15:32 <dnkl> rockorager: thought I'd see how much work it would be to implement the 'w' part of the text-sizing protocol. TL;DR there's now a WIP branch, text-width-protocol. No PR yet, because I don't yet handle w=0. But explicit widths seems to work (including text reflow on window resize)

15:59 <rockorager> dnkl: nice. I haven’t played around with it at the client side but it should be incredibly easy to implement there

19:37 <krobelus> I don't know if this is the best solution to the wcswidth-mismatch problem but it's very simple, which probably gives it the edge over the "ask the terminal how many cells this text is" approach.

19:41 <krobelus> AFAICT, being able to emit it without querying for support is more convenient but it's not necessary to prevent any glitches right (the extra network roundtrip should be insignificant in most cases)

19:44 <krobelus> I plan to use text-sizing for emoji+variation selector sequences. For ease of implementation, I might simply apply it to every non-ASCII code point

19:49 <krobelus> and ambiguous width too, as you mention. I guess PUA glyphs only need handling if the app explicitly supports them.