[NNTP] DATE: reader or general command

Clive D.W. Feather clive at demon.net
Thu Dec 2 01:36:44 PST 2004


Russ Allbery said:
>> I'm happy to shift commands between the various classes based on the
>> consensus of the working group. Personally I'd move DATE, HELP, and
>> NEWNEWS to general,
> 
> innd is definitely not going to provide NEWNEWS.  NEWNEWS requires a
> completely different type of database to work efficiently than innd would
> otherwise need.

Okay. That's a good reason.

> A case could be made for HELP, which is widely implemented in transit-only
> servers.  I'd be mildly inclined to move HELP to general, yes.

Shall I do it?

>> and I don't see the logic that says HEAD is general but ARTICLE
>> isn't.
> Because you can't actually read news with HEAD, just debug problems, and
> you don't get more information from ARTICLE that's useful for debugging
> problems.

No, I don't see it. We should not be making decisions like this based on why
we think someone might be using a command, but on what the command actually
does.

I can see how a server might be able to implement STAT without access to
the newsbase, so I can see how a transit-only server might do STAT but not
ARTICLE, HEAD, and BODY. But is there really any implementation for which
the code for those three is actually significantly different? Surely they
all look basically the same:

    find_article ()
    if (not_found) return correct 4xx code and exit
    send_response (command == ARTICLE ? 220 : command == HEAD ? 221 : 222)
    bool send_line = command != BODY
    while (line = get_line_from_article ())
    {
        if (line == "" && command == HEAD)
            break;
        if (send_line)
            send_dot_stuffed_line (line);
        if (line == "")
            send_line = true;
    }
    send_trailing_dot ()

So all are equally easy to provide (unless someone did a deliberately
perverse implementation, of course).

> It wasn't provided more as a security thing than anything else,
> originally,

Huh?

>> But I'm leaving it up to those who write split-function servers to
>> decide.
> I'd just stick with what INN has always done; there isn't anything
> fundamentally broken about how the division was done, and there's no real
> point in worrying about it.

If that's the consensus, that's also fine with me.

-- 
Clive D.W. Feather  | Work:  <clive at demon.net>   | Tel:    +44 20 8495 6138
Internet Expert     | Home:  <clive at davros.org>  | Fax:    +44 870 051 9937
Demon Internet      | WWW: http://www.davros.org | Mobile: +44 7973 377646
Thus plc            |                            |



More information about the ietf-nntp mailing list