ietf-nntp new draft of base document available

[Russ Allbery]

Charles Lindsey <chl at clw.cs.man.ac.uk> writes:
> Russ Allbery <rra at stanford.edu> writes:

>> Whoops, you're entirely right.  A pattern consisting entirely of !
>> patterns can never match anything.  I could have sworn that INN did the
>> implicit "*," thing, but looking at it it doesn't.

>> Thanks for the catch; I'll change my implementation and the upcoming
>> documentation.

> Now you've got me confused. Is the spec you just posted to both lists
> pre- or post- the above change?

Post.

> I.e. are you proposing that NEWNEWS !*.binaries.* will/won't work?

That wildmat would never match any newsgroup.

I agree that it would have made more sense to have an implicit *, at the
beginning; after all, that's how I thought it worked too.  But existing
practice in INN is that one defaults to not matching even with a negated
pattern, and I think it's more important here to stick with existing
practice.

It's not that hard for the client to add the *, to the beginning.

> I hope that it will (people will certainly be tempted to try it) but if
> it is thought to be too imcompatible with present practice, then OK.

Using the above wildmat wouldn't be possible for a client that needed to
talk to older servers, so I think it's of somewhat marginal utility
compared to the ease of the client adding the * itself.

> BTW, an implicit '*' at the beginning (whether starting with '!' or not)
> would mean that

> NEWNEWS

> with no paramters would mean "show me everything". That seems not a bad
> idea.

Likewise here.

> And don't forget that the usage of wildmats in PAT is not resolved yet.

I think "amazingly ugly" is a reasonable description of existing practice
for XHDR and XPAT.

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