access: in readers.conf

[Jeffrey M. Vinocur]

On Sat, 11 Jan 2003, Russ Allbery wrote:

> Jeffrey M Vinocur <jeff at litech.org> writes:
> 
> > Oh, certainly not.  I don't care about the bug (except for the general
> > icky feeling that known-but-not-being-fixed have), but there are a few
> > recurrent issues from access: that would also go away.
> 
> Yeah, that's a good point.
> 
> Well, the newnews one is already weird, since allownewnews is already an
> inn.conf parameter and a readers.conf parameter in its own right, but
> access overrides it.

That's the problem -- it isn't a readers.conf parameter at all.


> Adding allowapprove as an inn.conf and readers.conf parameter makes a ton
> of sense.  So does allowihave.  (Or allownnrpdihave, maybe?  The other is
> a bit confusing.)

All agreed (and yes, but I was feeling uninspired).


> L doesn't make any sense as an inn.conf parameter,
> though, since if you want everyone to be able to post to groups marked no
> local posting, why mark them that way at all?

Heh, true.  I guess defaulting to false would make sense for that one, 
especially since a default can be set at the top of readers.conf and 
inherit through.  (There's probably some corner case for the Perl/Python 
hooks where it might be relevant, but with perl_access and hopefully soon 
python_access I think we don't need to worry about that.)


> I wonder if we could get away with leaving access there, so that people
> don't have to change their existing configuration, documenting the order
> constraint, but deprecating it even more in favor of the above new
> parameters that we've added.

Ah.  I think this sounds like an extraordinarily sensible plan.


-- 
Jeffrey M. Vinocur
jeff at litech.org