Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!brl-tgr!wmartin
From: wmartin@brl-tgr.ARPA (Will Martin )
Newsgroups: net.news
Subject: More care needed in saving drafts
Message-ID: <3@brl-tgr.ARPA>
Date: Thu, 18-Jul-85 16:03:03 EDT
Article-I.D.: brl-tgr.3
Posted: Thu Jul 18 16:03:03 1985
Date-Received: Sat, 20-Jul-85 05:22:55 EDT
Distribution: net
Organization: USAMC ALMSA
Lines: 23

We *desperately* need a fix to rn's Pnews process so that it saves the
draft in /tmp until after _everything_ has completed successfully. I just
had to endure the loss of a completed and edited posting because, after
I replied "s" to the Send, abort,... question after finishing the
posting, the system came back with some nonsense message from "inews"
about the filesystem being full, and it couldn't post it. Maybe so; the
problem was that, when I then went to /tmp to try to copy my posting so
I could re-submit it, it *wasn't there*!!! It had been written out to
/tmp completely with no trouble when I ended the editing, so there was
plenty of filesystem space there. There is *no* excuse for the software
deleting the /tmp copy BEFORE the posting has been successful; if it had
put it in "dead.article" or somewhere, it would have been OK, but it
just flat *died* and threw away all my work!

It is unreasonable to expect people to have to remember to do special
manual "saves" or ":w filename" commands every time they post, just on
the chance that the software will screw them like it just did me. It
should be built in a "fail-safe" manner, so that any problems will still
leave the user his text or draft for future use.

Will Martin

UUCP/USENET: seismo!brl-bmd!wmartin   or   ARPA/MILNET: wmartin@almsa-1.ARPA