ietf-nntp [a-z] in Wildmats

[Russ Allbery]

Charles Lindsey <chl at clw.cs.man.ac.uk> writes:

> BUT there are some glyphs where there is no single character form. For
> example a TIBETAN LETTER PHA with a TIBETAN SIGN NYI ZLA NAA DA over the
> top of it (no, I don't know whether that would be good Tibetan or not,
> but you get the general idea). So what if you want to construct a [...]
> including that glyph (or do I mean those two characters)?

You pretty clearly mean two characters, IMO, and you can't; it's a
character range.  It matches Unicode code points, hopefully canonicalized.

> I can see situations where either would be useful (Clive's present text
> is strictly in terms of characters).

I'm a bit confused here... are you talking about having character ranges
apply to graphemes instead of characters?  That's heading in the direction
of collation-aware character classes, which is a *huge* amount of work and
is way more weight than the poor little rickety framework of wildmat is
going to be able to bear.  For stuff like that, you really want to go all
the way to POSIX regexes.

> So I think it would be better to leave the feature out.

Works for me.

> Note that such multi-character glyphs present no special problems in the
> rest of the wildmat syntax, provided they are properly normalized.

Yup.

> BTW, when are we going to see a draft incorporating these latest agreed
> changes?

We hadn't agreed on the wildmat stuff.  The last status so far as I know
was that the four of us who were actively participating in the discussion
all posted our opinions as to what should be included, Clive had text for
all the alternatives, and there was a general appeal for more opinions
that went mostly unresponded-to.

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