***** sp2 bREAKS eVERYTHING!!!!! *****

  • Thread starter Thread starter nAN
  • Start date Start date
On Fri, 3 Sep 2004 16:38:12 -0600, "Jupiter Jones [MVP]"
"Internet driving licence" would probably help

As would "Certificate of Roadworthyness" for computers!


-------------- ---- --- -- - - - -
"I think it's time we took our
friendship to the next level"
'What, gender roles and abuse?'
 
There is nothing crippled about an OEM copy of XP other than
that it can't be transfered to a new computer.

YMMV there.

DSP and generic OEM XP CDs are as OK as you claim, but MS allows big
OEMs the right to cripple the OS CD further, i.e. any of:
- inability to custom or repair install ("wipe and rebuild" only)
- non-bootable and/or no Recovery Console
- other missing items, e.g. no Backup utility
- no ability to install the OS at all; "call your vendor"
- no CD at all; only the HD contents

All of the above are legal, from MS's licensing perspective, though
they may offer the user considerably less value than warez CDR.


-------------------- ----- ---- --- -- - - - -
Reality is that which, when you stop believing
in it, does not go away (PKD)
 
Ok, where is the explanation for the string of folks who all have a
pretty similar description, "click on a file and Windows Explorer crashes"?

Let's get some detail there, else it's GIGO (Garbage In, Garbage Out),
e.g. I'd just answer "likely there is something wrong with the
persistant handler or default action for the file you clicked on"

Which file? Where is the file? In what way does it crash?

No, there are a number of different issues, many of which can be
anticipated and/or ameliorated.

Most go about things that SP2 is designed to block, i.e. apps that
rely on practices we now see as risky, and would prefer to block.
There's good documentation on these; lists of what software is
affected, and how, and how to fix.

But there are some specific issues that go about particular system
specifications or components, and that have effects beyond "everything
works except this particular application".

Then there are generic issues that can be expected to blow up any
attempt to install large wads of system code; flaky hardware, active
malware infections, and so on.
The chant "it is all viruses and spyware and stupid users" has been
posted in this newsgroup hundreds of times in the last week.

It's chicken and egg there. You want SP2 to keep these systems safer,
but many of these systems are already infested. So often the PCs that
would benefit the most from SP2, have adverse mileage because they are
riddled with malware before SP2 is in place and weren't cleaned first
I install SP2 and Windows Explorer in every use on the system immediately
locks up, refuses any keystroke or mouse click.

I'd really like more detail on that, as it's not a failure pattern
I've seen in posts here. How does YMMV if you:
- run Safe Mode?
- suppress all in MSConfig?
- use Shell Extension Viewer to suppress CLSID integrations?
- use something other than Explorer.exe as shell=?
- kill off Common Tasks and (if possible) Desktop.ini processing?
- disable your LAN card in this profile to suppress networking?
- suppress BHOs via IE's Tools, Options, Advanced?
- NTFS vs. FATxx mileage differences?

I'd be thinking of integrations into Explorer that are now blowing up
for one reason or another - not malware per se, but something added by
software or drivers, or broken by any number of things.

When Windows Explorer first starts, it may break on namespace
enumeration issues - things that aren't really drives but pretend to
be, that sort of thing.

Then, when it tries to list files in a folder, it may break on
persistent handlers, i.e. thumnailers or other code that runs when
files are enumerated, before any attempt is made to select or "open"
them. Antivirus scanners may impact at this point.

Then when you click or hover over something to select it, there may be
further persistent handler code run to populate the status bar info or
tooltip, e.g. things like .DOC authorship or .ZIP content file counts.

Then if you rt-click, anything that integrates into the contexct menu
gets a crack. Things that are listed "above the line" are usually
passive and don't run code unless clicked, whereas things that have
icons displayed, cascade out as additional menus, or are listed below
the top line in the context menu are running as code via a CLSID

Once you click a context menu entry, or press Enter on a file, or
double-click a file, you are not in Explorer anymore - extra code
beyond Explorer itself is more likely the problem.


------------ ----- ---- --- -- - - - -
The most accurate diagnostic instrument
in medicine is the Retrospectoscope
 
Let's get some detail there, else it's GIGO (Garbage In, Garbage Out),
e.g. I'd just answer "likely there is something wrong with the
persistant handler or default action for the file you clicked on"

I'll certainly agree with your "garbage" position, there are hundreds
of knee-jerk platitude postings here about what people should do when
they have SP2 problems that have nothing to do with details and almost
no evidence that doing them has accomplished the desired goal. I've
got to love the ones that think defragging enough might yield a miracle.

But, first off, someone who had SP2 problems shotgunned the long list
of newsgroups. McNulty, who responded to this didn't notice that
and responded thinking that this was about (pretty much) the only
things he has seen or dealt with, chipset problems in the hardware
group. But his claims were much more sweeping than that, until
both he and I resolved that he would scale back what he was claiming.
Which file? Where is the file? In what way does it crash?

I responded to his claim that everything had been explained by
asking for the explanation for the string of postings, and these
are still showing up, where users have "Windows Explorer" crash.
You can search the last 4000 postings in the windowsxp.general
newsgroup and look for Explorer to see how similar the descriptions
are. Most of these crash when a user clicks on an entry or double
clicks on an entry to open it. Scan all the postings an try to
see that they apparently fall into one or two, maybe three groups,
and the descriptions are actually pretty consistent.
No, there are a number of different issues, many of which can be
anticipated and/or ameliorated.

Again, he didn't think he was talking to anyone but the people dealing
with the odd problem that chipset drivers are being somehow lost with
SP2, apparently nobody knows why, he admits he doesn't know why, but
supposedly the guys in the hardware group think they have figured out
the other low level hardware problems.
Most go about things that SP2 is designed to block, i.e. apps that
rely on practices we now see as risky, and would prefer to block.
There's good documentation on these; lists of what software is
affected, and how, and how to fix.
But there are some specific issues that go about particular system
specifications or components, and that have effects beyond "everything
works except this particular application".
Then there are generic issues that can be expected to blow up any
attempt to install large wads of system code; flaky hardware, active
malware infections, and so on.
It's chicken and egg there. You want SP2 to keep these systems safer,
but many of these systems are already infested. So often the PCs that
would benefit the most from SP2, have adverse mileage because they are
riddled with malware before SP2 is in place and weren't cleaned first

Whether it is supposed to help protect against such crap coming in
is somewhat of a different issue from the omnipresent analysis that
has been posted litterally hundreds of times, you could post saying
"my computer is on fire" and you would get MVP's saying "scan for
viruses and spyware". That was what my sentence about "the chant"
was referring to. I have no objection to keeping crap off systems
but I think a disproportionate amount of the problems in the install
of SP2 have been "dealt with" by just pointing fingers at supposed
viruses and spyware, with extremely few reports that this was found
to then be the case.

And at the end of one of my postings, clearly anyone who was thinking
about doing an upgrade with naive users and who easily could have had
compromised systems would might have thought to do a scan for stuff
like this before doing the install. That might have easily saved 90%
of the people from having these problems, 100% if you believe some of
the more crazed folks. I cannot imagine why they didn't 1: virus scan,
2: spyware scan, 3: sfc run, 4: registry validate, 5: kill all processes
not absolutely essential to finish the install. All these are not rocket
science. I don't think that anywhere in the Microsoft "instructions"
did it say that you first needed to do a virus scan, fix any problems,
then do a spyware scan, fix any problems, ...
I'd really like more detail on that, as it's not a failure pattern
I've seen in posts here. How does YMMV if you:
- run Safe Mode?

done that
- suppress all in MSConfig?

checked that
- use Shell Extension Viewer to suppress CLSID integrations?
nope

- use something other than Explorer.exe as shell=?

nope, actually, trying to not mess too much with the system until
we can actually find the real root cause of this. I don't want to
use the new debugging model "reboot or reinstall and see if you
don't bother calling us back with your problem again."
- kill off Common Tasks and (if possible) Desktop.ini processing?

nope, but run very few things like that myself
- disable your LAN card in this profile to suppress networking?

none present
- suppress BHOs via IE's Tools, Options, Advanced?

I've done hijack this scans and have very few of those present
- NTFS vs. FATxx mileage differences?

Well, I'd really rather not risk blowing away my complete system
just to see whether the file system makes a difference unless you
can show me how much it is going to be worth to me
I'd be thinking of integrations into Explorer that are now blowing up
for one reason or another - not malware per se, but something added by
software or drivers, or broken by any number of things.

I understand and I'm hoping to help get this diagnosed and fixed.

I did escalate my Windows Explorer problem to Microsoft. They came
back and concluded "some file must be corrupted, repair windows back
to the original state and try reinstalling SP2 twice while you are
in Safe mode." Before I did that someone found that some folks see
Windows Explorer work when they create a new user and switch to that
user. I did that and in my case it works, also works in Safe mode.
I sent this additional info to MS, saying that this seemed to make
it less likely that "some file was corrupted" and asked for their
analysis. There has been no reponse from them for days now on this.
When Windows Explorer first starts, it may break on namespace
enumeration issues - things that aren't really drives but pretend to
be, that sort of thing.

I don't know how I could tell you whether that is the case or not.
Then, when it tries to list files in a folder, it may break on
persistent handlers, i.e. thumnailers or other code that runs when
files are enumerated, before any attempt is made to select or "open"
them. Antivirus scanners may impact at this point.
Then when you click or hover over something to select it, there may be
further persistent handler code run to populate the status bar info or
tooltip, e.g. things like .DOC authorship or .ZIP content file counts.

In my case Windows Explorer doesn't crash, it actually refuses to accept
any mouse click or keystroke and just "pongs" at me. But I'm in the
minority, most other folks report crashes on clicks or right clicks.
Then if you rt-click, anything that integrates into the contexct menu
gets a crack. Things that are listed "above the line" are usually
passive and don't run code unless clicked, whereas things that have
icons displayed, cascade out as additional menus, or are listed below
the top line in the context menu are running as code via a CLSID
Once you click a context menu entry, or press Enter on a file, or
double-click a file, you are not in Explorer anymore - extra code
beyond Explorer itself is more likely the problem.

I appreciate your description of all the details going on under the
surface. I'm hoping Microsoft is going to respond and somehow we
will get this problem resolved for everybody.

Thanks
 
Nathan McNulty said:
First off, where are these posts about SP2 that haven't been explained?

Well, for example in a post I made in this newsgroup on 8/28/04. After
waiting for 2 days without any responses at all, I finally posted again (in
that thread) to ask why no MVPs had anything to say about the questions I
posed. The only response I got was from one MVP, who stated that "I don't
see any questions listed in this post.". Now, that's a really big help!

Here is a copy of my post, which had a subject heading of "SP2 install delay
and other mysteries":

I just finished installing SP2 on my Gateway laptop, and ran into a few
interesting events.

Before doing the install, I was notified by the Automatic Update feature
that SP2 was available for install. I clicked on that icon and the screen
told me that I would be downloading a 260 MB file. Having heard that there
are smaller SP2 files, I cancelled out of that and clicked on the Windows
Update button, where I was told that the download of SP2 for my Home Edition
OS would be, if my memory is correct, 73 MB. So I opted to download and
install SP2 from the WU rather than from the Automatic Update feature. This
puzzles me. Why is the size of the SP2 file in Automatic Update different
from the size of the SP2 file in WU? (NOTE FROM OP - THAT WAS QUESTION
NUMBER 1)

Anyway, after initiating the SP2 download and install in WU, the download
went off without a hitch. However, the install procedure stalled for at
least 20 minutes (I use cable broadband) with only 3 tics on the progress
bar. I hit Ctrl-Alt-Del to see if there was a "Not Responding" message, but
there wasn't. So I "X"d out of that and waited another 15 minutes. Still no
more progress on the progress bar. Finally I hit "Cancel" to get out of the
install.

Then I clicked on WU again and re-started the SP2 install. I was pleased to
see that the system is smart enough to not do the download all over again.
Instead it began immediately to do the install, and this time there were no
delays and the install was successfully completed. I wonder why there was an
install delay during my first attempt. The download had already been
completed, so I don't see how an internet communication delay could have
been the culprit. (NOTE FROM OP - THAT WAS QUESTION NUMBER 2)

Any comments on these events are certainly welcome.
 
I like this sentence: "But his claims were much more sweeping than that,
until both he and I resolved that he would scale back what he was
claiming." I laugh... and agree :D

I would like to point out one of the many articles that are still being
written about SP2 (http://www.eweek.com/article2/0,1759,1642449,00.asp).
It has been stabbed at many times including by cquirke as to why these
programs inhibit the ability for SP2 to function correctly. It doesn't
really matter why it happens, just the fact that it does.

I agree that it sucks that software can sometimes install without a
hitch and can sometimes install with many errors even under the exact
same circumstances (identical hardware/software state). This is the
life of SP2. It works, it doesn't work, it works, it doesn't work.
Hopefully those who have problems will look at their computers are
figure out (with a little help) why Hotbar is not good for them, why
mass downloading unkown things off Kazaa is not good for them, why...
you get the picture.

As for this:
I appreciate your description of all the details going on under the
surface. I'm hoping Microsoft is going to respond and somehow we
will get this problem resolved for everybody.

Since these newsgroups are not monitored by Microsoft, though MS
employees may regularly volunteer their assistance, this will most
likely not happen. It would pretty much require a petition (and a large
one at that) in order to get anywhere.

I do challenege you though, everyone who is following this thread and
providing their input, to think about applying for the technical beta
programs when they become available if you have not already. These
testers are not just to find bugs, though that is a primary purpose, but
they also provide feedback and suggestions of which some of mine from
the past have actually found their way into the OS (with the voices of
many other testers as well).
 
See inline comments

----
Nathan McNulty

Well, for example in a post I made in this newsgroup on 8/28/04. After
waiting for 2 days without any responses at all, I finally posted again (in
that thread) to ask why no MVPs had anything to say about the questions I
posed. The only response I got was from one MVP, who stated that "I don't
see any questions listed in this post.". Now, that's a really big help!

Here is a copy of my post, which had a subject heading of "SP2 install delay
and other mysteries":

I just finished installing SP2 on my Gateway laptop, and ran into a few
interesting events.

Before doing the install, I was notified by the Automatic Update feature
that SP2 was available for install. I clicked on that icon and the screen
told me that I would be downloading a 260 MB file. Having heard that there
are smaller SP2 files, I cancelled out of that and clicked on the Windows
Update button, where I was told that the download of SP2 for my Home Edition
OS would be, if my memory is correct, 73 MB. So I opted to download and
install SP2 from the WU rather than from the Automatic Update feature. This
puzzles me. Why is the size of the SP2 file in Automatic Update different
from the size of the SP2 file in WU? (NOTE FROM OP - THAT WAS QUESTION
NUMBER 1)
The original fie size was 266 MB. If you have no patches on the system
at all, the file download should still not be that large. I would
venture to guess that Automatic Updates was simply specifying the
largest the file size would be. Microsoft is actually in the process of
modifying their update system as we speak.

Microsoft uses different databases for Windows Update, Automatic Update,
SUS, etc. This is why they are pushing for Windows Update Services
which all use one database. This will provide accurate results. For
more information on this, go to
http://www.microsoft.com/windowsserversystem/sus/default.mspx
Anyway, after initiating the SP2 download and install in WU, the download
went off without a hitch. However, the install procedure stalled for at
least 20 minutes (I use cable broadband) with only 3 tics on the progress
bar. I hit Ctrl-Alt-Del to see if there was a "Not Responding" message, but
there wasn't. So I "X"d out of that and waited another 15 minutes. Still no
more progress on the progress bar. Finally I hit "Cancel" to get out of the
install.

Then I clicked on WU again and re-started the SP2 install. I was pleased to
see that the system is smart enough to not do the download all over again.
Instead it began immediately to do the install, and this time there were no
delays and the install was successfully completed. I wonder why there was an
install delay during my first attempt. The download had already been
completed, so I don't see how an internet communication delay could have
been the culprit. (NOTE FROM OP - THAT WAS QUESTION NUMBER 2)
Generally when the download stalls during the downloading from WU, it is
because a system file (especially atapi.sys) was locked. I hate to say
it, but software can install differently each time you install it. I
have had completely screwy installs of XP before where notepad wouldn't
even open. Another format and install and it worked perfectly. Don't
know why, but something tripped up.

Here is where software and hardware meet. What happens if the
information stored in your memory was not retrievable and this locked up
Windows Update. There are many things involved with hardware that could
have caused this as well. It doesn't even have to mean that you have
bad hardware. I forget the specific statistic (but I'm sure someone
will correct me), but the average memory has at least 1 error every 20
days of running. Sometimes these errors do not affect anything, but
sometimes they do. This is why we have ECC memory on servers.
Any comments on these events are certainly welcome.
I apologize that no one answered your original post. Sometimes it is
that no one even read it, sometimes no one knew the answer, or they
started but never finished. I try my best to read every post from the
hardware newsgroup fully. I spend a couple hours in there every day.
There are days where I am gone, out of town, etc that prevent me from
seeing posts and then I try to catch back up. I have missed very few
posts in the last three months or so, so I am sorry that I did not reply
to yours.
 
Why not a test issued by the ISP (a standardized test but monitored by
the ISP) before allowing them online. After all, we don't just give
anyone a key to a car and let them drive. They have to learn how to
drive, pass a test, and do the work before they can drive.
 
The Microsoft Internet Connection Firewall (ICF) is enabled during the
installation of SP2.

This has become a godsend for the non-technical user, but for those with
home network and domain based enviornments, it has been an IT nightmare.

Don't forget to do you're homework and research any change made in a
production enviornment BEFORE making the change.

If you have a broadband router disasble the ICF if you are directly
connected to the internet leave it on.

If you run a network and need only a couple of services, go into the
advanced options for the ICF and select the services you want to allow and
any custom ports for any games or other special services.

Mike
 
cquirke (MVP Win9x) said:
On Sun, 29 Aug 2004 14:42:49 -0400, Barry Watzman
"Prescott" is the latest (3rd) internal core design of Intel Pentium 4
processors [the 1st was Willamette, the 2nd, still being sold, was
Northwood]. Prescott is relatively new, and is most easily evidenced by
an 800 MHz front side bus (but some 800 MHz FSB processors are
Northwoods).

Yes, the Prescott generation came out around June 2004.

Well, Mar 2004.
They double
the L1 data cache and L2 cache of Northwood, push the Celeron's base
speed up from 400MHz to 533MHz, and have a few other plusses (SIMD3
upcodes) and minuses (longer pipelines, which slow down the chip now,
but allow it to fare better at future higher clock speeds).


Or, if there's no BIOS update yet, uninstall SP2.


Well, apparently SP2 will run fine with NO microcode update and update.sys
renamed or using SP1's update.sys
 
See inline
The Microsoft Internet Connection Firewall (ICF) is enabled during the
installation of SP2.

This has become a godsend for the non-technical user, but for those with
home network and domain based enviornments, it has been an IT nightmare.

Don't forget to do you're homework and research any change made in a
production enviornment BEFORE making the change.

If you have a broadband router disasble the ICF if you are directly
connected to the internet leave it on.

This was the point of the firewall to work with broadband. Because
some third party firewalls won't with broadband.

If you run a network and need only a couple of services, go into the
advanced options for the ICF and select the services you want to allow and
any custom ports for any games or other special services.

Mike

:


Greg R







http://www.angelfire.com/in4/computertips/
 
Nathan McNulty said:
The problem doesn't point to the software, but more often the user and
junk they have allowed on their computer. Don't blame SP2 when you
can't take care of your computer.

I was with one user the other day, a retired lady barrister. She said
"Why are things so difficult? I only want to be able to send e-mails to
my friends in Australia. All these viruses and things are making me
think about going back to faxes!"

When you post things like "but more often the user and the junk they
have allowed on their computer" you need to remember that MS have
created a mass market piece of software and encouraged the non-technical
user to use it. Unfortunately it is so riddled with security holes,
that the average non-technical user is incapable of using it
effectively. Now SP2 comes along and more or less forces its way onto
their systems, systems that are likely to be full of viruses, and the
result is all too predictable.

It is not the user that "allows" junk on their computer - it is the
'operating system'.
 
Well, Mar 2004.

Which Prescott came out then? It may be that we in Rest Of World were
behind the pillar when the goodies got handed out, but we didn't see
Prescott until around June, starting with the P$ and then Celeron D a
week or few later. Are you thinking of P4 Extreme Edition? That's a
Northwood with heroic L2 cache and also prone to the same error.
Well, apparently SP2 will run fine with NO microcode update and update.sys
renamed or using SP1's update.sys

It seemed to run OK on my test system, yes, but I'd want to test
DirectX 9c games that use the new floating-point pixel shader feature
before being surer of saying so. I have a hunch that this new code
may be SIMD3-aware, thus Prescott-specific.

I'll soon be testing Prescoot Rev 0 + SP1a + Update.sys from SP2,
expecting that to break in the same way.

Then I'll BIOS-update the test PC and try Prescoot Rev 11 + SP1a +
Update.sys from SP2, expecting that to work.

That suppressing L1 and L2 makes full SP2 work fine, suggests it's not
purely a SIMD3 thing, but I suspect that new "is it a Prescott?" test
code within Update.sys (there to see whether DirectX 9c can use SIMD3)
is breaking on some ?timing assumptions that are sensitive to L1/L2.

I wonder if we will ever get the full scoop on this stuff :-p


------------ ----- ---- --- -- - - - -
I swear to god i must be the only true
 
cquirke wrote:
[/QUOTE]
Why not a test issued by the ISP (a standardized test but monitored by
the ISP) before allowing them online.

Ask yourself; how does one normally get businesses (such as ISPs) to
do things they don't want to do (because it costs money)?
- mandate they should do this, by law
- threaten to mandate by law, so they "self-regulate"
- set standards required for state tender business
- set standards required for international business
- set standards as part of a rating scale or product reviews
- direct government subsidy
- encourage new "value-add" products for consumers to pay
- build into insurance rating, thus subsidy from lower premiums
- accept as mitigator in anticipated litigation

Many of those require one to determine a set package of value that the
ISP is to add, and that is itself quite a minefield. Do you just
screen for MSware exploits, or do the same for Mac, Linux, etc.? Do
you factor in malware as well as exploitability? If you do factor in
malware, do you suspend service while the malware is in effect,
penalise by suspending service for a set period and then
re-test/re-punish, or drop the service alltogether?

The pain to ISPs is not only the overhead and resources required to
(say) screen user's PCs for exploitability, but also the loss of
customers to less picky ISPs when they redline what they find.


--------------- ----- ---- --- -- - - -
Tech Support: The guys who follow the
'Parade of New Products' with a shovel.
 
Let's get some detail there, else it's GIGO (Garbage In, Garbage Out),
[/QUOTE]
I'll certainly agree with your "garbage" position, there are hundreds
of knee-jerk platitude postings here about what people should do when
they have SP2 problems that have nothing to do with details and almost
no evidence that doing them has accomplished the desired goal.

Ai, be nice ;-)

Speaking for myself, I find ngs with > 500 posts a day to be more than
I can manage. Up to 20 posts a day, I can read every one and reply to
all that I can contribute to; I try to do that in the .za groups. Up
to 150 a day, I can read all subject headers and usually maintain
continuity. But xp.general's 1000-2000 a day is simply too much; I
follow up threads that my Free Agent is "watching", and I look for
"sp2" in the first few and the last few new subject headers, and the
rest I delete unseen. Brutal, but that's all I can do.

Because I've had detailed and early exposure to the Prescott issue, I
focus on that - but I don't just bang out that advice to all SP2
issues; I read to see whether it's relevant. It's relevant when the
particular failure pattern comes up (installs OK, locks every
subsequent boot with no error messages) or when it's a "should I
install SP2?" post ("do you have a recent system with Intel CPU?")

OTOH, there are lots of posters who base thier advice on a narrow
range of experience; either "works fine on the 12 PCs I installed it
on" or "I installed it on mine, and it was a disaster!". That becomes
a real problem when these posters arrogantly dismiss other poster's
mileage, e.g. "because it worked fine on my 12 PCs, there must be
something wrong with you or the way you installed SP2".

I don't reply to a lot of posts, but do spend quite a bit of time on
each. That's all very well, but may be less effective than someone
who helps more readers by replying to more posts, even if it is only
to paste in boilerplate text, or off-the-page URLs.
...love the ones that think defragging enough might yield a miracle.

Ah, that's one of my pet hates!

Please, dudes, raise your soldering iron in your right hand and
inscribe the following on the back of your left hand:

"Defrag is NOT a troubleshooting tool, and is dangerous
for unknown and troublesome PCs that may have at-risk
file systems, hard drives or other flaky hware e.g. RAM"

A fragged file system will slow things down to a mild to moderate
extent, but the only stability impact is that the critical window
period of file system operations may be prolonged.
You can search the last 4000 postings in the windowsxp.general
newsgroup and look for Explorer to see how similar the descs are.

Er, I can't do that easily, as it happens. Dunno how to do it in Free
Agent, not keen on wading around through a browser ;-)
Most of these crash when a user clicks on an entry or double
clicks on an entry to open it. Scan all the postings an try to
see that they apparently fall into one or two, maybe three groups,
and the descriptions are actually pretty consistent.

I'd love to harvest that info, but frankly it looks like too much
work. Normally, I'd suggest you post a new thread and I'd catch you
there, but this "xp.general" ng is so full I'd likely never see it -
and bulging a new sub-thread out of this one isn't great either.

If you like, you could email me directly, or maybe we could start a
new thread in a quieter ng such as...

microsoft.public.windowsxp.configuration_manage

....and if you tell me (by reply in this thread) what subject line to
look out for, I'll meet you there.
..."my computer is on fire" and you would get MVP's saying
"scan for viruses and spyware".

Ah, yes. Well, because malware is designed to defy troubleshooting
and mimic other things (including unrelated error messages), I see
great value in excluding this as part of the "prelim".

This is dogfood that I eat myself. Whenever an arbitrary PC comes in,
I don't even run Windows until I've done this...

http://cquirke.mvps.org/9x/bthink.htm

....i.e. checked for RAM and HD errors and formally scanned for
traditional malware. Only then do I go interactive, and usually the
first thing I do in Windows is to scan for and manage commercial
malware ("spyware" etc.). Then I know that the floorboards I'm
walking on are not going to collapse, and that the black boxes I reach
my hand into aren't going to contain snakes that bite me!
but I think a disproportionate amount of the problems in the install
of SP2 have been "dealt with" by just pointing fingers at supposed
viruses and spyware, with extremely few reports that this was found
to then be the case.

What I look for are stereotypical patterns of failure, and these
usually relate to specific issues, e.g.:
- specific STOP errors on start
- keeps rebooting on start
- locks up on black GUI on all boots (Ptrescott)
- runs OK normally but Safe Mode fails

OTOH, when the PC is infected with one/some of multiple malware, the
chances are it's also been subjected to bit-rot from AutoChk "fixing"
broken files are numerous bad exits. The chance of any stereotypical
pattern emerging out of this sprawl of permutations is slight.
And at the end of one of my postings, clearly anyone who was thinking
about doing an upgrade with naive users and who easily could have had
compromised systems would might have thought to do a scan for stuff
like this before doing the install. That might have easily saved 90%
of the people from having these problems

Yes, but it's not easy for MS to make specific recommendadtions here,
without Qs like; why recommend av product X and not Y? etc.

Also, MS may (should) find it embarrasing that NTFS lacks a proper
maintenance OS from which a formal virus scan can be done.
I cannot imagine why they didn't 1: virus scan,

Lack of mOS is a big obstacle for those who bought the hype and went
NTFS. You can hardly expect MS to say...
- use the WinPE boot CD (that we won't license to you)
- use Bart's PE
- use a Linux boot CDR
....and the usual advice to...
- use Safe Mode and Stinger (which scans for 60 out of thousands)
- visit an online av scanning site
....is pathetically weak for obvious reasons.
2: spyware scan,
3: sfc run

I wouldn't have thought of that ;-)
4: registry validate,

How does one do that, i.e. what built-in tools does XP offer?
5: kill all processes

Dangerous advice if online (need that firrewall!)
I don't think that anywhere in the Microsoft "instructions" did it
say that you first needed to do a virus scan, fix any problems,
then do a spyware scan, fix any problems, ...

Yes - there's a conflict of interest between wanting all PCs to patch
up, and warning on all the risks / precautions of patching up.

There's also an inherent tech-awareness mismatch between the low level
of tech insight that SP seeks to support, and the higher level of tech
insight needed to validate PC as safe for SP2 application.

This is the specific issue I want to flesh out off-thread...
done that
checked that

Yes, but what happened?

Thinking that this is likely to be where the body's buried. Most of
us have pet tools and things we like to install on all systems we
touch, and a conflict with one of these may make the problem look more
global than it is. For example, I always build with FATxx, not NTFS,
and use multiple volumes with XP installed in a path other than
C:\Windows; anything that fails on that will fail on "all" (my) PCs.
nope, actually, trying to not mess too much with the system until
we can actually find the real root cause of this.

I can relate to that. If you find that Safe and MSConfig suppression
both fail in the same way, then try Safe Mode Command Only, and see if
that works fine until you run Explorer. If it does, then you could
try normal mode with something other than Explorer as shell=
(something I've rarely done in XP but often in 9x), or just proceed
directly to Shell Extension Viewer - which shows shell integrations
that MSConfig doesn't, and which may be active in "safe".
nope, but run very few things like that myself

Other things may run from there, that's the point - not just malware,
but accepted software too.
none present

No LAN factor? OK. Most new mobos have LAN built-in, so it's still a
step worth trying, in case the PC goes off on a doomed "let's find an
IP address for this PC's (non-existant) LAN" quest.
I've done hijack this scans and have very few of those present

OK, but it still may be worth suppressing these via that setting.
Well, I'd really rather not risk blowing away my complete system
just to see whether the file system makes a difference unless you
can show me how much it is going to be worth to me

No, what I meant was; are your test systems NTFS or FATxx or both? My
own is FATxx, so if all of yours are NTFS and you wanted me to test on
FATxx, I could, as long as you can encapsulate the test suite.
I did escalate my Windows Explorer problem to Microsoft. They came
back and concluded "some file must be corrupted, repair windows back
to the original state and try reinstalling SP2 twice while you are
in Safe mode." Before I did that someone found that some folks see
Windows Explorer work when they create a new user and switch to that
user. I did that and in my case it works, also works in Safe mode.

Ahhh... OK, then what ye seek is mediated by per-account settings:
- HKCU
- HKCU overrides of HKCR (new XP feature)
- StartUp
- "other"
I sent this additional info to MS, saying that this seemed to make
it less likely that "some file was corrupted" and asked for their
analysis. There has been no reponse from them for days now on this.

OK. From what you've given me, I can't get very specific either.
I don't know how I could tell you whether that is the case or not.

Sure, it's not always easy to get tools to show you what you need to
know, and sometimes you need (hopefully free) 3rd-party tools such as
Shell Extension Viewer. It's a good start to know what you need to
find out, though, and that's the spirit of my advice.

Things that are relevant here:
- LAN drive mappings
- NTFS re-parse points (trickery such as "call C:\This\Path" "D:")
- other NTFS funnies (EFS, sparse files, compression, quota etc.)
- virtual CD, i.e. "copy CD to HD and treat as drive letter"
- 3rd-partyware that treats "private" files as drive letters
- 3rd-partyware that drop non-file, non-shortcut items on desktop
- anything odd that shows up outside drive letters in Explorer
- TweakUI settings to hide drives etc.
- cameraware that jumps in to manage USB plug-ins
- CDRW packet-writing sware that fiddles with CDRW drive

....as can viruses, of course.

Think any sort of indexers (including the Indexing Service), anything
that maintains thumbnail views, things that extract icons out of
graphic file content, etc.
In my case Windows Explorer doesn't crash, it actually refuses to accept
any mouse click or keystroke and just "pongs" at me.

Ah; use Alt-Tab to see if there's an "always in front" modal dialog
box that is not in front, thus presenting as a pseudo-crash.
most other folks report crashes on clicks or right clicks.

Clicks, and especially rt-clicks, smell like context menu integration.

Thar be many, many dragons; think WinZip, WinRAR, IOmegaware,
packet-writing CDRWware, av, all manner of fluff.

--------------- ----- ---- --- -- - - -
Error Messages Are Your Friends
 
Well, for example in a post I made in this newsgroup on 8/28/04.
Here is a copy of my post, "SP2 install delay and other mysteries":
I just finished installing SP2 on my Gateway laptop
Before doing the install, I was notified by the Automatic Update feature
that SP2 was available for install. I clicked on that icon and the screen
told me that I would be downloading a 260 MB file. Having heard that there
are smaller SP2 files, I cancelled out of that and clicked on the Windows
Update button, where I was told that the download of SP2 for my Home Edition
OS would be, if my memory is correct, 73 MB. So I opted to download and
install SP2 from the WU rather than from the Automatic Update feature. This
puzzles me. Why is the size of the SP2 file in Automatic Update different
from the size of the SP2 file in WU?

The "big" form that is described as "for networks" is complete, and is
your best (only sound?) bet if you want to dump it on CDR for re-use
on arbitrary PCs. I recommend this approach to those on broadband,
especially to those rather prone to "just re-install Windows" ;-)

For DUN users, I suggest giving up on WU entirely, and ordering SP2 on
CD-ROM (or looking out for genuine MS CD-ROMs at shops etc.). This
should offer the full "network" version.

In all cases, if your PC is a recent one that has an Intel processor,
then read http://cquirke.mvps.org/sp2intel.htm before you start. This
specific issue is significant, because if it applies, you may think it
is impssoble to get back into your PC.

The "small" forms delivered by WU are the subset of SP2 files required
to update your specific PC. Material that is relevant only to
subsystems you don't have installed may be left out, as will files
that duplicate patches you have already applied.
Anyway, after initiating the SP2 download and install in WU, the download
went off without a hitch. However, the install procedure stalled for at
least 20 minutes (I use cable broadband) with only 3 tics on the progress
bar. I hit Ctrl-Alt-Del to see if there was a "Not Responding" message, but
there wasn't. So I "X"d out of that and waited another 15 minutes. Still no
more progress on the progress bar. Finally I hit "Cancel" to get out of the
install.

During this time, was there any HD activity? If not, then presumably
it was navel-gazing in RAM (unlikely to take so long on a modern PC)
or waiting for some peripheral or network resource.
Then I clicked on WU again and re-started the SP2 install. I was pleased to
see that the system is smart enough to not do the download all over again.
Instead it began immediately to do the install, and this time there were no
delays and the install was successfully completed. I wonder why there was an
install delay during my first attempt.

I wonder, too. I might retry the test conditions with LAN card
"disabled in this profile" and peripherals unplugged.
The download had already been completed, so I don't see how an
internet communication delay could have been the culprit.

There may be post-install fiddling to be done, e.g. signature
verification and what-have-you. I've seen this when installing IE
upgrades that want to "call home"even when I'm installing from on
on-HD copy that was copied off CDR, with the read-only attributes
cleared. It may be a well-intentioned attempt to test whether the
code really is from MS, and is un-tampered with.
(NOTE FROM OP - THAT WAS QUESTION NUMBER 2)
Questions have question marks (?) ;-)

Jokes aside, an automated question-searcher may depend on ? to find
questions. Perhaps this approach is used by some advisors who still
try to find and respond to everyone's Qs, unlike cherry-picking me?

Should have found your question 1, though.


-------------- ---- --- -- - - - -
"I think it's time we took our
friendship to the next level"
'What, gender roles and abuse?'
 
Back
Top