an3223 has quit [Remote host closed the connection]
an3223 has joined #foot
radu242407821 has quit [Ping timeout: 240 seconds]
hexa- has joined #foot
radu242407821 has joined #foot
radu242407821 has quit [Ping timeout: 240 seconds]
Oleg has joined #foot
Oleg has quit [Client Quit]
Oleg has joined #foot
Oleg has quit [Client Quit]
radu242407821 has joined #foot
radu242407821 has quit [Ping timeout: 256 seconds]
Oleg has joined #foot
Oleg has quit [Quit: WeeChat 3.4]
radu242407821 has joined #foot
radu242407821 has quit [Ping timeout: 256 seconds]
radu242407821 has joined #foot
radu242407821 has quit [Ping timeout: 240 seconds]
an3223 has quit [Remote host closed the connection]
an3223 has joined #foot
radu242407821 has joined #foot
radu242407821 has quit [Ping timeout: 240 seconds]
<Nuc1eoN>
Oh indeed. My mistake sorry.
<Nuc1eoN>
Btw, is the scroll wheel not supposed to be able to drag with the mouse?
<Nuc1eoN>
*scroll bar
<dnkl>
Nuc1eoN: yes, it's not intended as a scroll bar. It's only an indication of where in the scrollback buffer you are. Its appearance can be configured (or disabled)
<Nuc1eoN>
dnkl, Ok that's what I suspected. I guess it would be nice to have a scroll bar optionally (for big amounts of text it's better), but Im fine with how it is
<Nuc1eoN>
One other thing I have noticed though, is that foot's glyphs rendering is not as crisp as on alacritty. Also it is clipped.
<dnkl>
Font fuzziness most likely comes down to different fontconfig settings being applied, as both alacritty and foot uses freetype to rasterize the glyphs
<dnkl>
Are you using a bitmap font?
<dnkl>
And make sure to use the same *pixel* size when comparing, as otherwise antialiasing (especially sub-pixel antialiasing) can look different.
<dnkl>
(not that I know how to configure a specific pixel size in alacritty...)
radu242407821 has joined #foot
radu242407821 has quit [Ping timeout: 240 seconds]
<manio>
Nuc1eoN: fwiw i had to enable grapheme-width-method=double-width in the [tweak] section to have same glyphs as in alacritty
<Nuc1eoN>
I am on the most recent version
<Nuc1eoN>
dknl normally yes, but only to proof the point here I switched to Meslo Nerd font here
<Nuc1eoN>
*dnkl
<Nuc1eoN>
dnkl, I can look into it again, but accorind to my tests so far alacritty was crisp in amy size, while foot was fuzzy in any size
<Nuc1eoN>
manio, thank you I will look into that option when Im back at the pc
<dnkl>
Nuc1eoN: imgur on phone sucks, so can't check your screenshots atm. But foot will scale bitmap fonts if you've enabled that in fontconfig. Not saying that's your problem, but worth looking into
<dnkl>
Could also be sub-pixel antialiasing vs. regular antialiasing
<Nuc1eoN>
Nah for this comparison I explicitely used vector fonts
<Nuc1eoN>
dknl could be
<dnkl>
Foot has support for Wayland specific protocols for communicating sub-pixel settings
<Nuc1eoN>
dnkl lmao
<Nuc1eoN>
Hm it seems a little bit compressed sadly (i should have used webp maybe) but this link should be better on mobile https://i.imgur.com/Si7hTX2.png
<dnkl>
Re. Clipping; it's usually not a problem anymore, but can happen in don't cases. For example, when the "wcwidth" of the character is 1, but the glyph itself is double-width *and* the fallback font is"too large"
<Nuc1eoN>
*shouldnt
<Nuc1eoN>
Yes but in this case I only used 1 font
<dnkl>
Foot supports overflowing glyphs into the neighboring cell. But in the example above, the glyph would need overflowing into two cells. And gets clipped
<dnkl>
Still can't see the screenshots. Maybe try 0x0.st instead of imgur
<dnkl>
?
<dnkl>
In any case, what manio suggests could work, because it changes how foot allocates cells. Instead of using wcwidth(), foot will always allocate 2 cells. This'll likely match the actual glyph width better
<dnkl>
But you may instead see cursor desynchronization issues, depending on where (i.e in which applications) those glyphs are used
<dnkl>
Btw, when you say "1 font", the emojis are must likely fetched from a fontconfig provided fallback font. Or are you sure they're in the primary font?
<Nuc1eoN>
dnkl, it is nerd font patched, so all fonts and emojis are included
<sochotnicky>
every once in a while I consider trying some new font on my desktop and then I see things like these and decide and that Source Code Pro works for me so why bother :-D
<dnkl>
Nuc1eoN: thanks, yes. That looks like the issue(s) linked above
<Nuc1eoN>
manio's tweak did not help sadly
<dnkl>
That's expected now that we know what the issue is
<dnkl>
Sadly, there's no workaround, and no plans to support this
<Nuc1eoN>
True, in my case the very left ones are also clipped I'll give u that :p
<Nuc1eoN>
dnkl, I switched to a more sane font now and my problems are fixed :p
<Nuc1eoN>
Thanks a lot
<Nuc1eoN>
What and unexpected ride
<dnkl>
Nuc1eoN: np :)
<Nuc1eoN>
dnkl, btw while fiddling with alacritty I have noticed that its changes (even font!) are immediately applied after saving the config file, kinda cool. Would that also be viable for foot? :P
<dnkl>
Nuc1eoN: sorry, no, config reload is not on the roadmap.. :)
<dnkl>
But you might want to use "foot -f fontname" while testing different font and sizes
<Nuc1eoN>
ok nice :)
<dnkl>
Nuc1eoN: and for future experiments, the more generic "-o" option
<Nuc1eoN>
Ok I will test that. Why is hot reload not planned? Preference reasons or technical? I think if I had an option to enable it I would do it :p
<Nuc1eoN>
Seems more common on wayland. Wayfire does it, and I was kinda concerned, but I started to like it being immediate :p
<dnkl>
Nuc1eoN: it adds complexity to the code base. It could obviously be done, but the ROI (hey, I do buzz words too!) value isn't worth the complexity, in my opinion.
<Nuc1eoN>
understood
<sochotnicky>
thumbs up on simplicity - code rots and maintenance burden adds up over time...
amk has quit [Remote host closed the connection]
amk has joined #foot
pranjal has quit [Ping timeout: 256 seconds]
pranjal has joined #foot
pranjal has quit [Ping timeout: 256 seconds]
pranjal has joined #foot
manio has quit [Ping timeout: 250 seconds]
manio has joined #foot
<mvdan>
I just wasted an embarassing amount of time debugging "command not found" errors on Bash which were ultimately caused by non-printable characters hidden around my spaces. I was only able to spot those via "cat -v". Is foot able to make those always printable?
<mvdan>
the characters in question were `M-oM-;M-?` as reported by `cat -v`
<mvdan>
I tried searching the issue tracker and man page but found nothing with the keyword "printable"
<mvdan>
I realise that a terminal's job is mostly to hide and interpret control characters, but from a naive point of view, I'd hope the ones that somehow end up as input by the user wouldn't fly under the radar like this. Perhaps this view is rather naive :)
cvmn has joined #foot
<dnkl>
mvdan: that's something the shell has to do. They are control chars that are interpreted by the terminal, not printed.
<dnkl>
Even if we wanted to print them, we can't, because they have a wcwidth of 0
<dnkl>
Meaning, we'd end up desynchronizing the cursor
cvmn has quit [Ping timeout: 276 seconds]
cvmn has joined #foot
<mvdan>
when you say "the shell has to do" - do you have an idea of how to do this with bash, for instance? my googling is turning up nothing, as I just find ways to "insert non-printable characters via shell script", which is the opposite
<dnkl>
I don't think you can make it. It would have to be explicitly added to bash.
<dnkl>
At least zsh has something like this for byte sequences that are invalid in the current locale
cvmn has quit [Ping timeout: 276 seconds]
<mvdan>
hmm, I see
cvmn has joined #foot
noteness has quit [Remote host closed the connection]