Stefan Monnier wrote:
> I'm not sure exactly what he wanted to say, but it sounds to me like
> it's going in the same direction as my earlier request to replace the
> hard-coded table by code that auto-generates the cases.
> There is already similar code in latin-ltx.el (written by yours truly).

OK, thanks, in that case this will need some thinking, since the code in 
latin-ltx.el suffers from the same problems I mentioned in 
<http://bugs.gnu.org/20499#105>: from a user's point of view the supported 
characters are a haphazard list.  E.g., it adds some chars for Pinyin tones but 
not others.  Partly the problem is that it adds "easy" Latin letters like ȳ even 
though nobody uses them, but not "hard" ones like ǚ even though they're actually 
used on occasion.

Fixing this will take some thinking, because we'll need to devise ways to type 
the "hard" Latin letters.  I suppose latin-ltx and iso-transl should use similar 
approaches here.

In the meantime, though, there is a need to type non-Latin punctuation like 
dashes and quotation marks.  That part of the patch seems relatively independent 
of the Latin-letter issue, so I installed the attached.  I hope to look into the 
Latin-letter issue later.