#4 premature deleted postings on post

[Dirk-Lueder Kreie]

fetchnews of leafnode 1.9.49 deletes postings too early
form out.going (after first server contacted), which is
bad, because of some situations in which a server has a
dodgy upstream but an acceptable downstream, so
you'd use it as "local" server for leafnode.
even Worse if the "local" server for leafnode only
accepts postings in particular groups and rejects
postings to other groups, (the article will be moved to
failed.postings), both leading to bad delays.

I'd suggest reverting the posting behaviour to the one
exhibited by earlier versions: if an error occurs posting
the article, keep trying on other servers, and delete the
article from out.going when the last server has been
contacted and the article has been successfully posted
to at least one server, otherwise move to failed.postings

[Matthias Andree]

Logged In: YES
user_id=2788

There is a reason why the posting behaviour was changed:
postings to moderated groups. As the servers leafnode
connects to will forward these postings to the group's
moderators by email, but not to other servers (the actual
posting is the task of the moderator), every server that
leafnode connects to would forward the post to the moderator
anew - who then gets multiple copies and is annoyed.

Leafnode-1 does not track which upstream groups are
moderated or not, so fixing the problem without introducing
the other would require major changes I am not inclined to
make for leafnode-1.9.

You can however work around the problem by using the
"nopost" server flag: Edit [/usr/local]/etc/leafnode/config,
writing "nopost = 1" directly below the "server =
dodgy-upstream.news.example.org" line.

I'll see if I can polish documentation a bit for 1.9.50.rel.