Uninstalling a package

Blair,

I did some fiddling with loose methods (packaged some of them, and
archived/removed others) and rebuilt a beta 2 image.  I can't see my beta 2
image at home right now to confirm it, but, I recall removing 'SUnitTests';
now I get a walkback (below).  Is this something that beta 2 does?

Have a good one,

Bill

----------

3:53:25 PM, Monday, February 18, 2002: 'STBFiler - Input stream not in STB
format'
STBInFiler class(STBFiler class)>>errorNotSTB
STBInFiler>>reset
STBInFiler(STBFiler)>>initialize
STBInFiler>>initialize
STBInFiler class(STBFiler class)>>on:
ResourceSTBByteArrayAccessor>>hiddenObjects
ResourceSTBByteArrayAccessor(ResourceAccessor)>>hiddenClassReferences
ViewResource(Resource)>>hiddenClassReferences
[] in Package>>tracePrerequisitesOfResources:
Set>>do:
Package>>tracePrerequisitesOfResources:
Package>>tracePrerequisites:
Package>>tracePrerequisites
Package>>calculatePrerequisites
Package>>buildPrerequisiteNames
Package>>prerequisiteNames
Package>>prerequisites
[] in PackageManager>>dependentPackagesOf:
[] in PluggableLookupTable(LookupTable)>>select:
PluggableLookupTable(LookupTable)>>keysAndValuesDo:
PluggableLookupTable(LookupTable)>>select:
PackageManager>>dependentPackagesOf:
PackageManager>>allDependentPackagesOf:visited:
Package>>allDependentPackages
[] in PackageBrowserShell>>confirmUninstall:
BlockClosure>>ifCurtailed:
BlockClosure>>ensure:
Cursor>>showWhile:
PackageBrowserShell>>confirmUninstall:
PackageBrowserShell>>uninstall:
[] in PackageBrowserShell>>removePackage
ExceptionHandler(ExceptionHandlerAbstract)>>markAndTry
[] in ExceptionHandler(ExceptionHandlerAbstract)>>try:
BlockClosure>>ifCurtailed:
BlockClosure>>ensure:
ExceptionHandler(ExceptionHandlerAbstract)>>try:
BlockClosure>>on:do:
[] in PackageBrowserShell>>removePackage
Array(ArrayedCollection)>>do:
PackageBrowserShell>>removePackage
Symbol>>forwardTo:
[] in Command>>value
BlockClosure>>ifCurtailed:
BlockClosure>>ensure:
Command>>value
PackageBrowserShell(Shell)>>performCommand:
CommandQuery>>perform
DelegatingCommandPolicy(CommandPolicy)>>route:
[] in ContainerView(View)>>onCommand:
BlockClosure>>ifCurtailed:

--
Wilhelm K. Schwab, Ph.D.
[hidden email]

"Bill Schwab" <[hidden email]> wrote in message
news:[hidden email]...
> Blair,
>
> I did some fiddling with loose methods (packaged some of them, and
> archived/removed others) and rebuilt a beta 2 image.  I can't see my beta
2
> image at home right now to confirm it, but, I recall removing
'SUnitTests';
D5 (and indeed D4 as far as I recall) traces into view resources to identify
pre-requisite packages. When uninstalling a package, it has to do this in
order to identify dependent packages that also need to be removed. The
walkback suggests that for some reason one of the view resources in your
image at the time of uninstallation is (for some reason) corrupt. If you can
supply a reproduceable sequence that would help.

Regards

Blair

Blair,

> D5 (and indeed D4 as far as I recall) traces into view resources to
identify
> pre-requisite packages. When uninstalling a package, it has to do this in
> order to identify dependent packages that also need to be removed. The
> walkback suggests that for some reason one of the view resources in your
> image at the time of uninstallation is (for some reason) corrupt. If you
can
> supply a reproduceable sequence that would help.

I think I found the view resource, and was able to fix it with the standard
trick of exporting from 4.01 and importing/saving in beta 2.  It's possible
that this has been lurking for a while because I was concentrating on
getting ready to deploy on NT/2k w/o breaking my current ability to deploy
on 9x - it's possible that this was my first attempted deployment on beta
2????  The odd thing is that the resource in question is one that I'd bet a
beer (maybe two<g>) that I had dragged into 4.0 using the same
export/import/save trick, with 3.06 as the source; I remember needing it and
being suprised that it hadn't crossed over to 4.0.  If my memory is serving
well, it seems like it should have already been saved in the 4.01 VC (after
the import from 3.06), so I'm not sure why exporting it separately from 4.01
would help.  Am I missing something unique about exporting a view resource?

A good starting point might be to make a new package in 3.06 and see if
installing that causes the problem in beta 2.

Have a good one,

Bill

--
Wilhelm K. Schwab, Ph.D.
[hidden email]

Blair,

> A good starting point might be to make a new package in 3.06 and see if
> installing that causes the problem in beta 2.

That works, but, loading the real package as saved by 4.01 seems to
reproduce the problem.  Sadly, it's tied into a lot of stuff, so it's hard
to call it an isolated example.  Is there a way to grab the view resource's
byte array from the 4.01 image and reload it into another image?   That
might bypass something that's masking a problem???  It'll be interesting to
see if I can reproduce this at home (Win98).  I'm having the problem on NT4.

Another idea is to try splitting the package; the part that's causing
trouble might be small enough to isolate if we're lucky.

Have a good one,

Bill

--
Wilhelm K. Schwab, Ph.D.
[hidden email]

Blair,

> That works, but, loading the real package as saved by 4.01 seems to
> reproduce the problem.  Sadly, it's tied into a lot of stuff, so it's hard
> to call it an isolated example.  Is there a way to grab the view
resource's
> byte array from the 4.01 image and reload it into another image?

Remembering the hex string stuff, I chopped a chunk out of the pac file and
emailed it to you.  Please let me know whether or not you get a walkback
after filing it in and editing the resulting view resource with the VC.
Failing that, I'll keep digging.

Have a good one,

Bill

--
Wilhelm K. Schwab, Ph.D.
[hidden email]