00:47 <soonToBeExHyprla> is anyone else getting some weird flickers on the terminals? i tried kitty and foot, both had the same issue. firefox seemed okay. it happens on both monitors, more frequently on the one with a higher refresh rate. nvidia gpu. no prior issues of this kind on other wayland compositors.

00:53 <adamcstephens> what version of River and wlroots? nvidia gpu is always a red flag for potential problems…

01:06 <soonToBeExHyprla> river 0.2.6-1 wlroots 0.17.1-1 . and yeah, i know nvidia sucks.. i hope there's a fix because i deeply miss tags

01:13 <adamcstephens> you could try the river master branch, as it has changed quite a bit

02:16 <soonToBeExHyprla> tried the same version on my amd thinkpad and its working perfectly

08:59 <Guest66> hi

14:45 <ifreund> man, tablets are way complicated

14:58 <LarstiQ> ifreund: how so?

14:58 <ifreund> LarstiQ: have a look at the tablet-v2 protocol if you're curious. Way more moving parts than a mouse or touch input for example

15:00 <ifreund> I'm making progress on the river implementation but it requires some hacks

15:00 <ifreund> for example, wlr_cursor doesn't really support what's needed to do things right

15:01 <ifreund> (sway for example uses the same cursor for the pointer and all tablet tools, which is bad UX IMO, mutter gets this more right)

15:02 <ifreund> relevant wlroots MR: https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/3974

15:02 <LarstiQ> ifreund: will do. I can imagine multi-touch at least making things more difficult

18:12 <adamcstephens> good to w

18:12 <adamcstephens> oops, ww

19:01 <leon-p> I wonder if we can just spawn a nre seat on demand if a tablet tool appears

19:01 <leon-p> *new

21:06 <ifreund> leon-p: no need for a new seat, we can have multiple cursors for a single seat

21:07 <ifreund> one pointer cursor and one transient cursor per tablet tool

21:58 <pkap> Hi, is there a character limit for float rules or something like that? I cannot make the following rule work `riverctl rule-add -title "Extension: (Bitwarden - Free Password Manager) - Bitwarden — Mozilla Firefox" float`

22:01 <pkap> It works with `-app-id firefox` but that makes all firefox windows floating of course.

22:22 <waleee> have you verified that it's the exact string used in the title?

22:24 <waleee> I wouldn't be surprised if there's some font/char trickery like using the longer "-"

22:26 <pkap> Yes, the title is copy/pasted from the output of lswt.

22:28 <pkap> Hm not sure. I remember having a rule this (with the long `-`) and it used to work. `riverctl rule-add -title "Firefox — Sharing Indicator" float

22:28 <pkap> `

22:29 <pkap> But does sharing indicators are not present anymore in firefox so can not not really tell. Just that the rule worked some years back.

22:31 <pkap> So handling the longer `-` may have changed since then..

22:33 <pkap> Ah this was before the globbing feature. So this could be related to introducing the globbing.

22:53 <waleee> pkap: I was mostly thinking about if you had manually typed the title, the strings wouldn't be indentical if you had used - instead of —

22:54 <leon-p> ifreund: I was under the assumption that there is only ever a single wl_pointer per wl_seat. Is that one swapped between the active devices? How would that work when the user moves both mouse and tablet pen simultaneously?

22:57 <leon-p> pkap, waleee: firefox sometimes changes title and app-id after mapping the toplevel, so rules trying to target firefox often dont work.

22:58 <waleee> ah the dwarf fortress kind of *fun*

22:59 <pkap> Ah I see. Nothing we can do then

23:01 <leon-p> this also sometimes changes between firefox versions, foradded spice

23:02 <leon-p> I have a rule to make the little video pop up thing floating and after every update its a gamble whether it will actually float

23:06 <ifreund> leon-p: cursor != pointer

23:06 <leon-p> I know, was just wondering what the client sees

23:06 <ifreund> by cursor I mean just the graphical thing that you move around on the screen

23:07 <ifreund> the client see a zwp_tablet_tool_v2

23:07 <leon-p> oh right, pointer emulation is optional, forgot that

23:07 <ifreund> If your asking about pointer emulation, that would indeed use the same cursor as the pointer

23:08 <leon-p> well yeah, then it's just putting textures on screen and nothing magic, I see