14:41 <wrl> alright i've managed to hit this xwayland crash a few more times

14:41 <wrl> gonna see if i can track it down

15:00 <ifreund> wrl: any information you can gather would be useful, is it always the same program that triggers it?

15:02 <wrl> ifreund: nah, i had it happen with sxiv earlier, which was new

15:03 <wrl> ifreund: i've got river built from the AUR, is there a convenient way to get a backtrace?

15:03 <wrl> you mentioned something about a debug build and redirecting the log somewhere

15:05 <ifreund> wrl: if you remove the -Drelease-safe in the PKGBUILD you'll get a debug build, which will give you the best backtrace

15:06 <ifreund> I start river like this in order to get logs: https://github.com/ifreund/dotfiles/blob/0e832a3354f631d7e47b7ef47244c0ddd966ec93/local/bin/start_river.sh#L8

15:07 <ifreund> er sorry that's out of date, -log-level debug is what you want

15:09 <ifreund> fixed: https://github.com/ifreund/dotfiles/blob/c2f16677468391f2d8b4b9d452199f27020cb8d6/local/bin/start_river.sh#L8

15:12 <novakane> ifreund: do you have any idea where the last known bug come from btw? the high polling rate one

15:17 <ifreund> novakane: yep: https://libera.irclog.whitequark.org/river/2021-07-23#1627061086-1627061712;

15:17 <ifreund> I'm busy with exams currently though :/

15:20 <novakane> ah right I probably saw it but forget about it. I was reading this issue again, I don't know how people can use their mouse with like 1000Hz 2400dpi, I would go crazy lmao

15:33 <wrl> ifreund: got it https://0paste.com/279988-6bdced95

15:37 <ifreund> wrl: oh shit, does makepkg strip debug info?

15:38 <ifreund> I'm guessing it probably does...

15:38 <wrl> ifreund: yeah that was my next thought

15:38 <wrl> let's see

15:38 <wrl> ifreund: hm, nothing i can see

15:40 <ifreund> wrl: what does file /path/to/river/binary say?

15:40 <ifreund> stripped or not stripped?

15:40 <wrl> stripped

15:40 <wrl> going to pass `-fno-strip` to zig build install

15:40 <wrl> oh, it doesn't like that.

15:45 <wrl> yeah i can't figure out where it's being stripped

15:49 <wrl> ifreund: ah hold up, looks like it's an implicit thing makepkg is doing

15:52 <wrl> ifreund: yep, adding `options=(!strip debug)` to the PKGBUILD fixes it

15:52 <wrl> just for future reference

15:54 <wrl> alright got it

15:55 <wrl> ifreund: http://paste.debian.net/1206564/

16:02 <ifreund> wrl: excellent, thanks

16:02 <ifreund> that should be enough for me to fix it

16:02 <wrl> alright cool

16:03 <wrl> ifreund: also – i was looking through river to add an "add before/after" for new window position (in addition to top/bottom), and i was curious as to whether there's a reason you didn't want to use a std linked list (eq TailQueue) for the ViewStack

16:08 <ifreund> wrl: I think I originally did that because I had a lot more of the "business logic" checking for matching tags, etc built into the type

16:09 <ifreund> I'll probabably switch it to something else when I implement https://github.com/ifreund/river/issues/357

16:10 <wrl> mm, that makes sense

16:10 <wrl> yeah i didn't want to start adding low-level linked list manipulation code without tests