=/home/terry/lo_hacking/git/src --disable-gstreamer --enable-gstreamer-0-10 built and running on debian-wheezy 64-bit.
The assertion is already present in daily dbgutil bibisect version oldest (of course: the bug report is older than the repository), and other bibisect repositories have product builds. Setting whiteboard notBibisectable.
Even though the backtrace has changed, this will only ever have happened on a dbgutil build so I think it's fine to go with the latest master backtrace. If it turns out there's another assert still hiding under the latest one, that can be dealt with later. I can reproduce the latter backtrace with current master on both Linux and OSX, so setting -> NEW and Platform: All
Adding self to CC if not already on
fixed on master
the bots are lazy today, commit is a4faad707c7fb27884fff5f20e5797b160d4f320
Michael Stahl committed a patch related to this issue. It has been pushed to "libreoffice-4-4": http://cgit.freedesktop.org/libreoffice/core/commit/?id=22671c1545f8bfff660acdf981d6d36bf73ded63&h=libreoffice-4-4 tdf#76174: basic: lock SolarMutex in BasicScriptListener_Impl It will be available in 4.4.1. 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.
The crash is gone with master commit dec9050d, fetched 2015-01-31 04:40 UTC. I am setting bug status to VERIFIED FIXED. Thank you, Michael.
Migrating Whiteboard tags to Keywords: (notBibisectable) [NinjaEdit]