🚀 go-pugleaf

In article <815151292snz@siberia.demon.co.uk>
           philip@siberia.demon.co.uk "Philip McNab" writes:

> Hmmm .. I've tried deleting my newsgroups when they get corrupted, and
> downloading all the articles again. Unfortunately, deleting them
> zapped the files, but didn't remove the entries from history.snw.
> Running expire after deleting the newsgroups (and before re-creating
> them) didn't help either.

I hope that when folks use the expression "deleting newsgroups" they
mean that they are removing them through DIS or Snews.

In terms of re-loading, my suggestion is to simply delete both history.snw
and history and then run NET with a narrow list of newsgroups in nntp.dat
and a sufficiently far back date and time in the first line.  The problem
that this will create is that all cross-postings to newsgroups that were
not removed and re-added will be duplicated.

Foolhardy folks like me might care to go the extra step of editing the 
active file in news to match the narrow list in nntp.dat and this will 
cause all the cross-postings to be avoided.

> My expire.dat contains three lines which I thought should make this
> happen:
> 
> [copy] 1000

Never used copy.  Anyone ever found it useful?

> [tail] 1000

Seems fine if you're swapping between unsynchronised Demon news servers.  
If you stick to just news.demon.co.uk and are happy to await it's return 
in the event of a crash, then you really only need 2 minutes worth + or - 
the typical clock error that is adjusted after connection.  If you also 
have access to newsservers outside of Demon, you may need far more.

> [life] 21

That seems high to me.  Although I use 9 days myself, I don't know of any 
reason why a number over 1 is strictly useful in our environment.  But any 
size is fine so long as you have the disk space and unbatch doesn't gag 
from lack of memory.

> Have I done something wrong?

Other than groping for replacement news which is still listed in history 
files and so is trashed in the nntp session or in unbatch, as you yourself 
discovered, nothing that I can suggest.

The one thing I'd urge is for you to find why you are ever getting the 
newsbase files corrupted.  I've only had this when time clocks pull the 
mains plug in mid unbatch or expire and with a large smartdrv of cached 
writes and open FCBs.  You must have some flakey hardware or software.  

The one strong point in favour of this platform software is its high 
stability, so you should be satisfied with nothing less.  You could have 
all of your same problems *plus* long lines and html elsewhere.

Regards,
-- 
   Tony - G3SKR / AA2PM                  email: tgold@panix.com
                                                tgold@microvst.demon.co.uk
                                        packet: g3skr@n0ary.#nocal.ca.usa.na