[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[leafnode-list] leafnode-1.9.4, groups expiring after timeout_short?


I use LeafNode 1.9.4 both at home (only having a dialup line) and at work
(because the main server is s-l-o-w, I prefer to have my own server to
read Usenet with).

However, at work, the groups always expire after two days (timeout_short)
of my not reading them (typically, the fetchnews in my crontab stops
downloading articles over the weekend). This problem doesn't occur at
home, even though the setups are almost identical: the only differences
are that delaybody is set at home, the underlying OS are Debian 2.1 (home)
and RedHat 5.2 (work), and the crontab runs daily at home and quadrihourly
at work. And I don't usually read news at home except on weekends.

I skimmed through the sources, and the only plausible reason I found is
that fetchnews discriminates between groups that have been opened just
once and actively read groups by comparing the ctime and mtime of the
corresponding file in <spooldir>/interesting.groups. Except that the
utime() call in markinterest() (nntpd.c:514) modifies both of them at
the same time...

OTOH, if this is the root cause of the problem, I can't explain why it
doesn't happen at home; one could believe that with enough markdownload()s
now and then the mtimes become different from the ctimes, but stat reveals
that most files in ../interesting.groups also have ctime==mtime.

Anyway, I can always increase my timeout_short at work, but I don't think
that's appropriate. Any advice? Perhaps markinterest() should cheat by
setting mtime one second or so below the current time?

						Thanks in advance,
						Cedric Ware.

leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list