On Wed, Jul 07, 1999 at 10:51:32AM +0300, Jarkko Hietaniemi wrote:
http://www.unicode.org/unicode/reports/tr18/, "Unicode Regular
Expression Guidelines". Some of the things we have, some we don't.
Maybe the most notable idea is *substraction* of categories (these
used to be called character classes). The syntax is neat, too, but I
have doubts about backward compatibility. Basically, '^' toggles the
"polarity" of a category: [A^B^C] = category A - category B + category C.
Here they also talk about "word characters" and equivalence classes ([=c=]
in POSIX), and about collation characters ([.c.] in POSIX).
I proposed something like [[: alphanum & ^lower :]] some time ago (I
hope you allow for whitespace in POSIXish syntax? ;-).
Ilya