08:22 <i9j8k7> hi, what is the equivalent of sway "shortcut_inihibitor disable"?

08:28 <NickH> What does sway "shortcut_inihibitor disable" actually do

08:28 <NickH> ?

08:31 <i9j8k7> NickH: it's for applications the inhibit the compositor shortcuts and start grabbing everything, like qemu. In sway you can say for example that a view with such an app-id should never be allowed to do that

08:31 <NickH> I'm not aware of such a feature in river

08:37 <i9j8k7> I see, I need it because unfortunately you can't disable it in qemu itself, and I don't want to press qemu ungrab shortcut every time I end in a qemu window in order to use compositor shortcuts again

09:14 <ifreund> i9j8k7: it's not possible for applications to inhibit compositors shortcuts in the first place in river

09:14 <ifreund> if you need functionality like that, a near-empty "passthrough" mode is a good way to implement it

09:54 <i9j8k7> ifreund: great, thank you

13:55 <adamcstephens> I'm seeing bugs around window border rendering on master. For example terminals open with a border that's too small and doesn't surround the entire window, vscode doesn't have a border at all. Should I file an issue about these?

14:00 <adamcstephens> maybe related, but not exactly what #807 is reporting

14:02 <tiosgz> adamcstephens: if something doesn't have a border at all, i'd suppose it's got client-side decorations. in that case `riverctl rule-add ssd` should fix that (as river no longer enforces ssd everywhere)

14:02 <tiosgz> wrt wrong border size, what's the particular program?

14:04 <adamcstephens> i can reproduce the issue with both kitty and wezterm

14:05 <adamcstephens> for example on an empty tag, i open wezterm and receive a border box in the upper left hand corner. the window fills the screen though, so content will overflow that box obviously. clicking outside the box, adding more windows and rearranging, will fix the border

14:08 <adamcstephens> here's a couple screenshots https://imgur.com/a/RdkpR6h

14:09 <tiosgz> i can't reproduce that myself with kitty... what you could do is grab a WAYLAND_DEBUG=client log and add that to the issue

14:44 <tiosgz> in case of wezterm, i can reproduce it. a supposedly atomic change is spread across multiple commit()s, of course with ack_configure coming first (i.e. it's the same client bug as with imv)

14:45 <tiosgz> adamcstephens ^

15:21 <i9j8k7> is it possible to map a key to more than one command without using spawn and multiple nested riverctl calls?

15:22 <i9j8k7> I guess not, but just making sure

16:10 <leon-p> adamcstephens, tiosgz: it's a known bug that river side view size and client side view size may de-sync in some instances. The fix is either to always defer to the client side (I'd vote for this one) or to always user server side and clip the surface if it's too large (arguably bad UX in a lot of cases)