Re: [leafnode-list] Problems with posting (was: no subject)

[krasel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx (Cornelius Krasel)]

Stefan Wiens wrote:

[fetchnews deletes articles from out.going/ when receiving a SIGINT
 or when jumping part of the code]

> If this happens, fetchnews deletes the articles from out.going/
> although they haven't been successfully posted. Because fetchnews
> links to failed.postings/ only after an error reply from
> the server, all unsent articles get lost.
> 
> The same will happen if fetchnews receives SIGINT while posting.
> 
> Fetchnews should delete from out.going only after postarticles() was
> successful at least once.

The reason I want to implement several posting attempts of an out.going
article over all servers is that some upstream servers sometimes suffer
of bad propagation. I don't see an easy solution to achieve this at
the moment. Otherwise, one could unlink() immediately after receiving
"240".

The way to go is probably to return a list of filenames that were
posted successfully and to work through this list at the end of the
program.

Another solution would be to test for the message id upstream. Articles
whose message id is available upstream on at least one server are already
deleted now immediately. I think this is basically what Matthias
suggested. Disadvantage: more overhead (a HEAD command for each article,
since STAT does not work reliably with all available NNTP servers, most
notably at least certain versions of nntpcache; maybe one could use
XOVER instead of HEAD?).

--Cornelius.

-- 
/* Cornelius Krasel, U Wuerzburg, Dept. of Pharmacology, Versbacher Str. 9 */
/* D-97078 Wuerzburg, Germany   email: krasel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx */
/* "Science is the game we play with God to find out what His rules are."  */

-- 
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list