Julian Assange [proff@iq.org] wrote: > > Since switching to the new provider, when?clients attempt to reset/refresh > > their list of available groups, they get all the way through the list (84,377 > > groups), and then hang. > > Try increasing the idle timeout in nntpcache.config Actually, I think it would make sense to have two different timeouts, one for lack of activity (just waiting for command) and another one for *each* send (perhaps in CFemit?). At the moment there is a timeout for the duration of each command's execution which is the same one used for when a client is waiting to enter a command which I'm not sure is in anyone's interest. Whilst I thought about which was the best way to go, I simply changed the timeout in client_command() to 3600 rather than idle which doesn't seem to have adversely affected the server. BTW. Is anyone interested in some patches I have made for posting directly to a news peer using IHAVE and also some small fixes to the radius and libradius code in 2.4b5? Faye -- Faye Pearson, Software Development Manager, ClaraNET Ltd. Tel 020 7903 3000 Your life would be very empty if you had nothing to regret.
