Mark Brown: > one of the situations I've encountered in the past is a group has a > long enough expiry time to have articles in it but not sufficient > volume going through it to cause readers to actually fetch articles > for it often enough to convince Leafnode that they are interested.
This should no longer be a problem, relevant bugs have caused three changes since April 2004. Newsreaders usually check with GROUP or LIST ACTIVE the article numbers in a group, and request overview or articles only if they find new articles (comparing against .newsrc for instance). 1.9.52+ refreshed subscriptions if a GROUP command for an interesting group was received. 1.10.3+ did the same after LIST ACTIVE with a non-wildmat group name was received (wildmat here := containing * ? or [) Brian Sammon then found out this didn't work for empty groups, so 1.10.8+ fixed the GROUP and LIST ACTIVE refresh for empty groups. > > I have put the wishlist item on the leafnode-2 TODO list so it can be > > reconsidered when configuration handling will be reworked. > > Are you considering looking at this for Leafnode 1? The wishlist item referred to is "group-specific timeout_short/long". This will not become part of 1.11, it is a stable branch of leafnode, and will change only for documentation and bugfixes, so that distributors can take patchlevel upgrades from this branch without worries. I am also very unlikely to open another 1.N branch, N >= 12; unless someone steps up to sponsor manpower or funding that is. Bugs that cause unsubscription from a newsgroup although leafnode can easily know that the newsreader is interested in that very newsgroup will be fixed in 1.11.M releases, of course. Given the data I have so far, I ask that the severity of this bug# be bumped to "normal" for the nonce. Should we figure that it's a situation leafnode cannot resolve, for instance Nikolaus' Gnus issued "LIST ACTIVE *" or such unspecific commands, we can still demote this to a wishlist, or better copy the relevant info to a new wishlist item and close this bug as invalid. -- Matthias Andree -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
