[MDGx]

UPDATED · 12-27-2007

Please see the top of this topic for most recent updates.

Enjoy.

[dencorso]

A long time ago, Geoff Chappell, author of the classic book DOS Internals, reported (in June 07, 1999) a math bug in both Win 95 and Win98 FE versions of SYSMON.EXE (here: System Monitor Rounds to Thousands). Geoff also provided detailed directions for a corrective patch of these versions of SYSMON.EXE. And, of course, it also applies to the Win98 SE version of SYSMON.EXE, because it is identical to that of Win 98 FE, except for its creation date and time, both reporting to be v. 4.10.0.1998. But, in June 08, 2000, one year later, SYSMON.EXE v. 4.90.0.3000 was released in Win ME, and my analysis shows it has the same buggy code inside as that of the Win 98 version, showing that Microsoft disregarded completely Geoff's heads up. His patch for the Win 98 version solves the problem for the Win ME version too, but, of course, has to be applied at different offsets (30C4, 30D4 and 30DA, respectively). Since this issue remained unaddressed up to now, I here offer you

Download-link: both patched SYSMON.EXE for Win 98 (SE and FE) and for Win ME

Inside SYSMON.7z one finds SYSMON.WSE and SYSMON.WME, which are, respectively, the files for Win 98 (SE and FE) and for Win ME, patched according to Geoff's instructions. Of course, before using one of these files to replace that in the %windir% folder, it is necessary to rename it from .W?E to .EXE. For both files, I've increased the version number by 1, to reflect the fact that those are patched files. Users of 98SE2ME should use the Win ME file, of course.

As always, the standard disclaimer applies: they work great for me, but YMMV and I can guarantee nothing whatsoever about these patched files, and about the use you make of any of them. By deciding to use one of them you fully accept that anything you do is of YOUR SOLE RESPONSIBILITY... Moreover, modding files voids the EULA, of course. You have been warned.

This post has been edited by dencorso: 21 October 2010 - 08:20 PM

[RetroOS]

dencorso, on Jan 9 2008, 04:14 PM, said:

A long time ago...
...
Since this issue remained unaddressed up to now, I here offer you...
...

Nice!
If MDGx is the Updates Guy, then dencorso is the Mods Guy!
Way to go dencorso.
Man, I recon that Microsoft must be complaining through their clenched teeth: "Why won't that old Windows die!"
Why?
Because there are too many people who figure that Microsoft made a bad judgement call by hauling their branched NotTolerant code line into the mainstream! (Remember in the NT 3.x days, you would install a Novell server because most people figured that Windows was not a real server...)
...And too many people who don't want to add another mile to the rubbish dumps just to "upgrade" their OS...
Maybe I'm over simplifying, but surely fifty million users can't be wrong?
The NT platform should have been an option rather than the rule!
Sure NT technology has some advantages like more fluid multitasking, multiple processors, better resource management (I agree!), better security and less vulnerabilities (yeah right!), better file system (unless you try recoverying a crashed NTFS!), improved API, more compatibility (huh?), and so on. You get the idea.
I won't cover the advantages of Windows 9x since those reading this are using Windows 9x because of them!

My Windows 98SE box is more stable than most of the out-of-the-box XP and Vista computers I've used.
Most pre-loaded Windows Vista notebooks that I have set up have blue screened a short time after completing the Mini-Setup.
That's Toshiba, ASUS, Dell, and others...
You decide!

[blackwire]

dencorso,

I noticed some problems with the patched SYSMON.EXE file. It always shows my cpu load at 100% even though I'm not running anything. The original version of the file does not have this problem.

A better solution would be to use process explorer from sysinternals. It works with 98 and is a vast improvement over the built in tools in the OS.

http://technet.microsoft.com/en-us/sysinte...s/bb896653.aspx

This post has been edited by blackwire: 12 January 2008 - 09:33 AM

[dencorso]

blackwire, on Jan 12 2008, 12:29 PM, said:

dencorso,

I noticed some problems with the patched SYSMON.EXE file. It always shows my cpu load at 100% even though I'm not running anything. The original version of the file does not have this problem.

blackwire, thanks for the report. I failed, however, to reproduce the issue you reported, in all test scenarios I tried.
Can you provide me with more detail, please? And thanks for the pointer to Process Explorer. I do use WINTOP and the System Monitor, but other tools are always handy. Cheers!

This post has been edited by dencorso: 12 January 2008 - 09:08 PM

[blackwire]

I'm not sure what the exact cause is. I am running a pentium 3 based laptop with acpi enabled on win98se with Usp3.0a2. I am also running powerstrip. Powerstrip and ACPI are the only things I can think of that would cause the cpu load to act funny with the patched file.

[soporific]

MDGx, on May 20 2005, 11:04 AM, said:

USBFDD98:
http://www.mdgx.com/web.htm#9SU
* Unofficial Windows 98/98 SP1/98 SE USB Floppy Lockups DISKTSD.VXD 4.90.3001 Fix:
http://www.msfn.org/...ic=43605&st=352
Direct download [112 KB]:
http://www.mdgx.com/files/USBFDD98.EXE
This DISKTSD.VXD 4.90.3001 Fix installs also as part of Native USB (NUSB):
http://www.mdgx.com/web.htm#NUS

Isn't the file version 4.90.3000 ? The file inside the hotfix is 4.90.0.3000 -- is it supposed to be that, or is it supposed to be 4.90.3001?

TAAFTU.

(TAAFTU = Thanks as always for these updates)

[dencorso]

soporific, on Jan 14 2008, 12:21 AM, said:

Isn't the file version 4.90.3000 ? The file inside the hotfix is 4.90.0.3000 -- is it supposed to be that, or is it supposed to be 4.90.3001?

Yes, soporifc, you're right, at least in part:
If you check the file version with getver.exe, it will report 4.90.0.3000, whereas if you select the file in Windows Explorer, right-click on it and select Properties, you'll find 4.90.0.3001...
Of course, as it loads OK in Windows 98SE, it is the patched version, so Properties is displaying the correct value,
but, at hexadecimal file offset 4717, B8 must be changed to B9, to update also the hexadecimal file version, the one getver.exe reports...

This post has been edited by dencorso: 13 January 2008 - 11:59 PM

[Drugwash]

Or if you have the FileInfo plug-in installed in Total Commander, you can see everything you need to know (and more ):

[dencorso]

blackwire, on Jan 13 2008, 01:41 AM, said:

I'm not sure what the exact cause is. I am running a pentium 3 based laptop with acpi enabled on win98se with Usp3.0a2. I am also running powerstrip. Powerstrip and ACPI are the only things I can think of that would cause the cpu load to act funny with the patched file.

Hi, blackwire!

I've done some more testing (always with the folowing options selected in the view menu: Numeric Charts, Always on Top and Hide Title Bar), and was unable to reproduce your issue on 10 different machines, by now. Of course, my testing has one bias, because all the machines I've tested it in had AMD processors, so I'd warmly welcome test results from other users of Intel processors. There might be something there. I doubt the powerstrip can be related to your issue, as I would any other kind of overclocking (overclock issues usually crash the machine, instead of causing unusual behaviour). It is important to test it in Numeric Charts Mode, because the patch corrects a rounding issue, which is difficult to see in this mode, but probably impossible to percieve in graphic modes... I've also disassembled the relevant part of the code from both files in my release, and can confirm the patch is applied correctly, and is identical to the disassembly provided by Geoff Chappell in his original bug report. I fail to see how the code change Geoff's patch implements could be causing the issue you reported, though. So, it seems that what you are experiencing must be due to some quirk peculiar to your installation. Did you try the Win ME version in your system? Does it cause the same issue? A side note: the only occasion in which I've had CPU usages close to 100%, consistently, was when AVG was running in the background, although it didn't stay solid in 100%, but did so for a series of short periods. Best wishes.

[soporific]

dencorso, on Jan 14 2008, 04:54 PM, said:

soporific, on Jan 14 2008, 12:21 AM, said:

Isn't the file version 4.90.3000 ? The file inside the hotfix is 4.90.0.3000 -- is it supposed to be that, or is it supposed to be 4.90.3001?

Yes, soporifc, you're right, at least in part:
If you check the file version with getver.exe, it will report 4.90.0.3000, whereas if you select the file in Windows Explorer, right-click on it and select Properties, you'll find 4.90.0.3001...
Of course, as it loads OK in Windows 98SE, it is the patched version, so Properties is displaying the correct value,
but, at hexadecimal file offset 4717, B8 must be changed to B9, to update also the hexadecimal file version, the one getver.exe reports...

i never thought to check properties--- you're really good with these 'only partially right' situations, i think this is my third with you

This post has been edited by soporific: 16 January 2008 - 01:46 AM

[MDGx]

Hi guys,

I'm back.
I had an unfortunate computer crash a couple of weeks ago, so I had to rebuild my PC from scratch.

I have changed the hex address in DISKTSD.VXD as instructed by dencorso [thanks again ], please install again to see if now getver displays the version correctly.
I can't tell because I don't have access to Win98 SE anymore, I'm running XP exclusively for now, my new mobo may [or may not] support 98 SE, so until I get on a computer with 98 SE [or fix my mobo Intel P35 chipset INFs to work with 98 SE], I won't be able to test anything.

* Unofficial Windows 98/98 SP1/98 SE USB Floppy Lockups DISKTSD.VXD 4.90.3001 Fix:
http://www.msfn.org/...ic=43605&st=352
Direct download [112 KB, English]:
http://www.mdgx.com/files/USBFDD98.EXE
This DISKTSD.VXD 4.90.3001 Fix installs also as part of Native USB (NUSB):
http://www.mdgx.com/web.htm#NUS

HTH

[soporific]

MDGx, on Jan 17 2008, 08:33 AM, said:

Hi guys,

I'm back.
I had an unfortunate computer crash a couple of weeks ago, so I had to rebuild my PC from scratch.

I have changed the hex address in DISKTSD.VXD as instructed by dencorso [thanks again ], please install again to see if now getver displays the version correctly.
I can't tell because I don't have access to Win98 SE anymore, I'm running XP exclusively for now, my new mobo may [or may not] support 98 SE, so until I get on a computer with 98 SE [or fix my mobo Intel P35 chipset INFs to work with 98 SE], I won't be able to test anything.

MDGx welcome back ... and welcome to my world . Ever since i switched to a mobo that didn't support Win98se, i've been relying on Virtual PC 2007 to do all my AP testing. Its sufficient for doing all the code testing, and then the last thing i do is a real world test on my 3rd machine (an old Pentium 3 machine). Virtual Additions works very well which makes access to the test OS very easy. It will also save you lots of time, my dual CPU handles the VM very well and Win98 inside it runs very fast, and i can also do lots of other stuff at the same time.

[MDGx]

UPDATED 1-29-2008

Enjoy.

[RetroOS]

MDGx, on Jan 30 2008, 06:54 AM, said:

UPDATED · 1-29-2008

Please see the top of this topic for most recent updates.

Enjoy.

Oooh goody, a new Java 6 update!
I'd been checking regularly but still the Updates Guy bet me to it!

Well, 2008 has been busy with updates for 9x so far!
Windows 7? Bring it on! I know where I'm staying!

MDGx, still holding up Windows 9x in 2008! Man you're good!

[soporific]

RetroOS, on Jan 30 2008, 07:40 PM, said:

MDGx, on Jan 30 2008, 06:54 AM, said:

UPDATED · 1-29-2008

Please see the top of this topic for most recent updates.

Enjoy.

Oooh goody, a new Java 6 update!
I'd been checking regularly but still the Updates Guy bet me to it!

RetroOS, don't feel bad that i rebuffed all your advances as to including Java 6 update 4 in AP but as soon as MDGx lists it as a bonafide Win98SE update, wham, its in AP faster than you can say "what would we do without MDGx", its because i'm so used to MDGx being the final word on these things ... then dencorso turned up, and now we've got you as an authority on what works and what doesn't. Yayus! But who do i trust?

[Drugwash]

Has anyone tried to edit their MySpace account profile under Win98SE, having Java 6 installed and set as default? I've had trouble seeing some of the applets (color pickers, sliders, etc) as well as CPU usage to 100% (400MHz Pentium II, 256 MB PC133 RAM). I had to uninstall it. Could anyone else confirm (or not) this behavior?

[RetroOS]

Drugwash, on Jan 31 2008, 06:00 AM, said:

Has anyone tried to edit their MySpace account profile under Win98SE, having Java 6 installed and set as default?...

Did you try Update 4?

[Drugwash]

Nope, that was 6u3 that I had at the time. I'd rather wait for other opinions before I decide what to do, as I'm already after a bad (unrelated) experience yesterday that could've gone VERY wrong and I'm trying to avoid a heart attack.

[RetroOS]

Java 6 Update 4 is working without a hitch in 98SE.

Here is the warning text from the install (note the 'for best results we recommend' bit):

Warning: This is not a supported Operating System

Thank you for choosing the Java(TM ) Platform. You may continue with this
installation, but for best results we recommend running on one of the following
operating systems: Windows XP Professional (SP2), Windows XP Home (SP2),
Windows Server 2003, Windows 2000 Professional (SP4+), and Windows Vista.