VMS 8.2 VMSINSTAL Bug with RUN_IMAGE Persists?
Richard Maher
There was/is a bug with VMSINSTAL's Run_Image callback on 8.2 machines that
exhibited the following symptoms at product installation time: -
* Do you want to purge files replaced by this installation [YES]?
%ANALYZE-E-OPENIN, error opening MISSING:[MISSING]VMI$NEW.EXE; as input
-SYSTEM-W-NOSUCHDEV, no such device available
%VMSINSTAL-E-INSFAIL, The installation of TIER3A V3.1 has failed.
The problem is that I've just had someone upgrading to 8.2 (Why not 8.3? I
dunno :-) and the bug appears to still be there :-(
VMS 8.2 is still supported, is it not? Has anyone got a patch reference:
Below is from one of COVs kind souls many moons ago stating that this bug is
*supposed* to have been fixed.
Any ideas?
Cheers Richard Maher
> 2. VMSINSTAL
>
> This failed miserably on VMS V8.2 on running T3$CHKACC, because there
> is a known bug in VMSINSTAL on V8.2 with the RUN_IMAGE vmi$callback.
>
> It craps out with:
>
> %ANALYZE-E-OPENIN, error opening MISSING:[MISSING]VMI$NEW.EXE; as input
> -SYSTEM-W-NOSUCHDEV, no such device available
> %VMSINSTAL-E-INSFAIL, The installation of TIER3A V3.1 has failed.
>
>
> This issue is allegedly fixed in VMS82A_KITTING-V0100, dated
> 10-JAN-2006, but that contains a copy of VMSINSTAL.COM from
> 10-AUG-2005, so doesn't solve the problem.
>
> Needless to say, I'm going to report that problem. I got Tier3
> installed by knobbling its KITINSTAL.COM thus:
>
> $! VMI$CALLBACK RUN_IMAGE "VMI$KWD:T3$CHKACC.EXE"
> $ tier3$missing = "true"
Robert Deininger
<mahe...@hotspamnotmail.com> wrote:
>Hi,
>
>There was/is a bug with VMSINSTAL's Run_Image callback on 8.2 machines that
>exhibited the following symptoms at product installation time: -
>
>* Do you want to purge files replaced by this installation [YES]?
>%ANALYZE-E-OPENIN, error opening MISSING:[MISSING]VMI$NEW.EXE; as input
>-SYSTEM-W-NOSUCHDEV, no such device available
>%VMSINSTAL-E-INSFAIL, The installation of TIER3A V3.1 has failed.
>
>The problem is that I've just had someone upgrading to 8.2 (Why not 8.3? I
>dunno :-) and the bug appears to still be there :-(
>
>VMS 8.2 is still supported, is it not? Has anyone got a patch reference:
V8.2 Alpha is still supported.
V8.2 I64 supported ended Dec. 31, 2006.
Peter 'EPLAN' LANGSTOeGER
>V8.2 Alpha is still supported.
>
>V8.2 I64 supported ended Dec. 31, 2006.
Yup, OpenVMS I64 V8.2-1 is supported parallel to OpenVMS Alpha V8.2 (Dec2008)
--
Peter "EPLAN" LANGSTOEGER
Network and OpenVMS system specialist
E-mail pe...@langstoeger.at
A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist
Richard Maher
>
> V8.2 Alpha is still supported.
Fabulous! So then one could be forgiven for thinking that bug-fixes
(especially those as simple as "Here's another copy of the 3.5K line DCL
command file VMSINSTAL.COM") would be forthcoming? Hell, it works again in
8.3; why can't they just put that version on a patch kit and say here it is?
This is my take on what's happened: -
1) Digital employees can't help but get all excited over bullshit concepts
such as "seperating Form and Funtion" or providing IT's Holy Grail of a
"Truly *NON*-procedural" software installation kit. (Yawn)
2) No real person shares their bollocks enthusiasm, so we are punished by
having our Developer's Guide to VMSINSTAL stuffed into the Archived
Documents section.
3) No longer content with merely spitting at us as he walks by, some wanker
decides that RUN_IMAGE: should actually do an Analyze/Image on the file
first (to see if it is really an image?) (to get that warm fuzzy feeling?),
but not bother to test it :-(
4) It get's fixed for 8.3 but he won't backport it to 8.2 'cos that famous
Digital arrogance tells him that "No one should be using RUN_IMAGE anyway"
or "Your code was always broken" or "It's a feature" or "The work around is
to upgrade to 8.3" you little shit of a customer! (Move to PCSI before I
fart in your general direction, you objectionable little man)
Sound about right?
All I want to do is, at install time, run a little program that checks the
UAF. Is there a bug/patch reference I can pass on to people so that their
path through HP support can be smoothed if they are lucky enough to come
across this issue?
Cheers Richard Maher
"Robert Deininger" <rdein...@mindspringdot.com> wrote in message
news:rdeininger-24...@dialup-4.233.173.47.dial1.manchester1.level3.net...
Robert Deininger
<mahe...@hotspamnotmail.com> wrote:
>Hi Robert,
>>
>> V8.2 Alpha is still supported.
>
>Fabulous! So then one could be forgiven for thinking that bug-fixes
>(especially those as simple as "Here's another copy of the 3.5K line DCL
>command file VMSINSTAL.COM") would be forthcoming? Hell, it works again in
>8.3; why can't they just put that version on a patch kit and say here it is?
I'm not familiar with the details of whatever VMSINSTAL problem you are
complaining about. That's ok, since this isn't a support forum, and
everyone understands that you won't get formal support here.
If you found a bug in VMSINSTAL, log a support call. I assume you have a
support contract. If you don't have one, then VMS V8.2 on Alpha is NOT
"supported" for you, and you are relying on volunteer/charity work to help
you with your problems.
I understand that it is work to log a support call, and more work if you
do it well than if you do it poorly. (Your problem description in the
original post definitely falls in the "poorly" category. But that's ok
because this isn't a support forum.) You need to supply answers to a
number of questions, both technical and bureaucratic. You need to
understand the problem well enough to explain it. ("It's broken!" is NOT
a useful problem report, no matter how many pages of flowery,
confrontational, non-specific prose are used to say "It's broken".)
Sometimes you need the patience to explain the problem to someone whose
understanding is much less than your own.
So I understand that it is easier to wait for someone else to log the
call, and hope that a solution appears for "free" in the form of a patch
kit. And it's fun to bitch in public while you wait for the solution you
hope someone asked for.
If you log a call for a bug in a supported product, and your problem
report is reasonable, and the problem can be reproduced, then you should
expect HP to provide a solution, TO YOU. Not every solution turns into a
patch kit, for a number of reasons. If you need the solution to turn into
a patch kit, the first, and obvious step, is to ASK for one, and perhaps
explain why it's needed.
Most, if not all, VMS patch kits are explicitly requested, either by a
customer or by a VMS developer. And VMS developers do NOT request kits
for every bug they fix. Doing so would overwhelm the patch delivery
process with very little benefit to anyone. In the worst case, the fixes
will ship in the next release. More often, batches of fixes will go out
in patch kits when they are collectively urgent enough to justify it.
>This is my take on what's happened: -
>1) Digital employees can't help but get all excited over bullshit concepts
>such as "seperating Form and Funtion" or providing IT's Holy Grail of a
>"Truly *NON*-procedural" software installation kit. (Yawn)
>2) No real person shares their bollocks enthusiasm, so we are punished by
>having our Developer's Guide to VMSINSTAL stuffed into the Archived
>Documents section.
>3) No longer content with merely spitting at us as he walks by, some wanker
>decides that RUN_IMAGE: should actually do an Analyze/Image on the file
>first (to see if it is really an image?) (to get that warm fuzzy feeling?),
>but not bother to test it :-(
>4) It get's fixed for 8.3 but he won't backport it to 8.2 'cos that famous
>Digital arrogance tells him that "No one should be using RUN_IMAGE anyway"
>or "Your code was always broken" or "It's a feature" or "The work around is
>to upgrade to 8.3" you little shit of a customer! (Move to PCSI before I
>fart in your general direction, you objectionable little man)
>
>Sound about right?
Er, no.
I'm having my usual trouble understanding your prose. (I don't remember
what misjudgement led me to actually read it.) Are you seeking a solution
to a problem or are you just venting? Do you enjoy writing these little
ditties that border on hysterical and offensive? So you think a lot of
people are entertained by them? Do you think your rants make folks WANT
to help you? Have you ever considered that you might get better results
asking for help if you weren't a pain in the <anatomy>?
To the extent I do understand what you wrote above, it's seems almost
totally disconnected from the real world. I detect nothing in your
description that matches how VMS design decisions and support cases are
handled. I can't address your points because none of my experience
connects with them.
>All I want to do is, at install time, run a little program that checks the
>UAF. Is there a bug/patch reference I can pass on to people so that their
>path through HP support can be smoothed if they are lucky enough to come
>across this issue?
That's a PERFECT question for HP support. Have you asked there?
Larry Kilgallen
> Hi Robert,
>>
>> V8.2 Alpha is still supported.
>
> Fabulous! So then one could be forgiven for thinking that bug-fixes
> (especially those as simple as "Here's another copy of the 3.5K line DCL
> command file VMSINSTAL.COM") would be forthcoming? Hell, it works again in
> 8.3; why can't they just put that version on a patch kit and say here it is?
I have not been following this thread.
Are you saying that you reported a problem through formal support channels
and they are failing to address it when you can show they know how to fix
it ?
If that is the case, you should demand the money back you paid for that
support contract.
FredK
> In article <f0md3l$4qk$1...@news-01.bur.connect.com.au>, "Richard Maher"
> <mahe...@hotspamnotmail.com> wrote:
>
>>Sound about right?
>
> Er, no.
>
> I'm having my usual trouble understanding your prose. (I don't remember
> what misjudgement led me to actually read it.) Are you seeking a solution
> to a problem or are you just venting? Do you enjoy writing these little
> ditties that border on hysterical and offensive? So you think a lot of
> people are entertained by them? Do you think your rants make folks WANT
> to help you? Have you ever considered that you might get better results
> asking for help if you weren't a pain in the <anatomy>?
>
I have to say that despite the odd chuckle I get when Richard actually lands
a bullseye, I for the most part don't read what he writes - because of the
invective. Participation in COV isn't part of our job descriptions - so I
don't have to respond to things that are just too far over the top.
Sometimes it isn't worth trying to pull out the nugget of a question or
suggestion to address.
Given the other streams of complete time-wasting-stupidity in COV the last
few weeks, it was easy to overlook. I only went *back* and read it to try
to figure out what this was all about.
I appreciate that Richard is frustrated, but his reaction to the frustration
doesn't win him many points.
>>1) Digital employees can't help but get all excited over bullshit concepts
>>such as "seperating Form and Funtion" or providing IT's Holy Grail of a
>>"Truly *NON*-procedural" software installation kit. (Yawn)
I don't work for Digital (RIP), I work for HP. Having had to deal with
VMSINSTAL on both ends (creating installations and installing products) -
PCSI with all it's faults is lightyears better. While I had nothing to do
with this particular issue - frankly I find nothing in this bullet except a
insult to all former Digital employees.
Feel free to look me up in the source listings.
The engineer that introduced the problem has retired - and was an excellent
kernel engineer and spent a lot of time working with customers issues. The
guy who fixed the problem is about as customer focused an engineer that I
know. Both of whom you would do well to be nice to, one of whom I believe
reads COV but seldom writes. Both of whom you can find in the source
listings as well.
>>2) No real person shares their bollocks enthusiasm, so we are punished by
>>having our Developer's Guide to VMSINSTAL stuffed into the Archived
>>Documents section.
So you don't like PCSI. It's like complaining that we don't use PIP for
backup. It isn't like we stopped supporting VMSINSTAL - heck we ported it
to IPF which is where the bug was introduced.
But clearly *we* believe that PCSI is a superior method for someone who
wants to develop a product installation. We invested a lot of time and
energy into creating something better than VMSINSTAL.
>>3) No longer content with merely spitting at us as he walks by, some
>>wanker
>>decides that RUN_IMAGE: should actually do an Analyze/Image on the file
>>first (to see if it is really an image?) (to get that warm fuzzy
>>feeling?),
>>but not bother to test it :-(
It was done to determine image architecture when used on IA64. Not tested?
Apparently not enough. Or at least not with a kit that used that feature.
Not really an excuse. But it was hardly done on a whim - it was a TYPO with
a symbol name. Frankly it appears that the code had been checked in for
several *years* before someone stumbled onto the problem.
>>4) It get's fixed for 8.3 but he won't backport it to 8.2 'cos that famous
>>Digital arrogance tells him that "No one should be using RUN_IMAGE anyway"
>>or "Your code was always broken" or "It's a feature" or "The work around
>>is
>>to upgrade to 8.3" you little shit of a customer! (Move to PCSI before I
>>fart in your general direction, you objectionable little man)
The guy who fixed the problem checked it into V8.2R. Backporting isn't an
issue. The sole issue is that it hasn't been included in any V8.2 patch
kit.
Have you requested one? I searched for the snippet you quote from one of
"COVs kind souls" and can't find it. So I assume that was from an e-mail.
Did he report it?
To be honest, we try to generate patch kits for the "mainstream" versions -
even when the mainstream is sometimes beyond support end of life (like
V7.3-2). V8.2 falls into the category of "lightly used" - so patches there
tend to get generated when a customer requests one.
Richard Maher
> I have to say
Having been aware of your aggressive disinterest :-) in my posts for some
years now, I personally fail to understand the compulsion, but regardless of
what has motivated you to break-cover on this occasion I'm certainly glad
that you're here! (And if you could stay around for one or two questions
later on that EXEC mode lib$*_vmABLE heap then that would be invaluable)
> I have to say that despite the odd chuckle I get when Richard actually
lands
> a bullseye, I for the most part don't read what he writes - because of the
> invective.
Look, no one reads the crap I write. I know this 'cos everytime I pen
something that's the slightest bit "edgy", I have a conga-line of bullshit
artists lining up to tell me how they only stumbled across it due to some
freak alignment of the stars. Maybe I didn't get a bullseye this time but
it's a big wunnnhundredaaandeiiightty! from my point of view 'cos (because
of your reply, and thanks for that) I've come away with what I was after.
Would I have preferred it if you had have replied to my first post - yes.
Would I prefer more friends than enemies - yes. Did I wish I didn't have to
flick your ear and tweak your nose to get the attention my request
deserves - of course. Do I wake up in the morning and say "Let's piss
someone off today" - very rarely.
I do not come here to hear my own voice regardless of how fascinating I find
myself. I come here for a result! And although I certainly don't get one
every time, I do alright thanks very much.
> Participation in COV isn't part of our job descriptions - so I
> don't have to respond to things that are just too far over the top.
> Sometimes it isn't worth trying to pull out the nugget of a question or
> suggestion to address.
Apart from the collective madness that seems to draw all of us here to the
COV well, I think it's fair to say that at least 90% of the people here
(most definately including myself) have VMS's best interests at heart, and
if we can proceed on that premise and are willing to view my art through
that prism then we should get along famously :-)
> Given the other streams of complete time-wasting-stupidity in COV the last
> few weeks, it was easy to overlook. I only went *back* and read it to try
> to figure out what this was all about.
This is the last time that I'm gonna bother defending myself (and is
indicative the level of respect I hold you and Hoff an almost all of the VMS
engineering team in) but just have a look at Google statistics for COV and
see exactly who is contributing what! I'm sorry if my technical questions
are not up to the Jerry Springer confrontationalism that you've come to
expect from COV, but you can alwasy just hit "block sender"
Is it the profanity that bothers you, and that has "Misery" Deninger doing
his Kathy Bates impression?
While I won't contend that contextualised volume phrases such as "useless
wanker" and "pointless knob-head" are exactly terms-of-endearment, they're
hardly an attack on everything you hold sacred in your life either. How did
you get through school if you wanted to fight someone every time they called
you a wanker? (Maybe it was only me that it heard it so often :-)
> I appreciate that Richard is frustrated, but his reaction to the
frustration
> doesn't win him many points.
Probably not; for winning points I rely on loyalty, perseverance, talent,
being the author of the software that will single-handedly turn VMS fortunes
around in the next decade, and of course charm.
> I find nothing in this bullet except a
> insult to all former Digital employees.
I'm not above doing that; I've done it before and I dare say I'll have
occasion to do it again. I have also been known to dish out the odd bit of
praise and grattitude as well you know?
> Feel free to look me up in the source listings.
No need for gratuitous Willy-Waving here Fred. We all know your
contributions to VMS and no one (certainly not I) is questioning your
credentials. (Bummer about DECWindows but eh?)
> The engineer that introduced the problem has retired - and was an
excellent
> kernel engineer and spent a lot of time working with customers issues.
The
> guy who fixed the problem is about as customer focused an engineer that I
> know. Both of whom you would do well to be nice to, one of whom I believe
> reads COV but seldom writes.
I just don't think it is unreasonable for a fix for a known bug to have gone
out in some form of consolidated patch set sometime in the last three or so
years. And for the last time, yes I can be nice to people and yes I like
friends, but respect is a two-way street and I don't give a monkey's about
having enemies either.
> Both of whom you can find in the source
> listings as well.
You gotta learn to let go, you seem bruised or scarred about my previous
attacks on Steve. "Go on - I'll take you all on then" Bittabiffo! You
really shouldn't let someone who is essentially a nobody get under your skin
like this. Ignore me - and I'll go away. Answer my questions - and I'll go
away. What's the big deal? It's a newsgroup; nobody dies here!
> But clearly *we* believe that PCSI is a superior method for someone who
> wants to develop a product installation.
With all due respect, who cares? Tell it to Rdb engineering. What does
Oribble use?
> We invested a lot of time and
> energy into creating something better than VMSINSTAL.
So it's "better" than VMSINSTAL? It's still crap IMHO and not worthy of the
development budget it was given. But who really cares about how you install
software on VMS? I just want you to fix what you break! (What a pain in the
arse that seems to make me)
> The guy who fixed the problem checked it into V8.2R. Backporting isn't an
> issue. The sole issue is that it hasn't been included in any V8.2 patch
> kit.
Is R for Remedial? Has there been no consolidated kit go out since it was
fixed?
> Have you requested one? I searched for the snippet you quote from one of
> "COVs kind souls" and can't find it. So I assume that was from an e-mail.
> Did he report it?
Yes it was an e-mail and the author looked to be away for a while but has
started posting again yesterday. I'm sure he wouldn't have minded being
named, but now that I didn't, I really can't. So hey you - "Did you report
it?". I'll mail him if he doesn't respond.
> To be honest, we try to generate patch kits for the "mainstream"
versions -
> even when the mainstream is sometimes beyond support end of life (like
> V7.3-2). V8.2 falls into the category of "lightly used" - so patches
there
> tend to get generated when a customer requests one.
I don't know why there going 8.2 at this stage, but the main thing is that I
can now fully describe what's gone on so, once again, thanks.
Cheers Richard Maher
"FredK" <fred....@dec.com> wrote in message
news:f0npo3$7sp$1...@usenet01.boi.hp.com...
Richard Maher
> Have you requested one? I searched for the snippet you quote from one of
> "COVs kind souls" and can't find it. So I assume that was from an e-mail.
> Did he report it?
Well I e-mailed him again to ask, and either he's ignoring me, I'm in his
kill-file, or my mail's just not getting through. Either way, I found a
previous reply confirming that the report did in fact get to VMS
Engineering. (See below)
Cheers Richard Maher
Sent: Friday, August 04, 2006 5:25 AM
Subject: VMSINSTAL problem now forwarded to OVMS Engineering
> Just to keep ya up to date, my report has been accepted and forwarded
> to the relevant OVMS Engineering folks.
> --
Richard Maher
I know you're all dying to know how this turned out. Apparently the first
mail (and quite a few other one's from more important people :-) didn't get
through, and the last one was late. In any event here is the feed-back: -
> I don't have a support contract, though I did make sure it got
> through to VMS Engineering.
> on an unofficial basis.
So clearly I am not the only one to experience the full brunt of that
delightful VMS engineering arrogance that I speek so much of. And there was
I feeling special :-(
"Don't have a support contact? Those stinking pigs! Get me a knife; I'm
going to cut VMS's nose off right now! There are procedures and protocols
and pecking orders to preserve here and we'd much rather let VMS sink than
fix a hole that hasn't been reported in the proper fashion. Lock that
degenerate, impertinent, filth down there with Leonardo Di Caprio - Their
unpleasantness is upsetting our view of those lovely ice bergs. I mean, how
much water is it letting in anyway.".
Cheers Richard Maher
"Richard Maher" <mahe...@hotspamnotmail.com> wrote in message
news:f1rgbn$qpq$1...@news-01.bur.connect.com.au...
FredK
> Hi,
>
> I know you're all dying to know how this turned out. Apparently the first
> mail (and quite a few other one's from more important people :-) didn't
> get
> through, and the last one was late. In any event here is the feed-back: -
>
>> I don't have a support contract, though I did make sure it got
>> through to VMS Engineering.
>> on an unofficial basis.
>
> So clearly I am not the only one to experience the full brunt of that
> delightful VMS engineering arrogance that I speek so much of. And there
> was
> I feeling special :-(
>
> "Don't have a support contact? Those stinking pigs! Get me a knife; I'm
> going to cut VMS's nose off right now! There are procedures and protocols
> and pecking orders to preserve here and we'd much rather let VMS sink than
> fix a hole that hasn't been reported in the proper fashion. Lock that
> degenerate, impertinent, filth down there with Leonardo Di Caprio - Their
> unpleasantness is upsetting our view of those lovely ice bergs. I mean,
> how
> much water is it letting in anyway.".
>
I'm trying to stifle my gut reaction to call you an... The bug was fixed -
not ignored. The guy who fixed it may even have given the guy
"unofficially" complaining a fix (or told him the one-line edit to the
command procedure). The fact that it didn't get released as a V8.2 fix has
nothing to do with the price of tea in China and was discussed in response
to your earlier screed.
We do not ignore bugs simply because they were found or reported through
non-official channels. However how we respond and release a fix for a
problem *can* be effected by it - because like it or not - customers paying
for service are the ones who come first. **IF** you or your friend had a
support contract and asked for a fix for V8.2 - then you would have gotten a
fix - and if you insisted on it - an official ECO kit for V8.2. Instead the
fix went into the mainstream changes and not into a special V8.2 kit -
because it is a low-priority problem on a version that isn't mainstream.
How do you believe we should respond to problems reported by hobbiests?
What then is the difference between someone who pays for a license and
service and someone who reports something on COV or via an e-mail to an
engineer?
I don't know what it is with you. Everything is some conspiracy. Everyone
is arrogant, or evil, or lazy, or stupid or some other nasty name you want
to throw out. If there isn't an immediate response to whatever you happen
to be hot-to-trot on - it must be because we're all idiots.
JF complained a long time ago about a CERT regarding a malformed image file
that could create a buffer overflow. Making it do something really bad
other than crashing the application on VMS would be nearly impossible.
Hey - there was no bug report - just a complaint in COV about a obscure
CERT. But I used my contacts to get the source to the fix and had the
engineers here fix the problem. Guess what? It too was never released as
an explicit ECO - because NO CUSTOMER has ever seen the problem. Heck, in
the real world no UNIX customer ever saw it.
How we respond to a problem is context dependent.