Problem with KB896727 update

Recently we installed the MS05-038 as part of Windows Update. After this we
found out that our Code behind HTML application created in VJ++ is not
loading properly in IE 6 browser. We chekced all the Security settings and
they seem to work fine.
If we uninstall the Hotfix update ( KB896727 ) the application again loads
properly. Is there any work around for this issue. This is very critical.

Thanx

hi mart,

did you find an answer to this problem ?
I am still looking...

"mart" <> wrote in message
news:A042F759-C3EC-4FE1-9A98-...
> Recently we installed the MS05-038 as part of Windows Update. After this we
> found out that our Code behind HTML application created in VJ++ is not
> loading properly in IE 6 browser. We chekced all the Security settings and
> they seem to work fine.
> If we uninstall the Hotfix update ( KB896727 ) the application again loads
> properly. Is there any work around for this issue. This is very critical.


Looks to me that the thing is misnamed. Hotfix *regression* might be
a better description: <eg>

<title>MS05-038: Cumulative security update for Internet Explorer</title>
http://support.microsoft.com/kb/896727

<quote>
and if you want to install the hotfixes that are included in update rollup 896727,
you must follow the instructions in Microsoft Knowledge Base article 897225.
Otherwise, all Internet Explorer hotfixes that you have installed will be removed.
</quote>

Note that its results depend on whether you already had hotfixes integrated
after a previous rollup. So another consequence of it, if you previously
had some, is that you could now be getting any more contained within those
hotfixed (aka RTMQFE) modules. Many hotfixes have compatibility switches
in the registry which activate them but I don't know if all do nor do I know
if their default value is always off. More about analysing that factor later.


In order to analyse this situation what I would do first is use filever
(from the Support Tools package) to generate a list of the versions
that you have and a list of the versions that you had.

E.g. starting from the uninstall directory created by 896727, in a cmd
window enter:

for %d in (*.dll) do filever -v %d | find "FileVersion"

I'd use the above pipe to find FileVersion in the verbose output because
the regular output does not highlight whether the version is RTMGDR
or RTMQFE and you want to know which of the those you previously
had for each module.

Getting the same information for the ones which an update installed
for you could be trickier if they weren't all in the same location.
Fortunately, for 896727 all of them will be in %windir%\System32\
so you can use a modified version of the above command
EXECUTED FROM THE SAME DIRECTORY to find them. E.g.:

for %d in (*.dll) do filever -v "%windir%\System32\%d" | find "FileVersion"


Alternatively use the KB896727.log to see if it contains enough
clues about current version. In fact, choice of version may be one
of the things that the installer's /verbose switch would highlight;
so if there was any puzzle about which set of modules was being
chosen you might have to reinstall it manually using that option
in order to generate that extra log information.


Once you know which modules are hotfixed you will be able to tell
which compatibility switches you can look at. Alternatively, I suspect
that you could get a jump on those details by tracing the problem case
with RegMon e.g. filtering on Compatibility. Tracing both cases
with the same filter could highlight differences as a result of applying
the update and give you something simple which you could change
and test to try to restore compatibility.


For a related discussion read this thread

http://groups.google.com/group/micro...052a29c07113d6

(Google Groups search for
(KB896727 OR 896727) (MSFT OR MVP) (compatible OR compatibility OR incompatible OR incompatibility) group:microsoft.*
)


Good luck

Robert Aldwinckle
---

Thanx Robert,

I'll be in a testing environment shortly, and then I'll try your suggestions.

However, the problem appears with our clients, which are basic users with no
real system experience, and they are located all over the world with systems
that we do not control. It is not practical to analyse their individual
system.
Even to ask them to uninstall KB896727 and never to install it again will
sound exotic to them, let alone to ask them to compare versions.

So we will need to find the simplest of remedy and the simplest of
instructions for them. Maybe just "uninstall KB896727" ? The problem is that
then KB891816 kicks in (about 30 MB), with sometimes a peculiar behaviour,
like keeping to ask over and over to install.

What I cannot find an answer to is 'how dangerous is it to remove KB896727
and to keep a system without it ?".
And also, it is going to make a mess with the clients, when WindowsUpdate
keeps asking them to install KB896727, and they are not familiar or not
confortable with hiding a particular update.

And also, Robert :

Many of those users do not have administrator privileges. Imagine their
request to their system administrator :
Please remove Cumulative security update KB896727 on my PC and hide it for
ever.

How credible and reliable will we seem to make this kind of suggestion.

And also :

If you try to remove this Software Update, the Software Update Removal
Wizard is so kind to warn you with the list of all user programs and hotfixes
that you installed since KB896727, warning you that all these programs "might
not run properly" !

And then the polite question "Do you want to continue ?"

All those system administrators in the world will treat me as the next
joker, wouldn't they, if I would insist to remove KB896727.