bug#81598: [PATCH] Add new input method for toki pona

> From: Joshua Murphy <[email protected]>
> Cc: [email protected]
> Date: Thu, 13 Aug 2026 19:44:15 +0000
> 
> I added to etc/HELLO as a future reference for myself, and also
> added a proper language environment. I'll combine the language
> environment and IME stuff into a single file when I make it into a
> package.

Thanks.

> What specifically would need to be done in a package to get
> font selection working?
> I see that I can do something like
> (set-fontset-font t '(?󱤀 . ?󱦺) "sitelen seli kiwen mono juniko")
> to get a specific font to work, but having it automatically pick one
> would of course be better. Does adding to UnicodeData.txt fix the font
> selection or is something else needed? If that is the problem, what's a
> reasonable way to handle that in a package? I'm guessing a package can't
> just override or append to that file.

This machinery works based on Unicode data files in complicated ways.
First, charscript.el is auto-generated from Unicode's Blocks.txt file,
which produces a char-table where each character is assigned to its
script; that should include this new script when Blocks.txt does.
Then the new script with its representative character(s) should be
added to the list in script-representative-chars in fontset.el, and
any rules of font requirements for the script added to the list of
scripts in setup-default-fontset there.  The list in otf-script-alist
should also be updated with this new script.  When all of that is
done, fontset.el will produce the correct set-fontset-font call
automatically.

As you see, all of this is only possible if Unicode supports these
characters and mentions them in its data files.