Overview data on rebuilding

Russ Allbery rra at stanford.edu
Tue Sep 9 17:21:43 UTC 2008


Julien ÉLIE <julien at trigofacile.com> writes:

> I think it unwise to have "-e" in makehistory with RFC 3977:
>
> -e Compute Bytes headers which is used for overview data.  This option
>    is valid only if -O flag is specified and overview.fmt includes
>    Bytes:.
>
> People will forget to use it and it will break the compliance of their
> news server.  What could we do?  Change "-e" to *not* compute :bytes?
> Remove "-e"?

I'd be inclined to remove it and always force that behavior.

> Note that overview rebuilding will be even more slower!

Yeah.  :/  Probably not a *lot* slower in the common case, but yes.

Yoy probably also want to double-check how makehistory is doing the byte
counting.  It might not be doing what RFC 3977 says to do.

> Another question:  Why does inn remove the Bytes: header of every
> article?  Is it harmful?

It's not harmful, but it's pretty much guaranteed to be wrong.  However,
we probably shouldn't be removing user-supplied headers even when they're
wrong.  innd definitely should not do so, and nnrpd SHOULD NOT do so per
USEPRO:

   6.   The injecting agent MUST NOT alter the body of the article in
        any way (including any change of Content-Transfer-Encoding).  It
        MAY add other header fields not already provided by the poster,
        but injecting agents are encouraged to use the Injection-Info
        header for such information and to minimize the addition of
        other headers.  It SHOULD NOT alter, delete, or reorder any
        existing header field except the Path header field.  It MUST NOT
        alter or delete any existing Message-ID header field.

-- 
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