12:44 <Nulo> For some reason all GTK apps running in Wayland show a CSD even when the app doesn't have functionality in it (Inkscape, Claws Mail); can I prevent that?

12:48 <ifreund> Nulo: you can get gtk to merge this: https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/2191

12:48 <ifreund> though it likely needs a bit more work, I don't have any time for it at the momement though

12:49 <Nulo> Ugh

12:53 <ifreund> alternative fork river and add support for the kde ssd protocol :/

12:53 <Nulo> Yeah no

12:53 <ifreund> I'm avoiding implementing it out of idealism even though it would be trivial with wlroots

12:53 <Nulo> Installing the patch would be recompiling GTK with it right?

12:54 <Nulo> If it would be easy we could implement it until GTK gets their shit together

12:54 <Nulo> > inb4 river patches and releases faster than GTK merging an months-old patch

12:55 <ifreund> Like I said, I won't accept kde's ssd protocol in upstream river out of idealism

12:55 <ifreund> but river is open source so anyone is free to patch it as they like

12:56 <Nulo> Fair enough

12:56 <Nulo> Then I'll probably patch GTK

12:56 <Nulo> With yours

12:59 <ifreund> Nulo: let me know if you notice any bugs using the patch with gtk, I've barely tested the current version

12:59 <ifreund> the wayland side of the patch is solid, but I'm not at all familiar with the GTK side of the code

13:01 <Nulo> Love how no one in GNOME is interested in making a patch for a standard protocol

13:04 <ifreund> as far as they're concerned, everyone should use CSD for everything

13:05 <Nulo> Oh also, a similar thing happens on Telegram-Desktop (Qt) but it's small so I don't mind

13:06 <ifreund> hmm, I don't really know what QT's store with SSD vs CSD is

13:06 <ifreund> s/store/story/

13:07 <Nulo> Keep in mind Telegram is a pretty special Qt app

13:07 <Nulo> Using the Telegram option to "use native window" (as opposed to CSD) shows a Qt CSD

13:08 <Nulo> https://0x0.st/-yp_.png

13:08 <dnkl> Nulo: do you have QT_WAYLAND_DISABLE_WINDOWDECORATION=1?

13:09 <Nulo> Nope, using that + "Use native window" fixes it, thanks

13:10 <ifreund> that would be a lot easier to implement for GTK tbh

13:10 <ifreund> the part of the gtk patch that probably doesn't work right is switching csd/ssd dynamically at runtime

13:11 <ifreund> the other problem is that I hate working with gtk's code

15:08 <snakedye> For me Telegram's titlebar is integrated in the app, kinda like Nautilus

15:09 <Nulo> snakedye: Yes, exactly. Using QT_WAYLAND_DISABLE_WINDOWDECORATION=1 fixed it

15:09 <Nulo> *and setting native window

15:14 <snakedye> I suppose I have this flag enabled by default also screen sharing on it worked ootb which was a pleasant surprise

16:38 <novakane> damn I finally took the time to implement true gaps in rivercarro, I spend waaay too much time on this, enough maths for 3 months :P

16:48 <ifreund> what exactly are "true gaps"?

17:00 <novakane> ifreund: like stacktile, instead of a padding around the view it's just a gaps between 2 views, so if I set outer-padding to 0 I have no gaps between a view and the screen edges

17:02 <novakane> ifreund: https://0x0.st/-yOz.png

17:03 <ifreund> ok so innner gaps instead of my lazy padding

17:03 <ifreund> padding does happen to be nice and simple though :D

17:04 <ifreund> working on a review of xdg-activation by the way

17:06 <novakane> ifreund: yep sure, although on small screen I like to have gaps to separate views but gain a bit of place with no outer gaps

17:06 <novakane> oh cool, stressful :P

17:06 <ifreund> I don't want it to be stressful, contributing should be as fun as possible

17:07 <ifreund> I do have really high quality standards though

17:08 <Nulo> What's the monocle view?

17:09 <Nulo> Also novakane what's the font on that screenshot?

17:17 <novakane> ifreund: no worries, I'm just joking, I never feel attacked here, I'm good with high quality standard, I know what I do is far from perfect but good reviews is great to learn :)

17:18 <novakane> Nulo: view that take all the usable screen, and the font https://github.com/rubjo/victor-mono

17:18 <Nulo> Nice, thanks

17:23 <ifreund> novakane: great, review is up :)

17:26 <novakane> great, I have a bit of time to look at it now

17:29 <Nulo> Tried stacktile and hit into a pretty ugly bug :(

18:07 <leon-p> Nulo: If you tell me what the bug is, I can fix it

18:12 <Nulo> leon-p: Setting outer-padding to 0 breaks in weird ways: no space is reserver for yambar and views are sticked to the right border instead of having inner padding to them (or maybe that's intended? but there's padding on the left, top and bottom)

18:52 <leon-p> Nulo: ah, I discovered that smae bug as well yesterday

18:52 <Nulo> :D

18:52 <leon-p> Actually I fixed it already once, but apparently it snuck in again

18:52 <leon-p> I'll give it a look later

19:18 <dagle> Hmmm. I have noticed that sometimes the tags in waybar+river lags behind, strange.

19:19 <dagle> Like I'm tag 3 but river thinks I'm on tag 2 for 1 sec.

19:21 <ifreund> dagle: hopefully just a few ms not a full second, but yeah that's racy currently

19:23 <ifreund> I hope to figure out a good way to expose the double-buffering of river's state to clients to allow for frame perfection there in the future, but it's a hard problem

19:23 <dagle> Idk, how long it is but long enough to notice it. So I would say more than a couple of ms.

19:23 <ifreund> by few a I mean a few hundred I guess

20:07 <ifreund> hmm, probably should've structed the river-status implementation a bit differently in retrospect

20:08 <ifreund> there's no reason to allocate a separate struct containing just a pointer to the output and to the resource for each resource

20:08 <ifreund> that was only the second protocol I implement server-side though, so I have an excuse for why it's bad

20:09 <ifreund> not worth fixing till after the xdg-activation PR though

23:31 <Nulo> dnkl: Just witnessed yambar go up to 100% CPU for not reason, not being able to be killed with a normal killall

23:32 <Nulo> Killing pulseaudio while running causes it.