=== guiverc2 is now known as guiverc
[21:43] <OvenWerks> are there any ways of prioritizing what an OOM event kicks out?
[21:45] <OvenWerks> killing part of the DE seems like the worst solution as it pretty much kills everything. But that is what I have had happen a few times. Using 23.04.
[21:45] <OvenWerks> of course it may be FF causing a crash in one of the DE componants...
[22:01] <OvenWerks> Ok, might be fixed sinse. I think this line in syslog spells it out: kernel: [396310.220870] kwin_wayland[7870]: segfault at 7f38958ca104 ip 00007f39271add91 sp 00007ffe835b4070 error 6 in libpipewire-0.3.so.0.365.0[7f392715d000+5a000] likely on CPU 1 (core 1, socket 0)
[22:02] <OvenWerks> It seems pipewire has segfaulted when wayland called it.
[22:02] <OvenWerks> (or wayland segfaulted whilst using a pw lib call.)
[22:04] <OvenWerks> studio doesn't (or didn't for 13.04) use wayland. So not worth a bug.
[22:05] <OvenWerks> I still think it is FF connected because it happened when starting FF. (which opens a lot of stuff all over)
[22:48] <Eickmeyer> OvenWerks: Just saw this.
[22:49] <Eickmeyer> Pipewire in 23.04 was... buggy.
[22:49] <Eickmeyer> Didn't know until after release.
[22:50] <Eickmeyer> By then it was too late, and even then the implementation felt hurried.
[22:54] <Eickmeyer> What's happening is, anytime a snap is opened (for the first time), it does a real fast squashfs decompression which takes quite a few CPU interrupts. Side effect is it steps on everything. I've noticed it causes xruns in jackd2, not so much in PW, but the PW in 23.10 is much more solid.
[22:55] <Eickmeyer> I have, in fact, brought this before the Ubuntu Desktop team, and they at least made a note of it.