"Paveza, Gary" wrote:
>
> Unless I am mistaken (and it is possible), SUBSYS tapes are usually
> released with powerpatches.
>
> But there aren't SUBSYS tapes without powerpatches.

Speaking of Express Releases you are correct, they either consist
of a PowerPatch or a SUBSYS + PowerPatch.
>
> Why not just let the HPVERSION be determined by the powerpatch level?
> C.55 should indicate that the user is on MPE/iX 5.5
> and .06 should indicate powerpatch 6
> so you end up with C.55.06
> Would definitely save on confusion.

A bit of more background:

When Jon refers to the PowerPatch and SUBSYS V.UU.FF, then this is the
one you will find on the tape sticker/label and in the reference manuals
(Communicator, Read before installing, etc.).

It does not necessarily equal - and speaking from a support perspective,
this is causing the largest part of confusion - the V.UU.FF shown during
HELLO or SHOWME or as HPVERSION.

Many of you will have seen systems that show a changed USER VERSION after
an Express Release installation, like this:

:SHOWME
USER: #S13,MANAGER.SYS,SYSMGR      (NOT IN BREAK)
RELEASE: C.55.00   MPE/iX HP31900 C.05.08   USER VERSION: C.55.03

while others install PowerPatch C.55.06 and still get USER VERSION: C.55.00
shown.

The reason for that is that our installation tools in their current form
use SUBSYS VUUFFs found in installation jobs from the Express SUBSYS tape.
I will not go into more details, but if you install an Express subsys
with the VUUFF Z.99.02 and a PowerPatch with the VUUFF Z.99.04 your
USER VERSION and HPVERSION will show the Z.99.02.

If you only install PowerPatch but no Express SUBSYS the last 2 digits
do not change.

The USER VERSION is user changable through sysgen, and we actually have
customers that use this for their own purposes of release control, typically
with large numbers of systems where a (fictional) USER VERSION of C.B0.10
tells them not only which OS release and patches, but also which partner
tool versions and application versions they have installed on their systems.

From my support point of view I would prefer that our tools actually change
the USER VERSION based on the PowerPatch version in general, while still
allowing the abovementioned user control of the USER VERSION easily.

BTW: for software developers (internal and external) we have introduced
two
other variables that are read-only, available since MPE/iX 5.0:

:showvar hp@version
HPOSVERSION = C.05.08     (that's the HP internal build ID)
HPRELVERSION = C.55.00    (that is reliable, cannot be overwritten)
HPVERSION = C.55.05       (this one was changed manually in SYSGEN)

It is arguable if HPRELVERSION should show the 'base' release VUUFF or
the PowerPatch VUUFF. Some developers may want to find out if a certain
new functionality introduced through an enhancement patch (think of recent
CI enhancements for example) is available on a system. But I would not
take
a VUUFF of C.55.03 as an reliable indicator that e.g. the CI word() function
is available. (I do not remember with which PP this was introduced, just
making up an example here).
After all users have to force enhancement patches, can veto other
patches now,
or one patch could have failed to qualify. And we certainly not go into
the
chaos of changing the VUUFF with each single patch...

Hope this clears some of the fog,

Goetz.