Re: Greek end sigmas cannot be up/down cased

[Richard Stallman]

    In Greek, downcase-word and upcase-word do not work correctly for
    the end sigma.

This is the absence of a feature, rather than a bug.  The existing
case table feature simply has no way of recording the idea that a
single upper-case character corresponds to two different lower-case
characters depending on position in the word.  To make this case
work right, it would be necessary to define a way to represent
this situation.

There are other problems with the current code.  For instance, it
can't handle the German lower-case s that corresponds to SS in upper
case.

Would someone like to volunteer to study the situation and find all
the cases that our current code doesn't handle?  We shouldn't
install such a change before the coming release, but you could
work on what we should do after the release.