ifreund changed the topic of #river to: river - a dynamic tiling wayland compositor || https://github.com/riverwm/river || channel logs: https://libera.irclog.whitequark.org/river/
elshize has joined #river
elshize has quit [Client Quit]
waleee has quit [Ping timeout: 264 seconds]
jao has quit [Ping timeout: 260 seconds]
eShaev9z has quit [Ping timeout: 246 seconds]
eShaev9z_ has joined #river
silv3r has joined #river
silv3r has quit [Quit: WeeChat 3.8]
angry_vincent has joined #river
angry_vincent has joined #river
Szadek has quit [Quit: WeeChat 3.8]
zdykstra has quit [Ping timeout: 268 seconds]
zdykstra has joined #river
Ordoviz has joined #river
taupiqueur has quit [Ping timeout: 255 seconds]
taupiqueur has joined #river
taupiqueur has quit [Ping timeout: 256 seconds]
Ordoviz has quit [Ping timeout: 264 seconds]
<novakane> I find that switching a view to tilling/floating mode is super smooth on the scene branch, never switched so much to floating in my life :P
<novakane> anyway seems like most bugs reported have been fixed, don't think I have anything with CSD to test the related problems but other than the map/unmap bug river run nicely
taupiqueur has joined #river
<novakane> oh wait, if I put firefox in fullscreen, when exiting it the tabs bar and address bar are not rendered anymore
<novakane> it seems that xwayland clients are only rendered if I switch to an other tag and switch back to the one with the view
Ordoviz has joined #river
<novakane> hmm actually it seems to be unfocus and refocus the view that render it
Ordoviz has quit [Ping timeout: 246 seconds]
<ifreund> novakane: should have that fullscreen bug you saw with firefox fixed now
<ifreund> I think that bug you're seeing with the xwayland clients is related to the map/unmap issue. I've actually hit that with wayland clients as well while running under valgrind xD
<novakane> yep that's fixed
<novakane> oh, have you any idea what's the cause of this issue?
<ifreund> something to do with saved buffers and the transaction system at a guess, I haven't taken a proper look yet though
<ifreund> xwayland views are treated specially in that we don't wait for them to respond to configures before committing transactions
<novakane> I feel like I need to relearn the whole river codebase now xD
<ifreund> yeah, a lot has changed but I think it's all for the better :)
<ifreund> I'm liking the kernel-style circular linked lists a lot more than the zig std ones
<ifreund> and I'm very happy to have gotten rid of ViewStack lol
<novakane> yeah the view code seems really cleaner, I need to read more about the scene API though
<novakane> is there more doc than the comments in the scene header btw?
<ifreund> not really no
<novakane> alright that's fine, I just need to really looks at this more than a glance, beens months I say this lol
<novakane> ifreund: also do you need some help with some part of the PR that you can delegate or your prefer to do it yourself?
<ifreund> novakane: I think the best way to help out is through testing/reporting bugs, I don't want to worry about limiting scope to avoid merge conflicts until this is merged
<ifreund> thanks though
<novakane> yeah that's fair, I intend to daily use it to test, if it runs fine enough, which seems to be the case
Ordoviz has joined #river
taupiqueur has quit [Ping timeout: 260 seconds]
jao has joined #river
taupiqueur has joined #river
taupiqueur has quit [Ping timeout: 252 seconds]
jao has quit [Ping timeout: 264 seconds]
taupiqueur has joined #river
taupiqueur has quit [Ping timeout: 260 seconds]
dvzrv has quit [Ping timeout: 256 seconds]
dvzrv has joined #river
mohan43u_ has joined #river
mohan43u has quit [*.net *.split]
zxtx_ has quit [*.net *.split]
mannerism has quit [*.net *.split]
zxtx_ has joined #river
mannerism has joined #river
waleee has joined #river
<angry_vincent> what is scene graph ?
<angry_vincent> ok, thx
Ordoviz has quit [Quit: WeeChat 3.8]
<waleee> tangentially on-topic, but how's dwl's scene graph support?
<waleee> the amount of code in the pr that added it was quite miniscule
<waleee> (so I imagine it has loads of possibly unhandled edge cases)
<ifreund> no idea
<ifreund> if you're trying to compare that to the size of the river PR note that the river PR takes the opportunity to do a significant amout of other refactoring
<waleee> yeah, I noticed that you had some transaction related stuff in there
<ifreund> for example rewriting the transaction system to fix some long standing issues and clean things up
<ifreund> > (so I imagine it has loads of possibly unhandled edge cases)
<ifreund> one I can spot pretty quickly is that the ext-session-lock implementation still has this race: https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/186
<scorpion2185[m]> ifreund is is possible to add a window above waylock?
<scorpion2185[m]> to use an on screen kbd on touchscreen devices, or a built-in keyboard
<scorpion2185[m]> *is it
<ifreund> scorpion2185[m]: that's up to the compositor, river doesn't currently implement anything like that though
<scorpion2185[m]> i mean in waylock since you're the dev
<scorpion2185[m]> does wayland have compositors?
<ifreund> scorpion2185[m]: river is a wayland compositor, so is sway, so is mutter
<ifreund> and this is not up to waylock, but up to the compositor like I said
<scorpion2185[m]> when I saw compositor I though about picom
<scorpion2185[m]> waht picom does on x11 on wayland it is done by river (that it is also the WM) right?
<ifreund> pretty much
<scorpion2185[m]> is it possible to have a built-in OSK on waylock?
<ifreund> again, not up to waylock but rather the compositor
<ifreund> river doesn't support that currently and I don't have any concrete plans to change that
<scorpion2185[m]> but it the waylock app has a OSK keyboard built-in inside its window that it is part of th window, isn't that a waylock thing?
<scorpion2185[m]> sorry if i am wrong
<scorpion2185[m]> *but if
<ifreund> waylock is feature complete, I'm definitely not embedding an entirely OSK gui and implementation
<ifreund> one can have separate OSK programs on wayland using the virtual keyboard protocol, e.g. squeekboard
<scorpion2185[m]> i see but river won't support showing it above waylock so you can't unlock the screen
waleee has quit [Ping timeout: 248 seconds]
<ifreund> yep, that's right
<ifreund> as I said, not supported
<leon-p> if this turns out to be an issue in real world use cases and if I get bored enough, I may push this along: https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/134
<ifreund> I don't think that that protocol is really necessary here, I think input methods should probably be universally allowed
<leon-p> which would require a working protocol for them first, which I don't see happening anytime soon unfortunately
<leon-p> so no T9 input in river for now
jao has joined #river
dbuckley has joined #river
taupiqueur has joined #river
taupiqueur1 has joined #river
taupiqueur has quit [Ping timeout: 246 seconds]
<ifreund> leon-p, novakane: just pushed a commit that seems to fix the imperfect frames on map/unmap to me. Can either of you still reproduce?
<novakane> let me try that
<novakane> map seems good but unmap is still wrong
<ifreund> ok, I was somewhat skeptical that that had really fixed it for unmap but I was never able to reproduce it for unmap reliably in the first place...
<ifreund> thanks for testing :)
<novakane> just need to remove the ability to close views and problem solved :P
jao has quit [Ping timeout: 256 seconds]
waleee has joined #river
Szadek has joined #river
angry_vincent has quit [Remote host closed the connection]
<leon-p> ifreund: can confirm, map works, unmap still flickers
<leon-p> other than that, this feels super smooth
<leon-p> can't remember any other wayland desktop feeling this great
<ifreund> leon-p: I figured out what's wrong with unmap, about to push a fix
<leon-p> nice!
<ifreund> (it's an annoying wl_signal ordering issue thing with the wlroots scene graph)
<ifreund> see #wlroots
<ifreund> I'm not really sure why it would be much smoother than sway tbh
<ifreund> river doesn't do any dynamic memory allocation as part of the transaction system, perhaps that helps with latency
<ifreund> sway allocates quite a bit iirc
<leon-p> I haven't used sway in a long time, it probably also feels nice now, but back then resizing always had some flicker
<ifreund> pushed the fix :)
<leon-p> Will try it out as soon as I am done with this lecture
<leon-p> I think I found another bug: when resizing via pointer and the client takes more than usual to update, trying to move the view with the pointer afterwards causes weird pointer movements
<ifreund> that's pretty much the same as https://github.com/riverwm/river/issues/519 I think, which should be fixable now
<ifreund> I'll probably have a look tomorrow, off for the night though o7
<leon-p> night!
jao has joined #river