[comp.os.vms] Photo/Eunice problems

BROSSEAU%SECAD1@engvax.scg.hac.COM ("LISA M. BROSSEAU") (10/26/87)

Please excuse the reposting of this question,  but I never saw the 
message come through... so just to make sure it gets out there here
we go again:


Hello all,

I am having a problem that no one (so far) has been able to explain.  


We have a cluster consisting of an 8650 and a 785 -- running VMS 4.4.

We are having a problem with Eunice and Photo.  When the LT or VT #'s 
rise above 1000 (yes our systems do stay up long enough to get there)
both Eunice and Photo stop working.  Eunice just does wierd things and 
Photo fails with the following error:

%SYSTEM-F-RESULTOVF, resultant string overflow

I believe the problem has to do with some definition in a system include
file that only sets up the field to hold the LT number as 5 characters 
when it should be 6. 

Who knows?

Any help on this problem would be greatly appreciated, since the only way
I can fix the problem now is to shutdown the LATdriver or shutdown the
system  (virtually the same since all of our terms are connected via 
terminal servers).


Thanks for any help

Lisa Brosseau
Hughes Aircraft Co.
Newport Beach  (Yes it is nice working by the beach!!!)

BROSSEAU%SECAD1.HAC.COM@YMIR.CLAREMONT.EDU
BROSSEAU%SECAD1.HAC.COM@YMIR.BITNET
BROSSEAU%SECAD1@ENGVAX.SCG.HAC.COM  ( I think this one works the best!!)

dirk@mvax.dn.mu.oz (Dirk van der Knijff) (11/01/87)

I got this patch from the local tsc (Sydney, Australia).
It's incomplete but you should be able to figure it out.
 
in image sys.exe
 
PATCH>REPLACE/DECIMAL/WORD IOC$CLONE_UCB+15=9999
NEW>999
 
unfortunatly ioc$clone_ucb isn't a known symbol - you have to
look it up in sys.stb (in hex) then define a patch symbol (or
do a conversion to decimal). I have applied this to VMS 4.2 to
VMS 4.5, I don't know if it still works for VMS 4.6.
(Kermit won't work on a LTA line with >1000 also.
 
Dirk van der Knijff
Melbourne College of Advanced Education
757 Swanston Street		ARPA: dirk@mvax.dn.mu.oz.au@seismo.css.gov
MELBOURNE, Australia.		UUCP: ...!seismo!munarri!mvax.dn.mu!dirk

carl@CITHEX.CALTECH.EDU (Carl J Lydick) (11/01/87)

 > We are having a problem with Eunice and Photo.  When the LT or VT #'s 
 > rise above 1000 (yes our systems do stay up long enough to get there)
 > both Eunice and Photo stop working.  Eunice just does wierd things and 
 > Photo fails with the following error:
 > 
 > %SYSTEM-F-RESULTOVF, resultant string overflow
 > 
 > I believe the problem has to do with some definition in a system include
 > file that only sets up the field to hold the LT number as 5 characters 
 > when it should be 6. 

The problem is not a "system" include file: it is something within the source
for the EUNICE and PHOTO software.  Both were written back in the days when
the worst-case terminal name was shorter than it is now.  One way to avoid
the problem is to change the point at which cloned UCB's wrap.  The following
patch does this (or used to, but since it checks the old value, it should
either work or fail harmlessly; my advice, though, is to backup your system
disk before applying the patch, just in case.  Good luck.)
$!	Patch to VMS so MINed works with VTA devices (for V4.1 of VMS)
$!
$!	to module SYS file UCBCREDEL.MAR routine IOC$CLONE_UCB
$!	V4.1 fiche card 389 J07.
$!
$!	Description: IOC$CLONE_UCB is called to create all template devices.
$!	Normally the sequence counts 0-9999 are used, but MINed dies when run
$!	on a terminal if the sequence count is greater than 999.  This patch
$!	changes IOC$CLONE_UCB so that the device numbers wrap at 999 instead
$!	of 9999.
$!
$!	Caveats: This patch affects ALL template devices.
$!
$ PATCH SYS.EXE
DEFINE IOC$CLONE_UCB=8000DC8F
REPLACE/INST IOC$CLONE_UCB+0A
'CMPW B^54(R2),#^D9999'
EXIT
'CMPW B^54(R2),#^D999'
EXIT
UPDATE
EXIT
$!
$!(C) 1985 Caltech Oddhack Committee, No Rights Reserved

snsrmwl@nmtsun.nmt.edu (M. Warner Losh) (11/04/87)

I have heard a lot of talk about PHOTO.  Sounds like something that we
has on DEC-20 years ago.  Could someone tell me how to get a copy that
works.  I've heard various rumors about DECUS tapes, but I couldn't find
it in the listings (my searching skills are quite poor).

		Warner