I just converted from 0.89 to 0.91 last night and have restored back
to 0.89 this afternoon.
I almost almost always run the "expire -r" process to clean up the
database, I check the totals at the end, and then check the directories
for the actual number of bytes deleted (usually more since other files
are affected besides the news files) and log the count.
The two expire -r's on 0.91 gave me:
articles expire bytes actual bytes
deleted reported deleted deleted
run a - 0.91 22,000 53,000,000 11,500,000
run b - 0.91 9,500 27,800,000 5,500,000
I then regressed back to 0.89, reimported the 3 soup packets since
the upgrade (but hadn't reread all of the same articles) and ran an
expire -r and got:
run a - 0.89 16,000 40,500,000 49,300,000
For me, this lack of space reclaimation is a problem.
I also had a couple of sys3175's in yarn.exe but I was unable to
screen print the registers (sorry). These always seemed to occur when
I had finished marking articles as read (via a K) in the article
selection level and was returning to the newsgroup selection level.
These were generally heavily crossposted newsgroups.
Thanks,
David
-- David Meade Internet=dmeade@netcom.com Oakland, CA dmeade@slip.net