Re: Better emoji support

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: emacs-devel@gnu.org,  kevin.legouguec@gmail.com
> Date: Mon, 20 Sep 2021 15:03:20 +0200
> 
> OK. Whilst updating the emoji sequences support to Unicode-14, I had
> to add the following ranges to the 'emoji' script, since otherwise
> they donʼt end up getting composed, because Emacs doesnʼt use the
> 'emoji' font for them. Theyʼre all Emoji_Presentation=No, but
> emoji-zwj-sequences.txt modifies them with FE0F
> 
> (why didnʼt they just change them to Emoji_Presentation=Yes? I see
> very little consistency between similar codepoints that nonetheless
> have different Emoji_Presentation values)

It has something to do with text segmentation, although I don't
entirely understand what exactly and how.  What confuses me is that a
lot of symbols are also in that category, and we definitely don't want
to treat them as emoji.

> (#x26F9 #x26F9 emoji)
> (#x2764 #x2764 emoji)
> (#x1f3cb #x1f3cc emoji)
> (#x1f3f3 #x1f3f4 emoji)
> (#x1f441 #x1f441 emoji)
> (#x1f575 #x1f575 emoji)
> 
> So we get to pick our poison: complete consistency with the
> Emoji_Presentation property from emoji-data.txt, or having complete
> coverage for emoji-zwj-sequences.txt. Eli? (Iʼll admit to having a
> slight preference for the latter, given that I find Emoji_Presentation
> to be somewhat arbitrary for quite a few codepoints)

I agree: we should use the latter, at least for now.

Thanks.