[NNTP] LISTGROUP extension

[Russ Allbery]

Matthias Andree <ma at dt.e-technik.uni-dortmund.de> writes:

> What happens if the group is modified between the client is sent the
> LISTGROUP reply and can send a subsequent ARTICLE/HEAD/BODY/STAT?

Then the article may not exist.

> I am wondering because section 8.1.1 LISTGROUP/8.1.1.2 Description of
> draft 25 snapshot 5 reads "On success, the list of article numbers is
> returned as a multi-line response following the 211 response code (the
> arguments on the initial response line are the same as for the GROUP
> command. The list contains one number per line, is in numerical order,
> and lists precisely those articles that exist in the group." (3rd
> paragraph).

> Note the "precisely", which IMO is an empty promise: Unless "LISTGROUP"
> implies some sort of "snapshot" (which it does not), the whole
> "precisely" is dead in the water when a write access hits the group -
> we'll then have listed an article that has been removed or we my have
> omitted a newly arrived article.

> Can we reword this paragraph? A first suggestion is "and lists precisely
> those articles that exist in the group at the time when the command
> arrived at the server. Note that articles may be removed or added to the
> group at any time, so a client MUST NOT assume that retrieval commands
> for article or details thereof will succeed."  but the standards gurus
> may have more background information than I do.

Note that we've closed discussion of this part of the draft, but I don't
want that to get in the way of making simple clarification changes if
everyone thinks they're okay.  I think the first part of your rephrasing
is fine, and if this seems completely uncontroversial to everyone else,
I'm fine with Clive making that change.  I don't think the second sentence
is really necessary.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>