Description: As per summary. Steps to Reproduce: (1) Create /tmp/a01.odt as empty document. (2) Open that file, setting OOO_EXIT_POST_STARTUP=1. My command line was (all one line, rewrapped for readability) ... SAL_USE_VCLPLUGIN=gen OOO_EXIT_POST_STARTUP=1 instdir/program/soffice --norestore --nologo /tmp/a01.odt Actual Results: In terminal, error messages ... *** ERROR: DO NOT call xmlCleanupParser() Application Error Fatal exception: Signal 6 and a core file. Expected Results: Return to command line without error message and without, in particular, a core file. Reproducible: Always User Profile Reset: Yes Additional Info: The soon-to-be-attached backtrace is from a local build of commit 7e03a22b, pulled 2017-04-08, configured ... CC=ccache /usr/bin/gcc CXX=ccache /usr/bin/g++ --enable-option-checking=fatal --enable-dbgutil --enable-debug --without-system-postgresql --without-myspell-dicts --with-extra-buildid --without-doxygen --with-external-tar=/home/terry/lo_hacking/git/src --without-package-format built and running on debian-stretch. I see similar results in recent versions of daily Linux dbgutil bibisect repository. Other LibreOffice applications also either crash or hang with low CPU usage. Other bug reports show that people are mostly using OOO_EXIT_POST_STARTUP=1 successfully. (Bug 104816 comment 13 is an exception.) I wonder what is different about my setup. User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Firefox/45.0
Created attachment 132572 [details] gdb on the core file, with backtrace
Created attachment 132607 [details] bt with debug symbols On pc Debian x86-64 with master sources updated today, I got a crash too when doing this: export OOO_EXIT_POST_STARTUP=1 make debugrun then on gdb "run"
Terrence: there's no comment 13 in tdf#104816, wrong ref?
I meant bug 86448 comment 5. Sorry about that, Julien.
(In reply to Terrence Enger from comment #4) > I meant bug 86448 comment 5. Sorry about that, Julien. No pb :-) This quoted comment is quite old. Do you have some refs about working examples just to compare? Indeed, I'm quite surprised to have a bt without creating any file, using gen rendering.
No, this has never worked for me.
I have found command-line argument --eventtesting, which seems to accomplish the same end, and it does not crash for me. Since I having this workaround and many users seem to be working without the crash, I am setting status RESOLVED WONTFIX.