Re: case-table functions clobbering extra slots

[Dave Love]

"Eli Zaretskii" <address@hidden> writes:

> AFAIK, that mechanism is a bunch of giant tables that tell, for every
> character, what other character are case-related to this one.

To a first approximation, yes, and the data are used for `the bunch of
giant tables' built by characters.el.  However, we didn't build that
properly systematically even for Emacs 22, and it probably still has
errors and omissions apart from the special cases.

> I don't see how this helps us devise our own solution.

Huh?  Those data are what I referenced for the rules for sigma, for
instance.  You shouldn't have your `own solution'.  Do what Unicode
says when it's easy enough/important enough.  As far as I remember,
Unicode recognizes applications not changing string length, for
instance.