lubot_ | [telegram] <RikMills> some people are getting the crash with realloc(): invalid next size. I guess that bug can be marked duplicate? | 10:05 |
---|---|---|
RikMills | BTW, as far as I can tell this was all started by mesa being bumped from 20.3.4 to 21.0.0 | 10:15 |
RikMills | also the gtk front end works fine. only the pyqt/kde front end crashes | 12:32 |
mitya57 | If it's related to mesa, then maybe it's something with the llvmpipe renderer. | 12:33 |
RikMills | just realiesd I haven't linked the bug | 12:34 |
mitya57 | In any case I recommend you to obtain a stacktrace from valgrind or gdb with symbols. | 12:34 |
RikMills | https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1920665 | 12:34 |
ubottu | Launchpad bug 1920665 in ubiquity (Ubuntu Hirsute) "Ubiquity KDE crash on try/install and from live session with "malloc(): unaligned tcache chunk detected"" [Critical,Confirmed] | 12:34 |
RikMills | mitya57: yes, I think llvmpipe also, as it doesn't happen on my intel laptop run with full graphics support, but does if I run in low graphics mode | 12:35 |
RikMills | mitya57: just setting up a VM with the symbols. I ran out of time to get to it yesterday | 12:40 |
lubot_ | [telegram] <RikMills> https://matterbridge.lubuntu.me/a233a1da/file_2179.jpg | 12:51 |
lubot_ | [telegram] <RikMills> That is what I get | 12:52 |
lubot_ | [telegram] <mitya57> @RikMills What about gdb? Or valgrind without PYTHONMALLOC? | 13:12 |
lubot_ | [telegram] <RikMills> I am just pastebinning the output without | 13:12 |
lubot_ | [telegram] <RikMills> https://paste.ubuntu.com/p/gGJVX28pwx/ (re @mitya57: @RikMills What about gdb? Or valgrind without PYTHONMALLOC?) | 13:14 |
lubot_ | [telegram] <mitya57> Still not much helpful :( | 13:19 |
lubot_ | [telegram] <mitya57> Maybe try gdb because it at least prints the arguments, which may contain information about what Python code is executed. | 13:19 |
lubot_ | [telegram] <mitya57> Also can you try running python3 -X faulthandler /usr/bin/ubiquity? | 13:20 |
lubot_ | [telegram] <RikMills> what would the gdb command be? | 13:23 |
lubot_ | [telegram] <mitya57> Forwarded from mitya57: gdb python3, then run /usr/bin/ubiquity (or whatever its name is) | 13:26 |
lubot_ | [telegram] <RikMills> ubiquity needs sudo so that is not working | 13:30 |
lubot_ | [telegram] <RikMills> or gain perms with pexec | 13:35 |
lubot_ | [telegram] <RikMills> https://matterbridge.lubuntu.me/2338dc0c/file_2180.jpg | 13:35 |
lubot_ | [telegram] <mitya57> sudo gdb python3? | 13:36 |
lubot_ | [telegram] <RikMills> I tried that. One sec | 13:37 |
lubot_ | [telegram] <RikMills> https://matterbridge.lubuntu.me/bbed2fe0/file_2181.jpg | 13:38 |
lubot_ | [telegram] <mitya57> Oh, forks | 13:38 |
lubot_ | [telegram] <mitya57> No ideas 😢 | 13:40 |
lubot_ | [telegram] <RikMills> Ok. Thanks all the same :) | 13:42 |
lubot_ | [telegram] <mitya57> Maybe add prints and find which line or import causes the error? | 13:43 |
lubot_ | [telegram] <RikMills> Mesa 21.0.1 bugfix is supposed to be out today, so I may try building that 1st (re @mitya57: Maybe add prints and find which line or import causes the error?) | 13:44 |
lubot_ | [telegram] <RikMills> But yes, I have done that previously for something I forget | 13:44 |
lubot_ | [telegram] <mitya57> Makes sense | 13:44 |
lubot_ | [telegram] <RikMills> I am also not sure that having glibc with new patches stuck in proposed which mesa built with is not a problem | 13:45 |
lubot_ | [telegram] <RikMills> Hmmm. maybe try updating my VM to proposed.... | 13:46 |
lubot_ | [telegram] <RikMills> Nope. That does not help | 13:56 |
lubot_ | [telegram] <RikMills> Giving up for a bit. Thanks! | 13:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!