aryak_ has quit [Read error: Connection reset by peer]
aryak_ has joined #river
orfeas194 has joined #river
orfeas194 has quit [Ping timeout: 255 seconds]
eShaev9z_ has joined #river
eShaev9z has quit [Ping timeout: 255 seconds]
sunn has joined #river
sunn has quit [Client Quit]
talismanick has joined #river
tiosgz has joined #river
<tiosgz>
ifreund: aye, thanks for the explanation
tiosgz has quit [Quit: tiosgz]
leopoldek has quit [Remote host closed the connection]
waleee has joined #river
waleee has quit [Ping timeout: 264 seconds]
traidare has joined #river
<novakane>
hmm seems like `Output: retry other modes if preferred fails` and `Output: fix regression of initial mode logic` break evertyhing on wlroots 0.17 on my system
<novakane>
with `Output: fix regression of initial mode logic` it seems it works if I launch river in a new tty but not in a nested session as all my clients can't find any outputs
<novakane>
could be clients at faults I guess
<novakane>
my bad it's `Root: fix faulty assertion, cleanup` not `Output: retry other modes if preferred fails`
<novakane>
but with `Root: fix faulty assertion, cleanup` I can't launch river at all
<novakane>
the other day I couldn't launch it because I didn't have backend then renderer, I fix it but didn't have time to test after and now I can't launch anymore again
<novakane>
wild adventure to try to test this pr xD
<ifreund>
yeah, so that's the thing right. I just haven't tested it at all myself
<ifreund>
I've been getting bits of time where I write code and then end up pushing what I have before I have time to test proprely
<leon-p>
I usually marks such commits with [WIP] so people know they are untested
<ifreund>
that might have been a good idea
<ifreund>
I did put it in the PR description though :)
<leon-p>
:] fair
<ifreund>
anyhow, it seems to work fine on the DRM backend now after fixing my typo
<ifreund>
the wayland backend doesn't work though, I guess because modes mean something different there
<novakane>
oh yeah, I saw it's untested in the description, that's why I test it :D
<novakane>
it wasn't to blame ifreund or that it bother me, I know what I signed for with testing this PR :P
<ifreund>
thanks for testing :)
<novakane>
seems like arch ci need libdisplay-info too
<novakane>
np, that's what open source is all about :)
<ifreund>
ah, sr.ht fixed the arch ci :)
<leon-p>
this much river work means I'll have to mute email notifications on my pinetime again :D
<leon-p>
don't really want to develop an urge to read a commit in the middle of a working group session
<novakane>
leon-p: at least reading commits would not cause you to write quickly a new wayland tools to fix something :P
<novakane>
ifreund: I was gonna say that the wayland backend works for me but I see you push a fix in the meantime lol
<novakane>
I can finally launch river with wlroots 0.17 \o/
traidare has quit [Ping timeout: 268 seconds]
TheAnachron has joined #river
traidare has joined #river
alexherbo2 has joined #river
<ifreund>
I think I'd like to move the issue tracker over to codeberg with the 0.3.0 release and start doing as much as possible there
<leon-p>
+1
<ifreund>
the 0.3.0 release seems like a good time as I can make the information prominent in the release notes
<novakane>
seems like a good time for that yeah
alexherbo2 has quit [Remote host closed the connection]
<leon-p>
it's both funny and sad that I can replace ~1k LOC over-engineered layout logic with ~450 LOC C and ~80 LOC scheme
<leon-p>
to be fair, most of that 1k is configuration and modularity, but still...
<ifreund>
that is in fact one of the reasons I want to get as much config shit as possible out of the compositor process xD
traidare has quit [Ping timeout: 276 seconds]
<LarstiQ>
I had two crashes today on `zoom`. Before I start digging any hints on how to debug that?
<leon-p>
arguably if we ripped out river-control and grafted a scheme interpreter onto river we could get rid of the command parser. Which in turn would allow us to get rid of keymap modes since they can be done in the scheme-config, basically "user space"
<leon-p>
LarstiQ: zoom is insanely badly programmed, at least on Wayland. if you value your sanity, use the browser version
<leon-p>
back when I tried it, it did not correctly set resize hints, for example
<leon-p>
and the way it did screencapture was a joke: it just took screenshots multiple times a second (only worked in GNOME anyway, but still a good thing to point out as a case-study)
<leon-p>
I don't like to talk badly about other peoples work usually, but how zoom behaved on Wayland is genuinely bad, as if the devs who wrote that didn't know anything about Wayland
<ifreund>
LarstiQ: do you mean the zoom riverctl command or the Zoom video conferencing program?
<ifreund>
in any case, getting a stack trace should be your first step
<ifreund>
if core dumps are configured on your system you may very well have a useful stack trace waiting to be examined
<ifreund>
oh, and is river crashing or is Zoom the program crashing? If it's the later there's not really anythng we can do...
<ifreund>
leon-p: I don't really want to have a scripting language in the compositor process tbh
<leon-p>
fair. I just don't really like the command parser and how restrictive it is especially w.r.t. keymaps
<ifreund>
oh the current river-control protocol approach of pasing string arguments over wayland and parsing in river has got to go
<leon-p>
the main restriction IMO is that you can't really do any control-flow in response to maps unless you use a script or something which is a bit annoying
<LarstiQ>
I meant the zoom riverctl command
<LarstiQ>
I presume it's riverwm itself going down since I'm bounced back to the display manager
<LarstiQ>
ifreund: cheers, I'll make sure to get coredumps
<novakane>
+1 for the current river-control to go
<novakane>
leon-p: have you seen there is a guile-wayland, example seems nice