<dnkl>
rockorager: looks mostly good. "screen" really refers to the screen/output/monitor? Curious what the use case is for that...?
<rockorager>
I am going to remove it
<rockorager>
Removed :)
<rockorager>
I had no use case for those - I just assumed it may have some use somewhere but the first two immediate comments I got were about that haha
<dnkl>
rockorager: might be a good idea to explicitly mention that "text area" does *not* include padding etc
<rockorager>
dnkl: wow you're fast - I haven't even gotten this into my libs yet
<dnkl>
got lucky and had a few spare minutes...
<rockorager>
I'll test it out a bit later today
an3223 has quit [Ping timeout: 260 seconds]
<dnkl>
rockorager: one potential issue, or pain point, for applications using the new size notifications is how to couple pixel and char notifications. Might be even more painful if there are a lot of un-throttled events during an interactive resize
an3223 has joined #foot
<rockorager>
dnkl: Yeah the throttling could be an issue, but I don't think it would be any different from SIGWINCH?
<rockorager>
And yeah, two separate notifications for pixels and characters could be problematic if there are images on the screen
<dnkl>
if you use the ioctl to read the sized you
<dnkl>
size, you get both pixels and chars in one
<dnkl>
gut feeling is a new reply, that combines the two mode would be better
<rockorager>
Ah that's a solution I hadn't thought of - you could use the new notification to trigger an ioctl call
<dnkl>
that's not what I meant though
<rockorager>
I think it could be handled on the application side fairly easily though
<rockorager>
If I subscribe to both, I can wait until I receive both in my read loop
<dnkl>
it'd be nice with a platform agnostic way to get the size
<rockorager>
before taking any action
<rockorager>
yeah
<dnkl>
right, but spec doesn't define the order, making this a bit more difficult for the application
<rockorager>
CSI n ; chars ; chars ; pixels ; pixels t
<rockorager>
I don't know what n should be but something like that would work
<dnkl>
yeah, that's what I imagined
<rockorager>
It could reduce to a single mode then
<rockorager>
if the terminal can't report pixels, they are 0
<dnkl>
yeah, might not make sense to have two modes
<dnkl>
unless you think more applications are likely to adopt the new mode(s) if we use the legacy replies, I think an all new reply format is better
<rockorager>
I don't think anyone would care - the new one is just as simple
<dnkl>
all new == CSI n ... t, as you described above
<dnkl>
then that has my vote:)
<rockorager>
I tend to agree with you on that one
diaspora3 has quit [Quit: Ping timeout (120 seconds)]
<rockorager>
fwiw alacritty has rejected this idea
diaspora3 has joined #foot
<dnkl>
they typically do, but they also tend to adopt things later, once they've become a bit more mainstream
<rockorager>
Yeah
<rockorager>
Alright let's do the single mode, single reply. I will try to find a nonconflicting value for the xtwinops response
<dnkl>
I do think it would be nice if we could be sure at least _some_ applications (that aren't too exotic...) will support the new mode(s)
<rockorager>
I will put support for this in my tui libraries, so that'll get aerc, senpai, and a few WIP zig things
<dnkl>
senpai I do use, so that's a plus point 😂
<rockorager>
I've also discussed a little with gpanders on nvim core
<dnkl>
was just going to mention neovim... would be nice if we could get them on the train
<rockorager>
Yeah
<rockorager>
gpanders is putting a lot of newer stuff into nvim so I think totally doable. OSC 8 as of 0.10 and I know he's working on mouse shapes as well
<dnkl>
is there a PR for OSC-8?
<dnkl>
one thing a lot of apps forget is to put unique IDs on the URLs...
<dnkl>
be nice if nvim could get that right from the beginning
<dnkl>
that'll break URLs, that line-wrap across panes
<dnkl>
and may also break URLs if part of an OSC-8 link is updated, at least on some terminals
<dnkl>
(foot included)
talismanick has joined #foot
an3223 has quit [Remote host closed the connection]
talismanick has quit [Ping timeout: 268 seconds]
talismanick has joined #foot
talismanick is now known as Guest4674
Guest4674 has quit [Client Quit]
talismanick has joined #foot
hexa- has quit [Quit: WeeChat 4.2.2]
hexa- has joined #foot
talismanick has quit [Ping timeout: 255 seconds]
an3223 has joined #foot
an3223 has quit [Remote host closed the connection]
an3223 has joined #foot
an3223 has quit [Client Quit]
<rockorager>
dnkl: how about `CSI 48 ; charh ; charw ; pixh ; pixw t`? 48 because it's just 4 and 8 jammed together
<jmcantrell>
foot doesn't seem to recognize title updates anymore. tested with vim (:set title), bash and zsh. using foot version: 1.17.2 +pgo +ime +graphemes -assertions