access: in readers.conf

Russ Allbery rra at stanford.edu
Sun Jan 12 04:04:15 UTC 2003


Jeffrey M Vinocur <jeff at litech.org> writes:

> I was fiddling on nnrpd/perm.c to make "post: !*" result in a 201 banner
> being displayed instead of 200, and observed from the code the following
> interesting behavior:

>    If access: is *before* newsgroups:/read:/post: lines, the presence
>    of R and P in the access string is ignored.

> I don't think it's an entirely trivial fix, and thinking about it I had 
> another idea entirely...

> Why don't we remove access: in INN 2.4?  The R and P functionality is
> already available with read:/post:, and it should be quite trivial to
> provide all the other parts of access: with individual boolean flags.

I wouldn't mind, although I also wouldn't mind waiting for 2.5.  I don't
really have a strong opinion either way.  It's been broken that way for
all of 2.3 and this is the first time I recall someone mentioning it, so
it's apparently not a frequently noticed bug.

We're hopefully going to be ripping out the current readers.conf parser
entirely for 2.5, so I'm not sure if it's worth people having to think
about this when upgrading to 2.4 when they're going to have to think about
configuration changes again in 2.5.  At the least, we definitely should do
this when upgrading the parser.

> We could make this seamless to the user by having innupgrade modify
> readers.conf (although the presence of multiple readers.conf files needs
> to be considered).

innupgrade has the ability to run on any file with an option to tell it
what style of file it is, so that should be okay.  We'll have to include a
note telling people to run innupgrade on other readers.conf files if they
have them.

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

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.


More information about the inn-workers mailing list