Subject: | |
From: | "Eric Thomas (CERN/L3)" <ERIC@LEPICS> |
Reply To: | Revised LISTSERV forum <LSTSRV-L@DEARN> |
Date: | Tue, 7 Jun 88 12:16:27 GMT |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
As sites start installing HPO 5, I've been getting a number of bug reports
about a "loop" in LISTSERV. As I have said already twice, this is not a
LISTSERV bug. LISTSERV gets an invalid file in its reader, and issues a 'CP
TRANSFER RDR spoolid TO userid' to transfer it to the postmaster. Now, some
level of HPO 5 breaks the TRANSFER and CHANGE commands, so that you have to do
'CP TRANSFER LISTSERV RDR spoolid TO userid'. I personally call this a bug, as
you had previously never had to put your own userid on the command and I don't
see why you should need to.
Now I've heard that some IBMers have come up with the notion that this works
as designed, and that the application should be changed. I am *certainly* not
going to even consider moving the tip of my finger to put 'LISTSERV' in all
the TRANSFER and CHANGE commands everywhere just because IBM have started a
massive "Privileged Command Reconstruction" (the first affected being the
RSCSV2 TRANSFER/CHANGE/PURGE commands). Especially as, from what I understood,
there is a PTF which fixes this "new feature" of HPO5.
Now we don't run HPO5 on any of the CERN CPUs, so I have very little info on
this. I tried looking into INFOMVS but didn't find anything. If one of the
HPO5 LISTSERV sites could post a short note to the list, indicating the
present official status of the problem along with the APAR number, I would be
most grateful.
Eric
PS: The mail forwarding package will not work under HPO5. Now if someone could
send me a console log with the exact message that CP prints when you
TRANSFER a file from someone else's reader, I could make a small change to
have it work under HPO5.
|
|
|