12:44 <pkap> For some reason waylock doesn't change the color anymore on my arch laptop (when typing or entering a wrong password). I think this is since I updated to the latest commit in the 0.3.x branch, but not completely sure. What could be the cause for that?

13:15 <ifreund> pkap: yeah I don't know yet, it broke for me too

13:16 <ifreund> the weird part is that waylock seems to be behaving as expected according to the WAYLAND_DEBUG log, river seems to have changed behavior

13:16 <ifreund> but I haven't figured it out yet

13:16 <ifreund> probably some wlroots breaking change I missed

13:39 <pkap> ifreund: thanks, good to know it's not just me.

13:41 <pkap> It's not a big deal for me, just wondered

13:56 <ifreund> pkap: huh, it's broken on sway too

13:57 <ifreund> I wonder if this is a wlr scene regression

15:19 <gbrlsnchs> What's the criteria for a window to be considered float? Non-zero dimensions hint? Context: writing a WM for River.

15:30 <ifreund> gbrlsnchs: the heuristic river uses currently is min-width == max width and min-height == max-height

15:31 <ifreund> pkap: the waylock thing is definitely a wlroots regression by the way

15:31 <ifreund> there's a new optimization in wlr_scene for single pixel buffers that is broken

15:31 <gbrlsnchs> ifreund: Thank you, that's good enough for my use case!

18:32 <leon-p> gbrlsnchs: you could also auto-float all windows which are child-windows. This includes f.e. file-open dialogs or tool dialogs

18:34 <gbrlsnchs> leon-p: interesting, that would be the parent event, right?