Expires: header

[Julien ÉLIE]

Hi Heiko,

> - How does inn handle articles with Expires: date in the past feeded by
>  peers?

They are accepted.
These articles will expire during the next run of expire/expireover,
if expire.ctl of course allows that.

innd currently only rejects articles whose Date: header field value
is too far in the past (artcutoff value, whose default is 10).
It will also check the Injection-Date: header field.


> - How does nnrpd handle articles with Expires: date in the past during
>  POST/IHAVE?

No check.
They will be accepted and handled as said above.


> - Does nnrpd offer articles with Expires: date in the past to readers?

Yes.

Nothing prevents it in RFC 5536 or RFC 5537 as far as I know.
Is it a *real* problem?


The fact is that everything is run by expire.ctl (you may want to
retain articles even after their expiration date) and the run of
expire/expireover.



> expire.ctl manpage should use examples with reasonable limits for min/max
> of Expire: headers. I propose some values in the attached patch. Maybe min
> should be even 0 instead of 1. As many people copy configurations from
> manpage examples, more reasonable limits could be useful. Ok, not
> important.

Yes, you're right.
"*:1:7:35" is better than "*:7:7:7"!

-- 
Julien ÉLIE

« En fait, je n'ai qu'une prétention, c'est de ne pas
  plaire à tout le monde. Plaire à tout le monde, c'est plaire
  à n'importe qui. » (Sacha Guitry)