ietf-nntp OVER extension

greg andruk gja at meowing.net
Mon Dec 31 09:54:01 PST 2001


On Mon, Dec 31, 2001 at 12:21:39PM +0000, Charles Lindsey wrote:
> Andrew seems to think they *are*. Please will people report what their
> servers actually DO with LIST OVERVIEW.FMT? (mine doesn't support it at
> all).

This is how it looks on virtually all servers that support it.  Xref:full
may or may not be there, and other Foo:full headers may be present at the
end.  The article number is not included, that is more an index or key.

 215 Order of fields in overview database.
 Subject:
 From:
 Date:
 Message-ID:
 References:
 Bytes:
 Lines:
 Xref:full
 .

> >I think we need to be clear on what is *supposed* to happen. Once we know,
> >it will be easy to write proper words.
> 
> >(1) Should LIST OVERVIEW.FMT include the 8 mandatory fields ?
>      NO

Not the article number, but yes to all the others.

> >(2) If so, what are the names for the last two of the 8 ?
>      Quite so, and also the first mandatory field (article number)

Answered above.

> >(3) Can the mandatory fields have the header name prepended ?
>      NO
> >(4) Can any optional fields have the header name prepended ?
>      YES
> >(5) Must any optional fields have the header name prepended ?
>      YES
> >(6) In LIST OVERVIEW.FMT, should "full" have a space before it ?
>      Pass.

No.

>      Anyway, I would rather remove "full" entirely if (5) is agreed. Could
>      INN survive that?

The actual purpose of overfiew.fmt is to tell INN how to generate overview
records.  "Headername:full" is what tells it to include the keyword.  Yes,
the file's format could (and IMO should someday) be changed to enforce
correct NOV syntax, but that's all internal implementation stuff and
shouldn't have anything to do with what appears on the wire.

There was once a Mac newsreader (Value Added Newswatcher maybe?) that
attempted to use this command and its earlier form LIST SCHEMA, but it was
able to function without it.

> But if (5) is agreed, then the whole LIST OVERVIEW.FMT command becomes
> redundant, as Greg Andruk has said.

Yes, but an improved description of the NOV format should be included in the
RFC.  Currently the only place it is fully documented is in the C News
distribution.  INN's overview.fmt(5) describes that file's syntax, but not
the ordering restrictions.



More information about the ietf-nntp mailing list