ietf-nntp wildmat routines and text

[Russ Allbery]

Andrew Gierth <andrew at erlenstar.demon.co.uk> writes:

> If you want to compare notes, I'll make available a copy of my UTF-8
> wildmat code.

I'd definitely be interested.  (Provided that the license is compatible; I
have to worry about that since I'm writing the INN implementation.  I'd
make my code public domain, but Rich's wasn't and mine is heavily based on
it.)

> One additional comment on wildmat syntax: especially when handling
> UTF-8, it would be useful to support numeric \-escapes (e.g. \040 or
> \x1234), even in []-sets.

Mmm.. yeah, although it's vaguely annoying to specify (in particular,
what the specification is for numbers over 127 decimal and whether the
character code is the literal UTF-8 octets turned into a number or is
instead some other Unicode character expected to be turned into UTF-8).

It's also not a backwards-compatible change, although I doubt anyone would
really care.  (Although it clutters the nature of backslash, which right
now escapes anything and everything indiscriminately.)

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>