14:25 <ccha> hello, do you know why sometimes there are 2 cursors? There is a true one and another one at top left corner not moving but replicate the same image as the true one.

14:30 <FireFly> that's not anything I've come across myself, but maybe inspecting `riverctl list-inputs` helps provide clues? I know wayland allows multiple independent mice in theory so maybe there's somehow a virtual one created for whatever reason on your system?

14:34 <ifreund> ccha: interesting, I haven't heard that one before

14:34 <ifreund> does it happen if you set WLR_NO_HARDWARE_CURSORS=1 before starting river?

14:35 <ifreund> FireFly: wayland allows multiple seats yes but only one pointer device per seat. also river only supports a single seat currently

14:37 <FireFly> ah okay

14:38 <ifreund> river will display separate cursors for tablet tools, but it doesn't seem like that's what's happening here

16:02 <lordmzte> In that same update that made river lose output configs after suspend, which I've fixed with kanshi, rivertile also started losing the `main-location` setting. Any way to fix that one?

16:55 <ccha> ifreund: with the variable, there is no issue. I used this variable when I run river in virtual machine, but I'm using on host. The double cursor happens when I start brave browser,not when I start river. And the topleft cursor is present on all tags. Then if I put the browser fullscreen, the topleft cursor disappears, and come back if I open new tab or close tab.

16:57 <ccha> the issue started with 0.3.10 I think. Before there wasn't this issue

16:58 <leon-p> lordmzte: not via configuration, since it too sees the output as completely gone now. You could add a list of recent output identifiers plus their set variables, if you wanted

17:03 <ccha> I just tested it, 0.3.11 and 0.3.10 there are 2 cursors and version 0.3.9 no 2 cusors

19:16 <lordmzte> leon-p: I see. I still overall haven't understood how this change isn't just a step backwards.

21:19 <ifreund> ccha: seems like a possible wlroots regression I guess