Re: ZWJ paste from clipboard problem (unicode)

On Tue, Apr 1, 2025 at 10:08 AM Daniil Rozanov <[email protected]> wrote:
>
> On Tue Apr 1, 2025 at 5:07 AM +04, Mikael Magnusson wrote:
> > On Mon, Mar 31, 2025 at 11:23 PM Daniil Rozanov <[email protected]> wrote:
> >>
> >> On Mon Mar 31, 2025 at 11:31 PM +04, Bart Schaefer wrote:
> >> > On Mon, Mar 31, 2025 at 12:13 PM Daniil Rozanov <[email protected]> wrote:
> >> >>
> >> >> When trying to paste "👩" (U+1F469 U+1F3FB U+200D U+1F9B0) emoji, zsh
> >> >> prints two emojis and "<200d>" in between.
> >> >
> >> > Try "setopt combining_chars"
> >> >
> >> > Additional tweaks may be needed if your terminal is using bracketed paste mode.
> >>
> >> Thanks! Now it pastes pretty emoji. But now there is another problem,
> >> which is character's width is more than expected. And with some special
> >> emojis it works wrong: duplicated display on paste, backspase erases
> >> kinda parts of emoji instead of remove it all at once. Btw this
> >> behaviour now is similar to bash.
> >>
> >> Maybe zsh has some group of settings which handle unicode?
> >
> > What you want is not theoretically possible as far as I know.
> > Combining emojis is optional according to the specs, and zsh has no
> > way of knowing if a terminal will combine two particular emoji in a
> > zwj sequence, and thus has no idea how wide the sequence of characters
> > will be or when the terminal will start drawing on a new line.
> >
> As far as I know it is possible on some bunch of terminals. This
> possibility called "mode 2027", briefly described here:
>
> https://mitchellh.com/writing/grapheme-clusters-in-terminals
>
> So at least zsh can test terminal to know does it support grapheme
> clustering or not and work around this.

This requires querying the cursor position from the terminal though,
see the thread starting at
https://www.zsh.org/mla/workers/2016/msg02326.html for why this is
unlikely to happen. (We need to know the width before printing the
text in many cases).


--
Mikael Magnusson