Libreoffice crashes while Choose Insert - Indexes and Tables - Indexes and Tables. gdb output: Program received signal SIGSEGV, Segmentation fault. 0x00007fffc10b2c58 in SwConditionTxtFmtColl::SetConditions(SwFmtCollConditions const&) () from /usr/lib64/libreoffice/program/../program/libswlo.so
This crash occurs with Polish User interface. When switching back User interface to English (USE) inserting indexes & tables -> Index/Table -> table of contents.
This crash occurs with "stock" LibreOffice binary distribution as well as with Gentoo package built from sources.
Thank you for your report – unfortunately important information is missing. Please comment when does this happen - always, with new document or with specific file only? Please read http://wiki.documentfoundation.org/BugReport as it will help you find out what information will be useful to reproduce your problem.
Created attachment 73819 [details] strace for LibreOffice crash with Polish user interface
Crash occurs always: 1) Create new document -> Insert -> Indexes & Tables -> Index/Table -> Table of contents. 2) Open existing document -> Insert -> Indexes & Tables -> Index/Table -> Table of contents. File format doesn't matter -> occurs with both .odt and MS .doc It doesn't even present the TOC inserting form - it only splashes with it and quits Writer.
There are many jvm errors like: open("/usr/lib64/jvm/oracle-jdk-bin-1.7/jre/lib/amd64/server/libiodbcinst.so.2", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) Is your Java properly configured? (Just a guess from a nondeveloper).
User and system java vm are set to oracle jre 1.7 LO works with the very same java env settings with English Ui and crashes with Polish Ui.
*** Bug 60015 has been marked as a duplicate of this bug. ***
NEW as bug confirmed by another user (bug 60015). Changed Platform.
Martin: your description shows the crash happens here: SwConditionTxtFmtColl::SetConditions(SwFmtCollConditions const&) To be even more precise and since you reproduce it also with Gentoo, could you try to retrieve a backtrace by following this link? (https://wiki.documentfoundation.org/BugReport#How_to_get_a_backtrace_.28on_Linux.29)
(In reply to comment #10) > Martin: > your description shows the crash happens here: > SwConditionTxtFmtColl::SetConditions(SwFmtCollConditions const&) > > To be even more precise and since you reproduce it also with Gentoo, could > you try to retrieve a backtrace by following this link? > (https://wiki.documentfoundation.org/BugReport#How_to_get_a_backtrace_. > 28on_Linux.29) Julien, reporter attached strace already: strace for LibreOffice crash with Polish user interface - https://bugs.freedesktop.org/attachment.cgi?id=73819
bfoman: strace and backtrace are different, at they seem too :-) Concerning Linux, there are 2 links: for backtrace: https://wiki.documentfoundation.org/BugReport#How_to_get_a_backtrace_.28on_Linux.29 for strace: https://wiki.documentfoundation.org/BugReport#How_to_get_a_strace_log_.28on_Linux.29
Sorry... But a debug build on my Gentoo laptop costs 6 hours... Is there any binary debug release of LO 4 for amd64 so I could test it?
Martin: I don't know but you can try to retrieve a bt even without symbols. It may give some hints.
Ok. Will try bt tonight.
Ok. Here they are attached: gdbtrace.log.bz2 and by accident java runtime fatal error dump file hs_err_pid9058.log.bz2
Created attachment 73947 [details] Backtrace log file
Created attachment 73948 [details] java hotspot crash dump file
Martin: thank you for your feedback. In java hotspot crash, I noticed these fonts present in your profile: /home/marcin/.fonts/i/iwonamedium_regular.ttf /home/marcin/.fonts/i/iwonalight_regular.ttf /home/marcin/.fonts/i/iwonaheavy_regular.ttf /home/marcin/.fonts/i/iwonacondmedium_regular.ttf /home/marcin/.fonts/i/iwonacondlight_regular.ttf /home/marcin/.fonts/i/iwonacondheavy_regular.ttf /home/marcin/.fonts/i/iwonacond_regular.ttf /home/marcin/.fonts/i/iwona_regular.ttf For the test, could you remove all of them, update font cache and give a new try?
All suspected fonts from .fonts and regenerating fc-cache crash still occurs. This time no java hot spot crash file present... Attached backtrace and strace log files after removing fonts.
Created attachment 73955 [details] strace log file after removing fonts
Created attachment 73956 [details] gdbtrace file after removing fonts
Martin: I forgot to ask you if you can reproduce this on a brand new file or just with a specific file? If specific, could you attach the file so we can try to reproduce this? Thank you for your feedback and patience.
It is reproducible on a brand new file and on any existing .odt .doc files.
Martin: should have read your comment 5 :-( Did you install any specific LO extensions? Did you install any specific fonts? (in addition to those I told you to remove for the test) Did you try to rename your LO directory profile? (see https://wiki.documentfoundation.org/UserProfile)
(In reply to comment #25) > Did you install any specific LO extensions? Extensions: *MySQL Connector *Polish dictionaries *Presentation Minimizer *NLP Solver *Wiki Publisher > Did you install any specific fonts? (in addition to those I told you to > remove for the test) There is a bunch of fonts, but crash occurs even with Times New Roman set to default font as well as for UI font > Did you try to rename your LO directory profile? (see > https://wiki.documentfoundation.org/UserProfile) Yep! Removing LO directory profile from my home directory was the first step, prior to reporting bug
Still for the test, could you disable all these extensions and give a new try?
After disabling all extensions LO still crashes with UI language set to Polish.
Martin: thank you for all your feedback. Cédric/Michael: pb on Writer (for real this time :-)), bt attached. One for you?
I just want to confirm that the bug still exists in released 4.0 stable binary build. Any hope for Polish LO users ?
Please do not update version field as it indicates the earliest version where the bug has been found. Just add a comment. @jmadero Joel, this is very serious issue for all Polish users and hopefully will be fixed ASAP. It has been confirmed by another user already in Bug 60015.
*** Bug 60649 has been marked as a duplicate of this bug. ***
*** Bug 61248 has been marked as a duplicate of this bug. ***
All the experts are already CC'ed on the bug so hopefully one of them will assign themselves soon, I agree, quite serious
Adding one more expert just in case he wants to tackle it :) Michael - interested?
In testing version 4.0.1.1 is all right.
per comment 36 closing as FIXED, if it still appears for other people in 4.0.1.1 or later release, please reopen this. Thanks!