Re: Character group folding in searches

[Eli Zaretskii <eliz@gnu.org>]

> From: Stefan Monnier <monnier@IRO.UMontreal.CA>
> Cc: bruce.connor.am@gmail.com, emacs-devel@gnu.org
> Date: Fri, 06 Feb 2015 19:05:28 -0500
> 
> >> I don't see how turning single chars into char-sequences can help us
> >> handle multi-char sequences (unless maybe you restrict it so that there
> >> can only ever be one such sequence per equivalence class?).
> > The decomposition sequence is unique, yes.
> 
> Elsewhere:
> > It will, if you only pay attention to the base character.
> 
> Hmm... could you characterize the kinds of equivalence classes you
> intend to handle?

That was one part of what I suggested to discuss.  We don't yet have a
list we decided upon.  However, it is quite clear that comparing only
the base characters is one feature we would want, for the same reasons
we wanted to ignore diacriticals in string-collate-equalp.

> Could it handle for example an equivalence class which includes
> →, ⇒, ->, and => ?

These are our application-specific equivalences, they are not in
Unicode.  So we will need to have a list of equivalences in addition
to Unicode, that we will want to add to the "folding" char-table,
either permanently or as buffer-local customizations.

But once the decomposition is in the table, processing it should be
the same as with others, although it's possible we won't allow only
base-character matches for equivalences such as the ones above.