[NNTP] Article Numbers Becoming Invalid (RFC 3977)

[Sabahattin Gucukoglu]

On 31 Dec 2009, at 08:27, Russ Allbery wrote:
> Sabahattin Gucukoglu <mail at sabahattin-gucukoglu.com> writes:
>> Right, this is fine.  But the question was really, how do I initialise
>> the pointer to something not the beginning of the group?
> 
> ARTICLE <number> or STAT <number>.
> 
>> I cannot use NEXT until I hit on a valid article.  Does this not mean
>> that I will have to find it by trial and error with article numbers?
> 
> Trial and error is an odd way to describe it, since that sounds like
> you're trying article numbers at random.  You start with one greater than
> the last article that you read and try each number in sequence from there
> to the high-water mark until you get a successful reply.  From that point
> forward, you can use NEXT.  (This is, in essence, what NEXT does
> internally in the server.)

I guess I meant that I would have to keep trying until I got a non-error response.  Yes, that's what's happening, thanks.  I have come away from this with the definite impression that NNTP clients should be very tolerant indeed of error conditions.

>> A coinciding question: if I use the ARTICLE command specifying an
>> invalid article number, has the state of the server changed so that NEXT
>> will now no longer work, or will it work as it did before I issued
>> ARTICLE, assuming a valid article before (as on GROUP)?
> 
>   The server MUST NOT change the current article number except when an
>   article number argument was provided and the article exists; in
>   particular, it MUST NOT change it following an unsuccessful response.
> 
> Nothing about an unsuccessful ARTICLE command changes state in the server,
> so if NEXT was working beforehand, it's still working afterwards.

If that's the case, then the only use for code 420 is after commands following entering an empty group, since it's impossible to invalidate the article pointer by any other means.

Happy new year, everybody!

Cheers,
Sabahattin