ietf-nntp Collected minor issues

[Russ Allbery]

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

> Indeed so. The use of [...] requires agreement on the collating
> order. If it is US-ASCII (or you are in the C locale, which amounts to
> the same thing) then all is straightforward.

> But if, in UTF-8, you ask for everything in between 'LATIN CAPITAL
> LETTER A' and 'GREEK CAPITAL LETTER OMEGA' you will get all sorts of
> weird and wonderful things. So writing that future extension would be a
> non-trivial task.

I don't understand quite what you're getting at.  I would assume that
UTF-8's collating order in the C locale is just the numeric order of the
characters.  Surely the client wouldn't attempt to negotiate a locale with
the server and would instead just assume C locale when constructing
wildmats?

There's a limit to what you can do with wildmats; if you really need to do
locale-aware matching, you're going to have to download the data and
interpret it yourself.  Particularly since different newsgroup names are
going to be in different languages (and therefore have different native
collating orders).

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