16:08 <ifreund> novakane: I'm not sure what's up with firefox extensions popups with regards to damage

16:08 <ifreund> I took a look with wlhax and they are a subsurface of an xdg-popup, which we should be handling correctly

16:12 <novakane> yeah that's weird I don't understand too

16:13 <novakane> It's not even laggy it's doesnt render text if I don't move my mouse

16:20 <novakane> I also try librewolf and waterfox, based on firefox, librewolf has the same but problem but not waterfox

16:20 <novakane> and waterfox is based on a less recent version of firefox iirc

16:21 <ifreund> is waterfox even wayland-native?

16:23 <novakane> hmm good question I didn't thought of this

16:25 <novakane> oops close river instead of river in river...

16:26 <novakane> looks like waterfox is wayland native so that explain that

16:26 <ifreund> is or isn't?

16:26 <novakane> isn't sorry

16:27 <ifreund> no worries :D

16:27 <ifreund> The issue is almost certianly due to a missing damage call somewhere, i just don't see where

16:31 <novakane> I don't see where too for now :/

16:48 <leon-p> ifreund: managed to crash river. I have an application on the float filter that does something fishy (haven't debugged that yet) that causes the application to not appear. Trying to move or resize it via pointer either causes weird pointer movement or a reaching unreachable. Resizing by keyboard works.

16:51 <ifreund> interesting, I thought I put both of those on the same code path for the actual resize

16:52 <ifreund> oh no, that was for moving

16:53 <ifreund> we should probably do the same thing there and factor the common logic and critical bounds checks into a View.resize() functions

16:53 <ifreund> s/bounds checks/clamping/

16:53 <leon-p> probably. But before that I'll figure out what exactly is happening

16:56 <ifreund> I wonder if using signed integers for width/height everywhere would make this less error-prone

16:59 <leon-p> What happens: Application creates view, river sends toplevel configure (size 0x0), river sends xdg surface configure, application renders, river sends toplevel configure (same size as the buffer already is), river sends xdg surface configure, application does _not_ render because the size did not change.

16:59 <leon-p> That makes river unhappy somehow

17:00 <leon-p> if I change it to always render on xdg surface configure, it works

18:58 <novakane> leon-p: oh you made a layout generator

19:08 <novakane> I think you makefile install/clean is wrong though

19:25 <leon-p> probably

19:26 <leon-p> It is nothing more than a few modifications to contrib/layout.c with the setup I used to test it

19:27 <novakane> oh ok you just didn't update the name in makefile then

19:27 <novakane> well I never try it, I should one day

19:27 <leon-p> the idea of the layout is that having too many windows should not cause small window sizes, hence the "card deck" and that I like to have three windows side by side on my ultra wide screen, hence the main area being devided into columns instead of rows

19:28 <leon-p> oh, and my gaps are a bit smarter than rivertiles. They are actual gaps instead of just padding

19:28 <leon-p> but that's it, really nothing all that special

19:29 <novakane> oh yeah makes sense, probably not super useful on my small screen then

19:29 <novakane> I rarely open more than 3 windows on a tag

19:29 <leon-p> yeah, not super useful then :P

19:31 <novakane> 3 columns on a 13" screen... yeah no :P