[net.news.config] UUCP truncates files

lepreau (10/08/82)

This may be a solution.   Please excuse any duplication-- I'm reposting it
cause of its importance and because we never got this article over USENET
at all.  We haven't gotten any other arpa-originated unix-wizards stuff
for a couple weeks.  (And there has been a lot of it!)  Someone at SRI tells
me that the problem must be "downstream" of them, because the stuff has made
it into their netnews.   Did the "rest" of you on our harpo side get it? 
I thought sri-unix used to send stuff to cca.  decvax again?
-Jay Lepreau

Mail-from: ARPANET site SRI-CSL rcvd at 1-Oct-82 0209-MDT
Mail-from: ARPANET host SRI-UNIX rcvd at 28-Sep-82 0339-PDT
Date: 28 Sep 82 2:31:50-PDT (Tue)
To: Unix-Wizards at SRI-UNIX
From: menlo70!sri-unix!fortune!Dave-Yost (Dave-Yost ) at Ucb-C70
Subject: UUCP BUG FIX -- uucico throws away data
Article-I.D.: fortune.588
Via:  Usenet; 28 Sep 82 3:36-PDT
Remailed-date: 30 Sep 1982 2107-PDT
Remailed-from: the tty of Geoffrey S. Goodfellow  <Geoff5 at SRI-CSL>
Remailed-to: Unix-Wizards@SRI-CSL: ;

Uucp bug:
A sends file to B,
A is master.
A logs the following:
  day randvax (9/9-13:58-7031) REQUEST (S D.hollywoX0367 X.hollywoX0367 day)
  day randvax (9/9-13:59-7031) BAD READ (expected 'C' got FAIL)
The file is not successfully transferred.
A deletes the D. file (WRONG).
Next time the connection is made:
  day randvax (9/9-14:26-7227) REQUEST (S D.hollywoX0367 X.hollywoX0367 day)
  day randvax (9/9-14:26-7227) FAILED (CAN'T READ DATA)
Et voila!  The file transfer never happens.

This problem has happened to me several times in one week!
(All because of GTE, of course.)

Fix: in cntrl.c, change
			unlinkdf(W_DFILE);
			RMESG(RQSTCMPT, msg);
			goto process;
to
			RMESG(RQSTCMPT, msg);
			unlinkdf(W_DFILE);
			goto process;

While you're at it, I recommend the following change in pk1.c:
  #define GETRIES 10      /* 5 in 4.1bsd -- d yost */
I watched uucico give up after 5 retries many, many times.
This got old, so I increased it to 10, and on the very next
connection I saw it pick up the conversation again after about
9 retries.

--dave yost