Bug#717112: hang debugging ...
Hi there, Can you install the relevant debuginfo packages, connect to libreoffice and get some backtraces: 't a a bt' in gdb. I suspect the thumbnailing piece myself, we do some call there that in theory is magic (and all that) but it's some synchronous magic and I'm suspicious of it. Looking forward to the trace. Thanks ! :-) Michael. -- michael.me...@suse.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#700728: Crash when starting Impress Remote
Hi Ben, On Tue, 2013-03-05 at 05:36 +, Ben Hutchings wrote: OK, I'll try again with 4.0.1 (if I remember). It works for me in 4.0.1-rc1, thanks. (Tested with Bluetooth only.) Great ! thanks for testing :-) hopefully it's a cool feature to promote LibreOffice with. All the best, Michael. -- michael.me...@suse.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#700728: Crash when starting Impress Remote
On Sat, 2013-02-16 at 23:26 +, Ben Hutchings wrote: What information do you need from me to debug this? Heh - thanks so much for trying it out; it really is in a bit of a state in 4.0.0 - a swathe of that stuff has been re-written for 4.0.1 and we're just working on getting that reviewed / in for 4.0.1rc1 - due this week; with 4.0.1 due in ~2 weeks or so. The wiki page has a long list of caveats for this version ... Sorry about that, Michael. -- michael.me...@suse.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#608732: stack trace ?
Hi Alex, On Wed, 2011-01-19 at 09:12 +0200, alex bodnaru wrote: i'm trying to make the soffice environment on my machine invoke gdb with the binary. Great. to date, i can see a problem related to file locking: Program received signal SIGSEGV, Segmentation fault. __pthread_mutex_lock (mutex=0x1) at pthread_mutex_lock.c:50 50pthread_mutex_lock.c: No such file or directory. in pthread_mutex_lock.c Well - if you get a SEGV; we need to fix it - LibreOffice should never SEGV. as my home dir is on nfs, i've turned FILE_LOCKING to no, with no visible difference. let me know where to continue. different type of locking - can you give us the results of: thread apply all backtrace ? Thanks, Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#608732: stack trace ?
On Mon, 2011-01-24 at 15:11 +0200, alex bodnaru wrote: well i've downloaded RC4 from upstream and it doesn't manifest any of the symptoms i signaled. So RC4 works for you ? If so wonderful; but we didn't fix anything in this area, that I'm aware of :-) ATB, Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#607715: [Libreoffice] [PUSHED] + review requests i#116187: report left/right page margin setting ignored on Unix 64 bit
On Sat, 2011-01-15 at 20:00 +, Caolán McNamara wrote: I pushed this to master and to the stable 3-3 branch, thanks for this. We'd need some more acks for 3-3-0 however, I'm +1 on that. Thirded, and pushed to libreoffice-3-3-0 Thanks ! Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#608732: stack trace ?
Hi there, On Thu, 2011-01-13 at 15:07 +0200, alex bodnaru wrote: i'd love to collect the backtrace, just be more instructive on the steps i should follow: needed packages, command lines etc. First - it would be much appreciated if you could use the freedesktop bug. Second - can you confirm that you can actually install your extension (move away your ~/.libreoffice first perhaps) - I cannot (I get errors wrt. missing registration methods). Third - if you get a crash you need to: cd /path/to/libreoffice/program gdb ./soffice.bin r # make it crash # :-) thread apply all backtrace and attach the backtrace - bonus points for having debuginfo packages installed before you run it ;-) If that shows memory corruption, then repeating with valgrind instead of gdb: valgrind --num-callers=25 ./soffice.bin 21 | tee /tmp/val-log may give a more useful trace. In the meantime, it is hard to see the bug (as is) as a blocker. Thanks ! Michale. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#608732: stack trace ?
Hi Alex, I'm having trouble reproducing this; perhaps it is bugs on our side; certainly installing the extension on master fails on install; I'm doing a fresh 3.3.0 build now to test; I was wondering; if you could run soffice.bin inside gdb - and get a stacktrace from that (thread apply all backtrace) with debugging symbols from your crash-on-exit. We have fixed a number of memory leaks in 3.3, and sadly some code, off the beaten path, may depend on the leaks still. Also - responses to the freedesktop bugzilla much appreciated. Thanks ! Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#604634: cause ?
It looks like the collector is still running for some reason; of course, while it is running and has that proc filesystem open - it cannot be unmounted. So I would look into why the collector is not quitting when it should - we should kill it (or ptrace kill it I forget which - and then it should die) ;-) -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#595229: openoffice.org crashed then froze with _XAllocID: Assertion `ret != inval_id' failed.
Hi Vincent, On Thu, 2010-09-02 at 15:57 +0200, Vincent Lefevre wrote: (gdb) thread apply all backtrace Ok - that's quite helpful. Probably the bug is more banal than a complex threading problem: a simple memory corruption breaking the (thread safe anyway) heap primitives. Is there any chance you can install the cups and openoffice debuginfo packages (so we get a better trace), and run it under valgrind [ this will take a while ;-]: cd /usr/lib/ooo3/program # [or Debian equivalent] export OOO_FORCE_SYSALLOC=1 valgrind ./soffice.bin -writer 21 | tee /tmp/val-log And send in /tmp/val-log ( probably best compressed unless you can supress the runaway java nonsense ;-). Hopefully that will pin-point the heap corruption to the line. Many thanks, Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#595229: openoffice.org crashed then froze with _XAllocID: Assertion `ret != inval_id' failed.
Hi Vincent, On Mon, 2010-09-06 at 11:32 +0200, Vincent Lefevre wrote: And send in /tmp/val-log ( probably best compressed unless you can supress the runaway java nonsense ;-). Hopefully that will pin-point the heap corruption to the line. It's only 34 KB. Attached. Wonderful ;-) thanks for that - it seems to show the bug in an interaction with freetype: ==21743== Invalid read of size 8 ==21743==at 0xC059F30: FT_Done_Glyph (in /usr/lib/libfreetype.so.6.6.0) ==21743==by 0x93DE72E: FreetypeServerFont::InitGlyphData(int, GlyphData) const (in /usr/lib/openoffice/basis3.2/program/libvcllx.so) .. ==21743== Address 0x18 is not stack'd, malloc'd or (recently) free'd which triggers the segv handler: ==21743== Conditional jump or move depends on uninitialised value(s) ==21743==at 0x4E5CF96: SignalHandlerFunction (in /usr/lib/ure/lib/libuno_sal.so.3) ==21743==by 0x5C771DF: ??? (in /lib/libc-2.11.2.so) .. Unfortunately, you didn't install the debuginfo symbols [ or you are using some architecture valgrind can't unwind ]. Any chance you can install the freetype and OO.o debuginfo and get another valgrind trace ? [ Rene what packages / commands would Vincent need for that ] ? Thanks, Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#595229: openoffice.org crashed then froze with _XAllocID: Assertion `ret != inval_id' failed.
Hi Vincent, On Thu, 2010-09-02 at 12:21 +0200, Vincent Lefevre wrote: #25 0x7f11af6e41f1 in EditEngine () from /usr/lib/openoffice/program/../basis-link/program/libsvxcorelx.so .. #29 0x7f11b059d532 in ScDocument::ContinueOnlineSpelling () from /usr/lib/openoffice/program/../basis-link/program/libsclx.so .. #33 0x7f11bb1236fc in SalXLib::Yield () from /usr/lib/openoffice/basis3.2/program/libvclplug_genlx.so #34 0x7f11c6596ebe in Application::Yield () from /usr/lib/openoffice/program/../basis-link/program/libvcllx.so .. #37 0x7f11c684c43a in psp::CUPSManager::authenticateUser () from /usr/lib/openoffice/program/../basis-link/program/libvcllx.so .. #44 0x7f11c933e8ba in start_thread (arg=value optimized out) at pthread_create.c:300 Is it possible to disable things related to cups, to see if this solves the problems (since the authentication dialog is something new)? I'm rather surprised to see this main-loop, and spell-check stuff running from another (the CUPS polling) thread - and I'm surprised at the suggestion that CUPS might be thread-safe. We really need a trace from all threads if that's possible: thread apply all backtrace in gdb :-) Thanks, Michael. -- michael.me...@novell.com , Pseudo Engineer, itinerant idiot -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org