chiselfuse has quit [Remote host closed the connection]
chiselfuse has joined #foot
ayushnix has quit [Excess Flood]
ayushnix has joined #foot
xenrox has quit [Remote host closed the connection]
xenrox has joined #foot
manio has quit [Ping timeout: 264 seconds]
manio has joined #foot
Biolunar has quit [Ping timeout: 260 seconds]
Biolunar has joined #foot
<dnkl>
just merged styled+colored underlines to master. Let me know if there are any issues. If you see a crash, a 'bt full' backtrace from a debug build would be much appreciated ;)
<j`ey>
dnkl: nice work!
<dnkl>
now, let's hope someone can come up with other use cases than just red, wavy underlines...
<j`ey>
that's reason enough! .. I dont even use diagnostics in vim
<j`ey>
but spellcheck in my git commit will look nicer!
<WhyNotHugo>
Current master crashes at startup for me: err: render.c:560: BUG in cursor_colors_for_cell(): assertion failed: 'term->cursor_color.text >> 31'
<dnkl>
WhyNotHugo: haven't made any changes to that... you probably had that before styled underlines, but you're not seeing it unless you run a debug build
<dnkl>
do you use OSC sequences to set the cursor color?
<WhyNotHugo>
dnkl: Yeah, via theme.sh
<dnkl>
ok, that's the problem then. Not sure what to do about it though...
<dnkl>
the issue is that OSC-12 only sets one of the two cursor colors
<WhyNotHugo>
The same setup works on foot version: 1.17.2
<dnkl>
you may however get a different cursor *text* color now
<j`ey>
dnkl: what are the two cursor colours?
<dnkl>
before, I think a release build would use black (which was just "luck" - it was using a zeroed-out variable
<dnkl>
background and foreground. Or cursor/text color
<j`ey>
oh
<j`ey>
that was obvious :P
<dnkl>
OSC-12 only sets the block/background color, not the text color
<dnkl>
WhyNotHugo: if you don't like it, you'll have to edit your foot.ini and set the cursor color explicitly. You can still use OSC-12; you just need to ensure the cursor text color is set to black in the config
alexherbo2 has joined #foot
alexherbo2 has quit [Remote host closed the connection]
alexherbo2 has joined #foot
<WhyNotHugo>
The above fix works fine, thanks!
<WhyNotHugo>
Regarding OSC-12 colours vs foot.ini: some applications (e.g.: neomutt) reset the colours to the terminal defaults when they exit. These resets colours to the ones in foot.ini.
<WhyNotHugo>
Can I have them reset to the OSC-12 ones instead?
<WhyNotHugo>
When my environment is too bright, I switch to light mode by piping OSC-12 to all my ptys.
<WhyNotHugo>
But in reallity, I'd want foot to re-load the theme. There seems to be differences in handling between "config colours" and "OSC12 colours"
<j`ey>
same for OSC-10/11 right?
<j`ey>
or maybe not, not sure how they reset actually. (I was thinking of the `reset` command)
sentriz has quit [Remote host closed the connection]
<dnkl>
WhyNotHugo: the only way I know of to "reset" the cursor color is through OSC-12... so no, you can't have it revert to another OSC-12 color
<WhyNotHugo>
I'm not sure that it resets the colour. It reset the rest of the colour theme.
<WhyNotHugo>
Wait, I'm setting the theme with OSC 4/11, not osc12
<j`ey>
so the rest of the theme goes back to foot defaults too?
<WhyNotHugo>
the entire theme goes back to the foot.ini values, yes
<dnkl>
WhyNotHugo: same with the other color OSC. You reset colors using the same OSC sequences
<dnkl>
there's a relatively new concept of color stacks, that'd I'd like to support in foot
<dnkl>
don't remember if they're completely new sequences, it variants of the existing OSC color sequences
<WhyNotHugo>
I don't quite understand that first message there
<dnkl>
you said neovim resets the colors on exit
<dnkl>
the way to do that is through the same OSC sequences you use to customize the colors
<dnkl>
question is why neovim is touching the color palette?
<j`ey>
isnt that an alternate screen thing?
spurgis has joined #foot
<dnkl>
colors have nothing to do with the alt even
<dnkl>
screen
<j`ey>
I see
<dnkl>
the only other way to reset the colors is to do a full terminal reset. but why would neovim do that?
<j`ey>
(neomutt, not neovim)
<dnkl>
aye, phone knows about neovim and keeps auto-correcting it... sorry
<dnkl>
but then new sequences to push and pop the entire palette
spurgis has quit [Remote host closed the connection]
qyliss has quit [Remote host closed the connection]
qyliss has joined #foot
<dnkl>
I remember why I didn't go ahead and implement it before; the way the explicit stores/restores into/from explicitly specified slots. Didn't like that at all, or the fact that the specification more or less forces you to implement exactly 10 slots
<dnkl>
but I think we can tweak our implementation a bit, to be less limited. while still allowing explicit slot numbers
<dnkl>
I'll give it a try next time I have some free time
<j`ey>
you mean that you should have: x = push(); and then pop(x)
<j`ey>
such that if someone didnt clean up in the middle, you still pop to the right depth>
<dnkl>
j`ey: not cleaning up in the middle is something I consider a bug, and I don't worry about that
<dnkl>
what's worse: if you're already at slot 10, and push, that'll silently fall
<dnkl>
fail
<dnkl>
but a subsequent pop will of course succeed
<dnkl>
so, I don't want my implementation limited to just 10 slots
<dnkl>
as for the explicit slot numbers, I mostly wonder _why_ that is allowed
<dnkl>
it could be used to clean up bad stacks I suppose
<j`ey>
shell could maybe use that to clean up after a application crashes
<dnkl>
an addition could query the current slot, push, and then restore explicitly to a known slot
<dnkl>
but I don't see a reason to allow storing to an explicit slot number
<dnkl>
s/addition/application
<j`ey>
oh storing to, I thought you could just *get* slot
sh1 has quit [Ping timeout: 260 seconds]
sh1 has joined #foot
<toast>
dnkl: took a second to test the updated underlines (what's in HEAD), the 4:5 looks even better, and the 4:4 (while still having gaps) only tends to have a gap of a single dot for my config
<toast>
so seems all good :)
ninewise has quit [Remote host closed the connection]
ninewise has joined #foot
sentriz has joined #foot
alexherbo2 has quit [Remote host closed the connection]
ninewise has quit [Remote host closed the connection]
ninewise has joined #foot
<WhyNotHugo>
I used a short `print` script to test undercurls and it works. However, neovim won't render undercurls.
<WhyNotHugo>
Neovim _does_ render undercurls with alacritty.
<WhyNotHugo>
I asked in #neovim and somebody mentioned 'it uses DECRQSS to check if undercurl is supported, so maybe foot is missing something there'
<j`ey>
(hi im the somebody)
<WhyNotHugo>
Oh, hi
<dnkl>
WhyNotHugo: could be. did they have any further pointers to the exact parameters they query?
<dnkl>
toast: good to hear, thanks for testing!
<j`ey>
\x1b[0m\x1b[4:3m\x1bP$qm\x1b\\
<j`ey>
dnkl: ^
<dnkl>
j`ey: ah, yeah, we only report legacy underlines there. Should be fairly easy to fix
mvdan has quit [Remote host closed the connection]
sentriz has quit [Ping timeout: 260 seconds]
mvdan has joined #foot
sentriz has joined #foot
boomboxnation has quit [Read error: Connection reset by peer]
boomboxnation has joined #foot
sentriz has quit [Ping timeout: 256 seconds]
sentriz has joined #foot
ollysmith_ has quit [Quit: ZNC 1.9.0+deb2 - https://znc.in]
ollysmith has joined #foot
sentriz has quit [Read error: Connection reset by peer]
ollysmith has quit [Quit: ZNC 1.9.0+deb2+b1 - https://znc.in]
sentriz has joined #foot
<WhyNotHugo>
Coloured undercurls work on neovim with 64e7f2512481d33fb46b1cd3eff4b4854d634a2c
sentriz has quit [Ping timeout: 255 seconds]
<dnkl>
WhyNotHugo: great, thanks for testing!
sentriz has joined #foot
<dnkl>
WhyNotHugo: https://codeberg.org/dnkl/foot/pulls/1754 implements XTPUSHCOLORS, XTPOPCOLORS and XTREPORTCOLORS. You might be able to work that into your shell to save/restore the palette when starting/exiting an application
<dnkl>
rockorager: are there any public discussions, issues etc that are related to the in-band resize notifications, that I can follow?
<dnkl>
(I know I already asked this before, but things may have changed ;) )
<WhyNotHugo>
So basically I'd need to `xtpush; neomutt; xtpop`
<dnkl>
correct
<dnkl>
I slightly more robust sequence would be "xtpush; xtreport; neomutt, xtpop-to-reported-slot"
<dnkl>
I'm sure I'm off-by-one in one or more places above, but you get the idea...
<dnkl>
but I think a simple "xtpush;;xtpop" will work for most applications
<rockorager>
dnkl: No, nothing beyond that spec...there was a small discussion in #alacritty and i've asked mitchellh to provide commentary directly on github so we can keep it all there
novenary has quit [Remote host closed the connection]
novenary has joined #foot
dustinm` has quit [Quit: Leaving]
<dnkl>
anyone ever seen XTPUSHSGR/XTPOPSGR being used (*not* the same as XTPUSHCOLORS/XTPOPCOLORS)?
<dnkl>
how do applications detect if they're supported or not? (this question applies to XTPUSHCOLORS/XTPOPCOLORS too...)
dustinm` has joined #foot
<dnkl>
well, I'm stupid... XTPUSHCOLOR support is detected by trying XTREPORTCOLORS and checking for a reply
alebastr_ has quit [Quit: WeeChat 4.2.1]
sevz has quit [Quit: WeeChat 4.3.2]
sevz has joined #foot
Arsen has quit [Quit: Quit.]
anarcat has quit [Quit: rebooting]
Arsen has joined #foot
Arsen is now known as Guest702
Guest702 has quit [Changing host]
Guest702 has joined #foot
Guest702 has quit [Quit: Quit.]
Arsen has joined #foot
sterni has quit []
sterni has joined #foot
alebastr has joined #foot
emersion has quit [Remote host closed the connection]