11:48 <mekeor> hi! i'd like to take a screenshot of a single view ("window"). i found a snippet that uses lswt to determine the position/region of the view that is then passed to grim. is this still state of the art? https://codeberg.org/JustineSmithies/river-dotfiles/src/commit/fc1287ac02fd52d20cc99b20370d490cabd89ead/.config/river/init#L141

11:49 <mekeor> (in particular i was wondering if i can avoid lswt and jq.)

12:11 <leon-p> mekeor: that snippet does not determine the geometry of the focused window, lswt never exposed that info because none of the wayland protocols expose it

12:11 <leon-p> in general, taking a screenshot of a single window by capturing the entire scene and cropping it is very ugly and hacky

12:11 <leon-p> there will be support in wlroots for directly grabbing a screen-copy of a single window

12:11 <leon-p> in the future

12:12 <leon-p> which has the advantage of also being able to capture windows which are hidden

12:13 <leon-p> you can use grim with slurp to manually select a screen region in the meantime

12:13 <mekeor> leon-p: oh, i just tried out the snippet and it does not work because there are no .outputs properties

12:13 <mekeor> alright, thank you, leon-p :)

12:14 <leon-p> that's true as well, I removed that property since the newer protocol for listing windows no longer exposes it :)

12:14 <leon-p> I might add it back if it turns up again, but those protocols are still in the nit-picky bike-shedding phase

12:16 <mekeor> i was expecting this to work because video-capturing single views (via desktop portal) already works on my stack :D

12:16 <mekeor> i'm still content with having moved to wayland and river; this is just my second non-functional edge-case :)

12:18 <LarstiQ> https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/4545 got recently merged fwiw

12:21 <mekeor> ah cool, it implements this https://wayland.app/protocols/ext-image-copy-capture-v1

12:24 <mekeor> would the roadmap now be like this?: implement ext-image-copy-capture-v1 in zig bindings to wlroots; implement the feature in river; implement support for the protocol in grim

12:25 <vyivel> there's one extra step before all this: add toplevel source support in wlroots

12:32 <LarstiQ> vyivel: did I understand correctly that needs a bit more agreement than where we are right now?

12:34 <vyivel> not sure what you mean by agreement here but wlroots has logic for capturing outputs and needs logic for capturing toplevels

13:44 <LarstiQ> right, and my impression was that it was considered out of scope for the previous PR so that it wouldn't be held up by more back and forth on toplevel

13:44 <LarstiQ> but very much from the sidelines so take that with a heaping of salt

17:32 <lbiaggi> Hello, how can I tell river to spawn a process on a certain monitor?

17:58 <ifreund> lbiaggi: you can use an output rule, see the rules section in the riverctl man page

17:59 <lbiaggi> ifreund: hum, okay! thanks