ietf-nntp OVER extension

[Clive D.W. Feather]

Russ Allbery said:
>> I think that 9.5.2 would be better written if it had the OVER command
>> first, describing the overview database and how the fields
>> appear. *Then* have the LIST OVERVIEW.FMT command, which describes which
>> optional fields OVER offers.
> I agree with this.

I have now done a complete rewrite of the OVER extension text. I will post
it to this reflector in a moment, and also put it at
<http://www.davros.org/nntp-texts/section-9.5.2.txt>

As well as reversing the order of the commands, I've tried to tighten up
the specification and to fix a whole load of minor errors (e.g. missing
digits in responses, missing fields in examples). As a result there are
far too many changes to include change bars.


>> Whereas the OVER example does *not* have 8 fields even though 8 are
>> mandatory ! Is it permitted for OVER to return less than 8 items ?
> I haven't looked at the draft, so I'm not sure what you're referring to,
> but it's allowable for some of the fields to be empty (but the fields must
> still be delimited by tabs).

And trailing tabs may be omitted. However, I was referring to:

          [C] OVER 
          [S] 224 Overview information follows 
              300234|I am just a test article|nobody at nowhere.to
              (Demo User)|6 Oct 1998 04:38:40 -0500|
              <45223423 at to.to>
          [S] .

which doesn't include byte and line counts. [The article number is also
incorrect, since the LWM for that group is 3000234.]

>> (2) If so, what are the names for the last two of the 8 ?
> Bytes and Lines.

Is it too late to change these, e.g. to omit the colon or in some other way
indicate that they're not headers ? I've put possible wording in my
rewrite.

>>> I am slightly surprised that the OVER command does not have a
>>> <message-id> alternative (since all the other commands with a [range]
>>> parameter seem to have one).
>> The overview databases (as I understand it) are per-newsgroup. There is
>> no easy way (without fetching the article) to map an ID to a newsgroup,
>> so this version is not provided.
> Correct.

>> I would not object to it being optional.
> How many current servers support this?  I know that INN doesn't; it's
> easy enough to fake up by retrieving the article and building the overview
> information on the fly, but it seems a bit pointless since I don't know
> why a client would do that rather than just sending a HEAD command.

It gives you the information in the same format as other OVER requests.
This might be useful.

> whole point of OVER is that it gives you the essential information that
> you need to thread and do similar things for a whole bunch of articles at
> once much faster than using HEAD, but none of those reasons apply to
> getting information for a single message ID.

I'm loath to use "not faster" as a reason for absolutely forbidding
something that people might find useful in the future.

I've put optional wording into my rewrite.

-- 
Clive D.W. Feather  | Work:  <clive at demon.net>   | Tel:  +44 20 8371 1138
Internet Expert     | Home:  <clive at davros.org>  | Fax:  +44 20 8371 4037
Demon Internet      | WWW: http://www.davros.org | Mobile: +44 7973 377646
Thus plc            |                            |