On 2017-05-06, Peter N. M. Hansteen <pe...@bsdly.net> wrote: > On my laptop, which tends to run the latest amd64 snapshot or one > that's at most a few days old, the following sequence of events will, > approximately one out of ten times, kill my X: > > $ soffice fehfeh.csv & > > where fehfeh.csv is the kind of file that went into the data for my pop3 > password groper columns - available via the blog referenced in my > .signature and just browsing for the obvious keywords) > > Then selecting a column (the username column) and using the Column Type > dropdown to select Text typefor the column. > > Then we have a black, empty screen for a moment and then the xenodm > login dialog box reappears. > > The most annoying thing here is that there is never an soffice.core > file, although other processes that were active at the time may leave > core files. This last time around it looks like only emacs was killed in > such a way that it left a corefile: > > [Sat May 06 09:50:22] peter@elke:~$ ls -ltr *core > -rw------- 1 peter peter 1528700960 May 5 22:56 firefox.core > -rw------- 1 peter peter 56259040 May 6 09:25 emacs-25.2.core > > (the firefox.core here is too old to be relevant here). > > So the question is, of course: How do I go about usefully debugging this? >
I've seen this once, but wasn't able to trigger it again. Check /var/log/Xorg*, and there are instructions in http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/xenocara/README?rev=1.39&content-type=text/plain about obtaining coredumps from X.
