<angry_vincent>
i did read this several times but cannot connect real example in pastebin i showed. there are some more riverctl spawn commands before keepassxc one. does scratchpad work only for last application spawned?
<angry_vincent>
maybe user did Super+Shift+P for that ?
n0r has joined #river
<fleischie>
The manpage `riverctl(1)` says for "set-view-tags": "Assign the currently focused view ..."
<fleischie>
And as in the pastebin you posted keepass is the last application to be started I would assume it was focused.
<fleischie>
So yeah, "does scratchpad work only for last application spawned" <- only if you do 1) create application 2) immediately call `riverctl set-view-tags ${scratch_tag}`.
<fleischie>
But I just learned about scratchpad. :shrug:
<angry_vincent>
thank you. so my guess was correct
Nulo has quit [Remote host closed the connection]
manithree has joined #river
manithree has quit [Quit: Client closed]
<tleydxdy[m]>
ifreund hmm I'm not sure if it's good or bad, but with the error check fix you did yesterday, setting adaptive sync still doesn't report any error
<tleydxdy[m]>
oops
<tleydxdy[m]>
1 sec
<tleydxdy[m]>
wth, so now setting adaptive sync works again
<tleydxdy[m]>
that's why there's no errors
groknull has joined #river
groknull has quit [Remote host closed the connection]
jao has joined #river
groknull has joined #river
groknull has quit [Remote host closed the connection]
groknull has joined #river
groknull has quit [Remote host closed the connection]
Nulo has joined #river
<edrex[m]>
i'd like a command to "fake fullscreen", ie tell the application it's in fullscreen mode so it doesn't render toolbars etc while keeping the window size the same.
<edrex[m]>
It's really nice for browser-based video players.
<edrex[m]>
imo coupling the two makes a lot less sense in a tiling shell
<edrex[m]>
thoughts?
<tleydxdy[m]>
should be relatively easy to do
<leon-p>
technically a fake fullscreen thing would be a violation of the xdg shell
<leon-p>
also, since the view is not actually fullscreen, it may get resized, which the client may not expect
<edrex[m]>
it's true. i haven't had any trouble with it testing with chromium, firefox, qutebrowser over the years.
<leon-p>
and, since the only use case is working around bad UI in some applications, I am inclined to repeat that I don't think fixing other programs bugs is our responsibility
<edrex[m]>
i think the spec is float biased :D
<leon-p>
true
<leon-p>
still, my vote goes to following it to the letter
<tleydxdy[m]>
some floating wm also have fake fullscreen iirc
<leon-p>
you can get around firefox UI issues with a custom user chrome
angry_vincent has quit [Remote host closed the connection]
<leon-p>
so technically this is already solvable by more idiomatic / canonical means
<edrex[m]>
i tend to think of the two states (internal client UI "distraction free mode" and window fullscreen) as distinct axes that have just been glommed together by the dominant big chunky ui shells made for grandma
<edrex[m]>
i will patch if it doesn't go in. would rather not tho.
<tleydxdy[m]>
to keep within the spec, the compositor could just have a virtual output right?
<edrex[m]>
usability is primary, for me. fuck a bunch of specs if they don't serve my use cases
<tleydxdy[m]>
output can resize too, so app ought to handle it
<edrex[m]>
i spend a lot of time plumbing state between layers of system architecture that aren't "supposed" to talk to each other because it's necessary for some UX need.
<edrex[m]>
leon-p: i've been using stacktile for a few weeks now, and I'm wondering how you use the three areas (primary, secondary, remainder). Also, have you thought about introducing a layout command to resize the current area, rather than pri or sec? Running out of keys :)
<leon-p>
edrex[m]: if you check the master branch instead of the old C version, you'll find that stacktile now supports an arbitrary amount of areas
<leon-p>
I don't resize any of them though
<leon-p>
as for the spec thing, _everything_ has a lower priority than compliance. Not doing what the protocol says is a bug, and introducing a bug on purpose is a big no no. hyprland is setting a /really/ bad precedent here, same as with the workspace protocol issue
<leon-p>
I get not liking browser UIs, I don't either, but this is not the way to fix it
<edrex[m]>
leon-p: alright, i'll try it now that llvm15 is landed in nixpkgs. is the zig version at all close to an initial release?
<leon-p>
it's pretty different, but it's the same in spirit and it's all documented
<edrex[m]>
are there area resize commands implemented in it? i will look anyway
<leon-p>
areas are not resizable, except for the main one
<leon-p>
I found that I never resized anything except for the main size, so I did not bother implementing it. Also I don't think it fits with the new design really
<leon-p>
btw, this is probably a good time to say that I'll likely won't maintain stacktile after river has a more powerful layout protocol. I want to try something new, something less traditional.