17:36 <sterni> that is also a wrong abstraction though because the application doesn’t know how big the “character” is

17:36 <sterni> that is determined by the size of the glyph in the used font technically which the application doesn’t know about

17:38 <sterni> i guess the only improvement of this protocol is that terminal and application no longer need to check wcswidth independently from each other (with the possibility that different implementations are used)

18:31 <cbb> sterni: that part is dealing with how many columns the text occupies, not the glyph size specifically

18:31 <cbb> https://github.com/kovidgoyal/kitty/blob/cell/docs/text-sizing-protocol.rst#fixing-the-character-width-issue-for-the-terminal-ecosystem

18:41 <cbb> allowing the client to specify the width should be solution to issues like this: https://codeberg.org/dnkl/foot/issues/1665#issuecomment-1734299

18:51 <sterni> cbb: yes absolute size doesn’t matter but relative size does which is still controlled by the font

18:52 <sterni> even if the guidelines you can derive from Unicode about which grapheme clusters are supposed to be “double width” is respected by the font, missing glyphs and font fallback also need to be considered

18:55 <sterni> e.g. if your font doesn’t have a glyph for 👨‍❤️‍👨, it would be rendered via the three glyphs that make up the ZWJ sequence

18:55 <sterni> i.e. you would expect this to take 2 cells but it may also take 6 lacking font support

18:56 <sterni> there is no way for the application to know whether this glyph exists

19:12 <cbb> sterni: from 2 days ago: https://libera.irclog.whitequark.org/foot/2025-01-19#1737294306-1737294514;

19:13 <cbb> I think this spec is trying to solve the cursor synchronization problem, not the "exactly as I want it" rendering problem

19:13 <cbb> although it seems like it'll solve one and improve the other

19:15 <cbb> I've thought about the problem quite a bit from the application side of things and this spec is the most practical/realistic solution I've seen so far

19:42 <sterni> that is true

22:08 <xaltsc> Hey, I've been trying to use fontfeatures, but, 1. it doesn't work (loglevel info shows it's been taken into account), 2. I can't find the syntax for multiple fontfeatures.

22:22 <xaltsc> Oh, I didn't see that foot didn't support ligatures, nevermind.

22:37 <xaltsc> Also, I can't really decide whether I should run foot as a server. Are there genuine benefits when resource usage is not really a problem ?