Conference evms::spia

365.0. "PCSI$SOURCE problem" by COMICS::MILLSS ("Jump! Jump now!" ...Kosh) Fri Feb 07 1997 11:41

Hello.

Here's a good 'un...

According to Section 1.5.5 of the PCSI Utility Developer's Guide (from the V7.0
documentation CDs), 

>  PCSI$SOURCE defines the location of the software kits to
>  install. By default, the user's default device and directory are
>  used.

But experience has just shown that this is *not* the case,

> $ SHOW LOG PCSI$SOURCE
> %SHOW-S-NOTRAN, no translation for logical name PCSI$SOURCE
> $ DIR *.PCSI
>
> Directory DISK$USERS2:[MILLSS.SCRATCH]
>
> DEC-AXPVMS-DQS-V0103-1-1.PCSI;1
>
> Total of 1 file.
> $ PRODUCT EXTRACT RELEASE_NOTES DQS
> %PCSI-E-OPENIN, error opening
> PCSI$SOURCE:[MILLSS.SCRATCH]*-*-*-%%%%%-*-*.PCSI*; as input
> -RMS-F-DEV, error in device name or inappropriate device type for operation
> %PCSI-E-S_OPFAIL, operation failed
> %PCSIUI-E-ABORT, fatal error encountered - operation terminated

This clearly shows that PCSI is still trying to use the PCSI$SOURCE logical 
name even though it isn't defined.

Has anybody else come up against this? Any helpful hints or suggestions? Maybe I
should just IPMT this...?

Many thanks,

Simon R. Mills
OpenVMS Group, UK CSC

   This one is known behaviour -- donno if it's fixed in V7.1 (there
   were various updates for V7.1), but it's been this way for a while.
   See 182 amnd 286 for some previous similar discussions.

   It's worthy of a medium-low QAR, if nothing else.

   PCSI should likely $trnlnm the logical name, and $crelnm the logical
   name with the translation, or -- if no translation is found -- with
   the translation SYS$DISK:[].  Or some such similar code path...

This problem has been fixed in OpenVMS V7.0.  I just logged on to a V7.0
system to verify the behavior.  You mentioned that you referenced the V7.0
documentation, but what version of OpenVMS are you running?  Your error
messages look like they came from a V6.1 or V6.2 system when PCSI incorrectly
used pcsi$source as the device component for the RMS search call without
checking to see if the logical was defined.

Jim Krycka

Re .1 and .2

Thanks for the replies. I just tried this out on a V7.0 system and it worked as
documented. It was a V6.2 system that I reproduced it on in the base note,
that's because my customer is running V6.2

Thanks and regards,

Simon.