Path: utzoo!utgpu!watmath!clyde!att!pacbell!lll-tis!helios.ee.lbl.gov!pasteur!shambhala.Berkeley.EDU!ricks
From: ricks@shambhala.Berkeley.EDU (Rick L Spickelmier)
Newsgroups: news.software.nntp
Subject: Re: Bug in GROUP command in NNTP
Message-ID: <7427@pasteur.Berkeley.EDU>
Date: 15 Nov 88 21:13:00 GMT
References: <43126@yale-celray.yale.UUCP>
Sender: news@pasteur.Berkeley.EDU
Lines: 17


The NNTP server reads the newsgroup information once on
startup and does not look at it again.  So the GROUP information
is consistent with the state of the news system on NNTP
start up.  The LIST command goes directly to the 'active'
file rather than dumping the internal list.  This is the 
inconsistency.   In XRN we have a 'rescan' button that drops
the connection and restarts it (keeping XRN's internal state, but
getting a new NNTP server) so the new information can be seen....

Maybe the NNTP server should stat the 'active' file on each 
NNTP request and 'rescan' if necessary.


			Rick Spickelmier
			UC Berkeley
			ricks@berkeley.edu