[leafnode-list] is it bug?

[Matthias Andree]

Urke MMI schrieb am 2003-12-21:

> I install leafnode-2.0.0.alpha20031218a and form now, every time I fetch
> news, fetchnews every time download complet group list from servers. I
> watch couple servers and every connection wast my time for 20 or more
> minutes.
> 
> In /etc/leafnode/config I set timeout_active option to 365, but this not
> affect, I think.
> 
> Some tips, please?

It's unlikely that it is a bug. Once fetchnews has decided it has to
download the full newsgroups list, it will retry until it has received
the full list from all servers in one run. Try running "fetchnews -n" so
you'll only see the error messages, and watch for messages like:

connecting to news.freshmeat.net:nntp: cannot connect: [some error]

Some error varies, connection timed out, connection refused, network is
unreachable, to name just three possibilities.

We can only fix this problem by locally caching the active file of ALL
servers, which is a bit wasteful. No such code exists as of this time.

As a workaround, I can forward port the verbose code that answers the
"why does fetchnews re-fetch the full active" question from leafnode
1.9.

HTH,

-- 
Matthias Andree