I left my game open overnight, paused (the menu you get from pressing "esc"); came back in the morning to find it had crashed. The last thing I did before leaving it was save the game; I have asynchronous saving enabled. Save file and log attached.
I noticed when I got back that pipewire had switched to only having "dummy output" available instead of the normal output (which is HDMI audio via my RX 580 GPU). I suspect this crash might have been caused by the active audio output device vanishing; not sure why that happened. It fixed itself when I restarted pipewire and wireplumber.
I am running Arch Linux with pipewire for audio. Factorio is running from Steam. Restarting factorio doesn't seem to immediately reproduce the crash.
[2.0.28] Crash in long session, unreadable stacktrace (potentially audio-related)
-
- Manual Inserter
- Posts: 3
- Joined: Mon Nov 11, 2024 6:20 pm
- Contact:
[2.0.28] Crash in long session, unreadable stacktrace (potentially audio-related)
- Attachments
-
- installed-package-versions.txt
- This is the output of `pacman -Q` -- which shows the installed version of all packages on my system, in case that is relevant.
- (33.35 KiB) Downloaded 11 times
-
- factorio-current.log
- (88.15 KiB) Downloaded 15 times
-
- Space Age.zip
- (60.61 MiB) Downloaded 12 times
Re: [2.0.28] Crash in long session, unreadable stacktrace (potentially audio-related)
Lack of readable stack trace makes this quite hard to sort or decide anything what is to be fixed, so unless you have some reproduction steps there is not much that can be done with this.
Re: [2.0.28] Crash in long session, unreadable stacktrace (potentially audio-related)
Just a note, active audio device disappearing shouldn't be a problem and it shouldn't crash (I've tested that many times by just removing active USB headset). So the audio stuff might be just a consequence of the actual issue (maybe something strange happened with the GPU, but I'm just guessing wildly).