sp193

Hey guys, just to let you all know on my findings. I DID complain that my Simplified Chinese IME on my USA Windows 98SE machine was not working properly (Some dialogs do not appear, only the "click" error sound is heard from the PC speaker). I found that by disabling KernelEX extensions for firefox.exe, I solved my problem. I use the last version of Firefox (2.0.20 I believe, but I know that it's the last version officially released), but I am not sure what is it's actual version number as I did not boot into Windows 98SE. I complained about this error last year, posted as the 2nd last post of the old KernelEx 0.36a thread.

Glad to hear that you've got no new issues with v0.0.8. Your problem is really quite persistent. You're using the USA version of Windows 98 right? That means that our software setups should not be too different (At the core level). Can you run the dependency walker on iexplore.exe and do a profile? Then maybe we can exchange the logs later on (Details via PMs). Also, please note down the missing exports (Names and/or Ordinal number if applicable) from SHELL32 that are linked to iexplore.exe. The dependency walker may be able to help us troubleshoot what is your installation's problem... You used IEradicator after uninstalling other software like KernelEx and SH95UPD right? Then did you install Internet Explorer 6.0 BEFORE installing SH95UPD and KernelEx again? If you did not do any of the above then you've quite likely changed nothing on your system. Optional: Did you try restoring your SHELL32.DLL, COMDLG32.DLL and EXPLORER.EXE after uninstalling SH95UPD (This would rule out the possibility of corrupted/contaminated system files)? There is, however, quite a big bug in SH95UPD. I don't know how long it's been there, but you cannot access ftp sites by typing ftp addresses in Internet Explorer 6.0. Doing so will cause a crash due to the missing export 256. I added that in a test, but that caused another crash somewhere else. Other than that, I don't see any other bugs. Hence, if you don't use internet explorer for ftp transfers, you won't have to worry.

It's not a HDD problem if Win98 can already boot from it. It seems more like it's just a driver problem. That's an alternative, but try to solve the driver issues in Windows 98 internally if possible as that's easier. We really need you to let us know that following: 1. Your computer's specifications. Check your device manager to get a list of installed devices/find out what hardware is installed in your computer from your OEM's site. 2. The model and manufacturer's name of the devices in question. Are all the other devices (Graphics card, sound card etc) installed and working properly? Are the correct(Not generic) drivers installed (e.g. Your graphics card should not be a "Standard PCI Graphics adapter (VGA)"? Note that for Windows 98 to detect an optical drive, the bus that the drive is connected to has to have it's driver installed. If your drive is connected to a SATA or SCSI bus, you need to find and install the appropriate drivers. If it's connected to a IDE bus than Windows 98 should have automatically installed the device. If all your other devices (e.g sound card) are installed properly, but only your network adapter is not installed, then you may have: a. a resource conflict b. The adapter does not have it's driver properly installed. Make sure that there are no device conflicts in your Device Manager (No yellow exclamation marks visible). Did that error suddenly appeared? Windows XP was there right? The Operating System didn't even start right? Then it may be: a. The HDD is not properly connected. b. The HDD is dead. Check if the old drive was properly recognized by the BIOS. Offtopic: I've got a question: What are the signs of a Northbridge/Southbridge failure? This PC which I am using is starting to have: 1. PCI adaptor detection issues (As if all sockets have become loose). The devices would not be properly recognized in Windows/causes crashes) 2. the USB systems (Both onboard and the EHCI adaptor) seems to have trouble supplying sufficient power to some USB devices (Some of my USB HDDs seems to malfunction if I don't plug in the USB Y-connector or use a external power supply) 3. I get "USB Bus Resets" when I burn CDs/DVDs. (From a USB drive to a USB Optical Writer) 4. I will sometimes get "The device cannot be read" or something similar when doing large file transfers to/from a USB device. 5. The Mouse PS2 port has gone dead (No fuse for the Mouse PS2, but only for the Keyboard which works). 6. COM1 (Onboard) is more or less dead too. Devices plugged there sometimes malfunction. COM2 is a connected to a riser. 7. AGP graphics card plugged into AGP slot seems to have some detection issues (Or had caused some driver crashes in Windows XP/98SE before, but no more). 8. The RAM modules (Both Kingston ValuRam 2x256MB) seem to have detection issues (The BIOS beep code for a RAM problem, or sometimes not all RAM is visible to the BIOS) but "touching" the modules seems to solve the problem temporarily. When I mean by it "seems like a socket is loose", I mean that by reinstalling that piece of hardware into it's socket, the problem is usually solved. I thought that it was a bad PCI/DIMM slot (Both PCI and RAM sockets), but that that's impossible. Almost all the slots have that problem, even if they were unused by me before. The problems I've listed above are intermittent, and do not frequently occur. But when they do... my PC is almost totally disabled. It won't even boot up (Either it won't clear the POST screen, or Windows would crash). The last one seems to be the biggest problem which is troubling me. The Mainboard is a Gigabyte GA-6OXM7E. It does seems to be on it's last legs. Examples of not all RAM being detected (Total RAM shown by BIOS should be 524288K): 133660K (Or somewhere around there) 262144K (But removing either RAM modules would cause the PC to give a "RAM error" beep code) 4096K (This was the worst)

You should run the Hardware Detection Wizard, and set it to detect non plug-and-play hardware. Install all new devices and reboot. Windows 98 should at that point at least install the generic "PCI bus" device driver. If it does not then you may need to download the chipset driver installer from your hardware magnufacturer's web site. You may need to reboot a few times to get Windows 98 to detect and install all PnP hardware later on. BTW: Well, you should wipe that harddisk clean of all data, and reinstall Windows on it. I'm not saying that you can't just do a HDD transplant, but this isn't really clean. There may be even complications later on. But if this does not solve your problem, could you please post here your computer's full specifications?

SH95UPD v0.0.8 is now released. As it has major core changes, it IS an "Ultra beta" - a much bigger beta than v0.0.7 with it was initially released. Changes for SH95UPD v0.0.8 24/12/2009 -Added the following functions: -IsUserAdmin (Export ordinal 680) -SHFlushSFCache (Export ordinal 526, alias of the SHFlushSFCache that was already present in older SH95UPD releases) -GetRealDriveType (Export ordinal 524) -SHBindToParent -Fully implemented SHGetSettings -Seperated the uninstaller from the main program. -The uninstaller is now the resident manager of SH95UPD, which has a smaller footprint. This may seem unimportant, but now the full uncompressed SH95UPD is already 366KB in size (Compared to the uninstaller’s 144KB uncompressed). -Added the ability for SH95UPD to alias a function that was only exported by ordinal. -Added the ability for SH95UPD to export a function by ordinal. IMPORTANT! SH95UPD v0.0.8 has some extreme core changes done to it (View the changelog for details). Hence you can consider it to be a “Ultra Beta”, a much bigger beta than v0.0.7 when it was initially released. Frankly, I have not really heavily tested this, but I have yet to encounter any flaws. Please report any _new_ crashes you may encounter here. Please refer to the first post of this thread for downloads. A Merry Christmas and a Happy New Year to the MSN community!

Thanks guys, but I won't be able to try that any time soon. My HDD case got wrecked (USB port got ripped off, and I can't solder it back on), and I won't be getting a replacement as that HDD is just too old, too small and I don't have $$$. Not to mention that it may already have suffered damage from my PC's faulty USB 2.0 hardware (It just seems to be the cause of all the damaged USB enclosure electronics of other cases) I DID boot Win9x (Windows 95, 95A, 95B, 96(Nashville) and 98SE) from a USB device before, even from PCs that do not support booting from an external USB device. I can't clearly remember, but I think that that error was so bad that it extended into safe mode (Meaning that I can't even access safe mode without hitting that error). But I CAN boot into MS-DOS mode, just not any mode that loads win.com (Running "win /D:M" would also cause that error to appear). This was what I meant by 3GB can kill this installation, not that it can kill any proper Win98 installation with the right adjustments made. At any rate, I think that I'll just find another device (Somehow), and install Win98 again. Then I'll limit the memory, or patch it. BTW: Won't just limiting the size of the swap file also fix this? Somehow I got this impression that I can from a Microsoft KB. Upcoming SH95UPD v0.0.8 features: -Some core modifications to make modifications easier to make -Separate uninstaller(?) -Add export 680 (IsUserAdmin). It was actually already in Win98 as export 680.

Posted a link to LitePC's site where IEradicator 2001 is hosted at on the 1st post. You can't hack the version number, as it's not a string, but an array of variables hard coded into SHELL32 after SH95UPD is installed (Unless you use a disassembler to check where it's offset is). But I've more or less comfirmed this theoritically: The version number should NOT confuse any program anymore (i think that the one in v0.0.7 did). If it does... then some other function added by SH95UPD is broken, or some incompatible function in the Windows 95 SHELL32.dll was called (e.g. Export ordinal 194). All modifications that uses/modifies the Windows SHELL32.dll must be installed on top of SH95UPD. Failure to do so may result in some strange, unwanted behaviour. This applies even when upgrading SH95UPD to a newer version. The bottom line: SH95UPD must always be installed below all modifications that are linked/dependent on SHELL32.dll. This means that you have to uninstall all modifications installed on top of SH95UPD before uninstalling/upgrading SH95UPD. example: If I want to upgrade SH95UPD v0.0.6 to 0.0.7A but I have KernelEx installed over it: I have to: 1. Uninstall KernelEx (And any other modifications that are linked to SHELL32.dll) and reboot. 2. Uinstall SH95UPD and reboot. 3. Install the new SH95UPD and reboot. 4. Re-install KernelEx and other modifications you wish to have. I know that this is just a blind guess... but I just want to be sure that you really did NOT break your system by upgrading SH95UPD (You mentioned that you found a old version there before upgrading)... and that there really is a major issue caused by SH95UPD. Tips for those who experience crashed with certain programs after installing KernelEx: Make sure that the "Disable KernelEx Extensions" is checked for that program in question. This applies exclusively to programs that experience crashes that are related with UNICOWS.dll. Good to know that my software has helped you. Thanks for your feedback. I use the hacks listed at his site. I didn't download his pre-hacked version... but modified EXPLORER myself. Interesting... I thought that the reboot routine was fool-proof. However, since it's a "WinOldApp" error... I got a feeling that the SHExitWindowsEx function introduced by SH95UPD is responsible. Could you please provide a screenshot, or post here the entire error message contents? Somehow I think that a "WinOldApp" error actually means that the MS-DOS console subsystem in your system had encountered an error... but the SHELL32 functions do not use a MS-DOS console window... Hence I somehow think that it was 98SE2ME's setup that had a problem. Someone correct me if my hypothesis here is incorrect... Offtopic: Thanks to all who were concerned about my exams. I did my best, and the exams were so-so. Right now I can have fun while waiting for the release of my results next year... However, now I have to live with the "normal" loading speed that comes with Windows NT 6.1..... I'm going to miss my Win98 installation. (More on what I mean here is on the 1st post). I found a old Windows 98SE (Japan locale) installation on a old laptop's HDD, but I can't boot it from my laptop's USB connection as this laptop has WAY too much RAM for Windows 98 to handle. I tried adjusting the page file sizes...but the notorious "Out of memory" error message that occurs at startup keeps appearing. Guess that 3GB of RAM can really doom a Windows 98SE installation.

What is your mainboard model? I am aware of a few models of mainboards that *cannot* boot Windows 98 with a Geforce card installed. I had a Biostar M6TBA mainboard(With a very slow, old PII 400MHZ CPU). I tried 3 GeForce cards on it. Windows always crashes/has issues after the Nvidia driver is loaded. 1. Geforce 2 MX-400 (Second hand) Windows 98 started, but would not progress further after the Windows 98 startup screen clears (Before the desktop appears). 2. Geforce 2 GTS (Sparkle SP-6600, second hand like above) Windows 98 would also fail to start like above, but would sometimes start, abeit with only 640x480x16 support. This was not really a fair test as this card had a problem (It overheats). I could not detemine if it was over overclocked by it's previous owner, or if it was physically damaged. I could not flash it's BIOS either as Windows 98 would not boot up with the Nvidia driver loaded. 3. Geforce FX 5200 Windows 98 did start, but some games that use HW T&L had terrible graphics (Attributed by the mainboard's AGP 1.0 slot?). I think that it did have occasional problems starting up too. The cause of the failures was that my Mainboard had a problem managing IRQs (A BIOS fix was availiable for the M6TBA version 1.2, but mine was a 1.0). The Geforce often conflicted with "PCI IRQ steering" or some other system device. Pity that I did not do boot logging when I still had that system working, or I could have seen whether your problem is similar. In a local magazine, a user with a similar mainboard(Biostar M6TZA) with a Geforce card also had a similar problem. In short, you may need to try to update your mainboard's BIOS if a IRQ conflict exists.

I can't really remember too clearly, but I think that I DID try the Winternals NTFS for Windows 98, but I don't think that it was V2.... or maybe it was. I remember that it did need some real files from a legit Windows XP installation, but seemed to be quite stable. Although they said that it only has "Read-only" support, I think that it seemed that I could also write files to the NTFS drive (But I did not check whether the files were really written to the drive). Also, I don't think that there was a problem with the recycle bin. I tried it last year, and many details are quite hazy. So if anyone sees any incorrect information posted here by me, please correct me. Yup, and if that was ever real, Windows 95/98 would not have become popular in it's year (Who buys an OS that causes you to lose data!?).

Strange... you said that you tried to launch IE6 throgh a shortcut and directly.... but nothing happens? I do that all the time, and it works perfectly. After my "final" 2 papers (Tomorrow, there is actually one more but it's only MCQs), maybe I'll try to re-enact this issue which you've encountered. I'll try to purge IE6 (with IEradicator 2001A), then reinstall it on top of a fully running system(WIth KEX and SH95UPD). I'll let you all know what happens. About the version number returned by SHELL32: It returns it's version as 4.72.3612, which is Windows 98SE's SHELL32.dll version, so it should actually not really cause any problem with Win9x (And it didn't for me...yet).

This means that "SH95UPD.dat" in your "X:\WINDOWS\SYSTEM" is missing/damaged. SH95UPD looks at that file and determines whether it's installed. Maybe I should implement a permanent version information function in SHELL32.dll like KUP did for KERNEL32.dll to prevent your issue from occurring again. I am using the same version. This is adequate evidence that SH95UPD may be the root cause. So that means that there IS still a major bug in SH95UPD. However, I forgot to ask you earlier: how did you launch your IE6? Is it: 1. Through the IE6 icon on your desktop (Actually a registry entry which was installed by the IE6 setup program)? 2. Through a shortcut on your desktop? 3. Through a Quicklaunch bar command (Leftover from the old Windows 98 shell)? 4. Through a shortcut from your Startmenu? 5. Launched directly? If it's through an icon that was placed by IE6's install, then yea, there may be a bug (I can't verify that as the IE6 setup never creates that icon for me anymore after I deleted it, but I remember that it did work when I was developing older versions of SH95UPD). I launched my IE6 through a shortcut from my Startmenu. Somehow, it seemed like after I added the ability for SHELL32 to return it's version, IE6's reaction to the version number actually varied. When I got SHELL32 to return it's version as 5.10.2800 (Windows NT 5.1), all the icons in IE6 disappeared. When I set it to around 4.00, IE6 would display the IE startup logo (Windows 95 behaviour). If I set it to 4.72.3612, then I got Windows 98 behaviour. XD Maybe SH95UPD isn't really ready to allow SHELL32.dll to return it's version as a v4.72 compatible... Offtopic: I am now upgrading my shell files to: EXPLORER.exe -> 4.00.951 SHELL32.dll -> 4.00.1112 (No change) COMDLG32.DLL -> 4.00.951 EDIT - I've finally uploaded a link on my Linkbucks blog. Check the first post of this thread for links

Well, I thought that it'll be interesting for some of you to know: I DID try to replace MS-DOS 7.10 with MS-DOS 7.00 before(I think about 4-5 years back). Use "sys c:" from an Windows 95B/C/D/E(4.00.950B/C/D/E) startup diskette. Don't use 4.00.950/4.00.950a/4.00.950A as they do not support FAT32. Windows 98 still booted properly, but at startup there was an ominous error message that said something like "Invalid version: 4.00.950"... I can't remember if the Windows 98SE MS-DOS tools still worked though. I did that a long time ago as I felt (At that time) that the Windows 95 startup logo looked better than the Windows 98 startup logo.

SH95UPD v0.0.7A would only tamper with SHELL32.dll, and no other file. Only those extremely old versions of SH95UPD (Like v0.0.1) would also tamper with KERNEL32.dll. Your COMDLG32.dll is newer than mine. I never knew that there was a v4.00.951 (I want one! XD). However, I have EXPLORER.exe 4.00.951 and SHELL32.dll 4.00.1112. I won't rule out that there is a damaged/incorrect registry value issue here though. I know what you mean. Maybe this problem isn't really caused by KEX nor SH95UPD. What other modifications did you install? I think that I also encountered something similar before, when Scandisk trashed all the long filenames on my PC(Including IE6's). The repair won't complete, even with SH95UPD v0.0.6 and KEX uninstalled(Not a system file problem, but just fails with no error I think...). So I had to format the entire installation...but I think that was more severe than your case. However, I am not sure whether SH95UPD or KEX is really to blame for this as I only install/uninstall/reinstall IE6 without any modifications installed. Can any other SH95UPD user confirm if this IS a real problem? Offtopic: Maybe I should post links for SH95UPD to upgrade all their Win95 files... (Newer is better in this case) For now, please search MDGx's site for updates.

Yea, it's a pity. However, you also need good software(That's properly written), and not only good hardware. Hmm, not missing files, but I think it was files that you had to copy to your C:\ after applying the "MS-DOS mode hack". I think that it was Windows ME reading the AUTOEXEC.bat and CONFIG.sys and copies them to (Or was it a file?) the system registry, and deleted the originals. Quite likely you'll succeed, provided that MS had a MS-DOS mode in pre-release versions of Windows ME. However, I have to disagree that the worst thing that could happen is that it won't work. The worst thing is that it will still work, but causes great instability and/or mass data corruption. Doesn't seem like, seems to be roughly the same. Blame the program's programmers if you actually experience real bad memory leaks (That are not the "natural memory leaks" in Windows but caused by badly written programs) . I tried removing IE, and using the Windows 95 shell. This also disabled System restore and perhaps PcHealth too, but ME DID become a lot faster and a LOT more stable. I did that about 4-5 years ago, way before I started programming SH95UPD. It won't run, plus almost all of it's MS-DOS tools won't work(As they are tied to MS-DOS 8.00). However, there is a hack out there that "enables" Windows ME's MS-DOS mode(This was discussed in a thread here, but I can't remember where, nor the hack's name), but I think that it isn't as good as the real option provided by Windows 95/98. It isn't as good as there isn't a solid feel to it, plus my MS-DOS games won't work on Windows ME's MS-DOS(After applying the hack), but works properly on Windows 98's MS-DOS 7.10. Well, after tweaking with Windows ME in the past, I do think that it isn't so bad after all. It actually has better startup speed than Windows 98, does not use the MS-DOS scandisk to scan your HDD after every bad shutdown (Windows 98's MS-DOS scandisk corrupted ALL my long filenames). Stability can also become "rock solid" if ME was tweaked (Maybe even on par with my modded Windows 98 SE system... but I did not keep that ME installation long enough to see if my modding actually had any long-term side effects). Yes, I'm more than willing to share with others what I gained from my old research... and that also makes me feel better after knowing that all the hours I "wasted" are still useful, and can help others.

Yes, that's quite probable cause, since SHELL32 would now report itself not as version 4.0.x but 4.72.3612. I suppose that since 98lite/ROM/ROM2 users have the ability to install Uberskin, the programmers should detect and patch their program pior to installation. Otherwise I would have to rewrite the SHELL32 functions that cause that issue... which may break EXPLORER.exe. That's NOT by design, and is certainly a problem! IE6.0 SP1 is working properly for me, maybe you should first try to reinstall it(Run IEridicator 2001A, then reinstall IE6.0)? I've also install KEX 0.40 on my system, still no problems. I am using a SHELL32.DLL version 4.0.1112, EXPLORER.exe 4.0.950 (But I also tried 4.0.951) and COMDLG32.DLL version 4.0.950. IE 6.0 is still working 100%. I formatted my system last year, during the development of SH95UPD, and I did not have your issue; So I suppose that this incompatibility is not widespread. Thanks for letting me know; It may actually be a highlight of a dawn of a new big error... XD Yea, I also noticed that certain programs would crash (Illegal operation caused in UNICOWS.dll; The program I tested was FFDSHOW) if I don't disable KEX extensions. Seems like the only proper way is that I somehow "integrate" KEX with SH95UPD properly, so that there won't be such a resource problem again... currently the only solution is that I don't ever link SHELL32.dll with UNICOWS.dll, or risk serious crashes with KEX is installed, hence all functions have to be UNICOWS-independent. Yes there is. Originally, SH95UPD v0.0.7 was intended to have a much more robust unicode system(From KEX0.36a), but that conflicted badly with KEX. It seemed to be because UNICOWS was already loaded by KEX. (Hence all crashes would disappear with KEX extensions were disabled) So I redesigned SH95UPD v0.0.7(The original and rev. A) pior to it's release to greatly reduce that side-effect, although I was cetain that I had totally resolved that issue beforehand... now I think that I'm wrong. Thanks! BTW: Geocities SG would be going offline from 2009/010/26 onwards, hence my files would be unavailiable for a while. When my exams are over(In the 1st week of November), I'll upload them to my Linkbucks blog.

You don't find any problems with the IAA? IAA actually affected my Benq CD-RW and DVD drives, so I had to dump it. I use a MOBO with an Intel i815 chipset. I also agree. Also, it's quite pointless to continue if you BIOS does not properly support 48-bit LBA as your system is on grounds on mass data corruption(Beyond the 137GB range). So in other words: If your BIOS SUPPORTS 48-bit LBA properly, then the ESDI_506.pdr can actually help prevent data corruption by Windows (Excluding it's disk management tools like Scandisk and Defrag which don't work well on my setup, since they've corrupted my filenames from time to time)? I used to use IAA, but now can't as it kills my CD-RW/DVD-ROM drive support, so I hoped that the 48-bit ESDI_506.pdr patch was a equal substitute(Excluding the "performance gains" that don't exist on my badly fragmented, un-defragmentable HDD). Luckly, my MOBO is able to handle a 200GB ATA-133 HDD with no problems(No detection problems in the BIOS), even if it was made in 2000/2001.

Wow...that problem shown in the Screenshot you posted looks bad... After you uninstall v0.0.7A, the taskbar works properly? You didn't install\uninstall anything else? Hmm... if you didn't alter your Windows installation other that uninstalling SH95UPD v0.0.7A then I can conclude that some function that I implemented in v0.0.7A isn't working properly. If that was really the case, then I would suspect the dummy SHGetSettings() function. It supposing returns the current settings of the Windows shell, but I didn't fully program it yet, and it returns dummy settings (That are still valid, but may be not the same as your current setup). These setting include whether hidden files are shown, and whether to display a confirmation dialog on deleting files from the recycle bin etc. About the links problem... I'll see if my Microsoft Word 97 has any problems. It also uses Hyperlinks and e-mail addresses. Windows 95 had this ability, but certain extra components had to be installed; For example: Internet Explorer 4.x/5.x. The Pre-OSR2.5(Win95C) Windows 95 components were largely unaware of Internet Explorer's existance, or it's shared functions. Hence it's SHELL32.dll did not contain any internet access capabilities. Agreed. In fact, I find it irritating to find any part of my Windows 98/95 hybrid not working properly.

Congrats on fixing your SHELL32.dll. I forgot to mention earlier, but the "incompatible version" you had might not have been SH95UPD, but a version of KernelEx\KUP. SH95UPD also reports them as "Incompatible versions" if detected(And the backup files of KEX are *.W01). I've never encountered this problem personally(The programs I use do not use the mailto: and http: handler), but in theory this might happen as the Windows 95 SHELL32.dll is not very "integrated" with the rest of the Windows OS(And lacks certain functions) compared to the Windows 98 SHELL32. Your solution is probably the best option, as the Windows 95 shell is missing the necessary code (And is not aware of URL.dll or any other IE related files). However, to help solve this, could you please be more specific to list programs that malfunction because of this? Other missing functions include the SHCreatePropSheetExtArrayEx, which is implemented in the Windows 98 SHELL32.dll as ordinal 194 (Hence the Desktop control panel crashes, as Ordinal 194 is RealShellExecuteExA in the Windows 95 SHELL32) You mean that the Taskbar actually goes off-screen or out of it's normal positions? I'll look into that... might have been caused by a "broken" function... if any still exist. Uberskin runs on a normal Windows 98SE installation right? There are still unimplemented functions in SH95UPD v0.0.7A, so this might be a side effect.

Well, if you installed SH95UPD v0.0.3 properly there then there should be an entry in your "Add/Remove programs" control panel called "SH95UPD v0.0.3". However, if it's missing then you could try the following in pure MS-DOS mode(Not the MS-DOS box in windows, but you MUST reboot into Windows 9x's MS-DOS): 1. Run the batch file SH95RECO.bat in your WINDOWS\COMMAND folder. (e.g. C:\WINDOWS\COMMAND\SH95RECO.bat) 2. SH95UPD backs up your system files as *.S95, so rename SHELL32.s95 as SHELL32.dll and you're fixed. 3. If that fails, then you have to restore a copy of SHELL32.DLL from your Windows 95 setup disc\diskettes. Also, if your copy of SH95UPD was old to have been the version that modified KERNEL32.DLL, then you should also restore that. The last version I think that still did that was v0.0.2 so I think that you're safe. If you use Windows 98 then you can use SFC for that; If you are using Windows 95 then you need to extract the files from your \Win95 folder on your setup disc/diskettes yourself. 4. Another way(But is highly unrecommended) is to run SH95UPD.exe with the --debug-reinstall parameter. I use this during development. But please, use this only when you really can't fix your installation with the above methods as this is risky! However, this would preserve all previous modifications in your SHELL32.DLL. It's almost the same as method 2, but "hot-patches". All the best! BTW: My Geocities page would be going off-line on the 26th October 2009 - the first day of my exams. Therefore I think that I would upload everything onto my Linkbucks blog... until I find another suitable host.

It also lags on my system, almost to the point of a system crash (This is on Win98SE, and on both Firefox 1.5/2.0 and IE6). It seems to be a resource leak by Flash 9 in Windows 9x according to some other thread here. However, this lag is less severe on other operating systems like Windows XP/Debian 5.0. Hence this is not counted as a performance problem as the official Adobe Flash player 9.0 seems to be unstable on ALL Win9x versions (Despite the fact that it's meant to run on it). It also lags badly on my system, even if I do not use Windows ME now (I am using Windows 98SE). I am not very familiar with those games, but if has an option to set the graphics renderer, set it to Direct3D (aka D3D/Directdraw) or OpenGL, but NOT "software". Also, try reducing your resolution in the game, and turn off Antialiasing, if it's enabled. Also, try to force the game to use Pixel Shader 1.1/1.4 instead of 2.0 if possible. This should improve performance with GeForce 5 FX cards. I once thought that the 82.69 drivers would offer better performance and stability (Windows 98 sometimes crashes with it comes out of standby) with my GeForce FX5200, but the people at that thread said that there were no improvements for my model in that driver set, and that I was better off using the official 81.98 driver. And yes, the GeForce 5 FX series is very weak in Pixel Shader operations (It's performance is very poor). Over-clocking won't fix this because it's a poor hardware design problem. The 77.72 driver might offer better performance? Hmm... I may want to try that, but does it have more bugs than 81.98? I already have enough troubles with my buggy 81.98 driver, and introducing some more is just uncalled for. BTW What is your PC's specifications? Performance is also hindered by old hardware and improperly installed or generic drivers. For example, installing the Intel Chipset drivers if you have an Intel based chipset really improved performance for me. Also, when I looked at the CPU usage of certain programs on WinXP, the CPU is always at 100%. So this shows that the "high CPU" usage you noticed may be because of old hardware, not unoptimized drivers that were installed (Windows 9x, including ME, need to have proper drivers from their manufacturers installed, or the "generic" drivers from Microsoft would slow the entire system down(Especially graphics card and Chipset drivers). Did you do tests on the same PC, but on different OSes? My PC's specifications: CPU: 1.0 GHz Intel Pentium 3 HDD: Hitachi Deskstar 200GB ATA-133 (Operating at ATA-100 transfer mode) Mainboard: Gigabyte GA-6OXM7E (Version 1.0) RAM: 2x256MB PC-133 Kingston Valueram sticks (Already the maximum for this mainboard, which is low thanks to the Intel 815 chipset's design flaw) Graphics Card: Nvidia Geforce FX5200 OS: MS-DOS 6.22, Windows 98SE(With Windows 95 shell), Windows XP Home and Debian 5.0 "Lenny" CD-ROM: SONY CD-ROM CDU571-Q (Apparently a 16x CD-ROM drive)

Windows Media Player 6.4 CAN play DVDs, but you need to change some registry keys. I can't remember how to do that, but I know that it CAN be done. Another way is to install FFDSHOW(Last stable version for Windows 9x if you do not have KernelEx installed), and enable "DVD decoding" under MPEG2 decoding under the FFDSHOW video configuration. Try opening all the files one by one in WMP. One of them is the DVD's main video file. I played DVDs on my Windows 98 machine running WMP 6.4 with no problems. That DVD has multi-filesystems support. Somehow it feels like that DVD has a problem... did you copy it from another disc? My guess is that you should open "VTS_01_1.VOB" in WMP, as it should be the MPEG video file on that DVD being the biggest. BTW: Make sure the "File type" is set to "all files" NOT "Supported video formats" or you WILL NOT see any of those video files in WMP. You do NOT need to associate any of the file types to play DVDs.

Thought that I would give a little surprise before my "all important examination". I present to the MSFN community SH95UPD v0.0.7. It is a major rewrite of SH95UPD v0.0.6, and now uses the KernelEx 0.3.6a architecture instead of the old KernelEx 0.3.2a architecture. It has many of the missing exports implemented and fixed. However, that makes it a genuine "prototype" as it may have introduced many new bugs. I didn't extensively test it though, but it hasn't caused any malfunctions on my system yet. The sad thing is that Firefox 3.0 still doesn't run on my system. Tested with KernelEx 0.4.0 Final installed on top. Post here the issues you have encountered in v0.0.7, but were not present in v0.0.6. Also, post any new issues to SH95UPD here, but please mark them as being "fresh". Visit the project site for a more complete update list. Edit 2009/08/27: I found a major issue with SH95UPD v0.0.7. I can't save web pages in SH95UPD due to a faulty SHGetSpecialFolderPathA implementation... Therefore I have released SH95UPD v0.0.7A that should resolve the issue...

How do you know that it wasn't Windows ME at fault? Windows ME (Out of the box) does cause programs to become unstable often, leading to crashes. I used it before, and programs that worked fine on Windows 98 do crash... and lag. If the program does not support Windows 9x natively, then you are running them at your own risk. Potential side-effects of doing so include instability, crashes and other undesirable behavior. As stated above, this may be one of those side-effects, or may just be caused by Windows ME. Program performance on Windows 9x in general is almost on par with performance on Windows XP, if not faster (But really depends on the program you're running, and the hardware you're running them on. IOW Program performance would not really be faster if you computer is very "modern"). But this is only possible when Windows 9x IS supported by the program. It really depends on why the program crashed. Do try KernelEx though... but it may not solve your problem.

It won't be too much. In fact, DOS/Windows 3.1 won't use most of it. Most machines in it's era only had ~8MB RAM. Windows 3.1 does not require to have Virtual memory, and doesn't have the problem Windows 9x has. I don't think DOS can access beyond the 8GB barrier(I had a partition on the ~60GB area of my HDD and MS-DOS 7.00 didn't like it). What version of DOS are you using? MS-DOS 6.22 only had FAT16 support (Partitions must be<2GB, and are FAT16 formatted). MS-DOS 7.10 doesn't work well with Win3x without patching. Also, DOS would INSIST to install to C:\DOS.... so beware how your disk is partitioned. It would ignore other partitions it can't read though... The PC-Speaker driver would allow you to have sound in Windows and Sound Recorder, but many programs would not have any sound at all. When you install DOS, it would overwrite your HDD MBR, so you need to know how to restore and configure WinNT 5.x (aka 2000, XP etc) bootloader. Yo may also need to select your PC as a "Standard PC with APM" if your loading @ the Windows screen takes a long time, or vice-versa. I had mixed responses to this selection on different machines. Wow... Windows 3.1 won't fit there, unless you cut out almost everything..then it's practically almost useless... If you have a old <8GB HDD then try to fit it there.

Corrent me if I am wrong, but didn't Windows 95C(OSR2.5) have IE4 integrated into it's shell? (Like Windows 98) I never heard of such a statement before.... where did you get such information? I thought the USB support build into Windows 95 OSR versions were the same(After all updates) ..... Nashville was a late Windows 95 build (Identifies itself as "Microsoft® Windows Nashville" in the command prompt and is Windows version 4.00.999) but was build before Windows Detroit (Windows 95B), also in 1996. As stated somewhere else, Microsoft intended it to be a "bridge" between Windows 95 and Windows 98. I also saw screenshots of a Windows Nashville version 4.00.1xxx, but I forgot where it came from, and what was it's actual version, but I DID see a version of IE(IE3?) bundled with it. It also possibly came with the same shell. UI was mostly the same as Windows 95, but with some "web" features built into it (But still without internet explorer yet). Such features included: -Underlined link-like filenames -Single click to open feature It also had a graphical bug with windows's scroll bars. It was largely the same as Windows 95 vesion 4.00.950, but had a "dummy USB driver"(Stated below), and a different shell (As stated above). I can't remember if it was bundled with a version of Internet Explorer though.... but I DID saw "USB device" or something like that under the device types list - possibly pre-Detroit (Windows 95B) USB support for Windows. It didn't have those USB drivers though, only it's name. Like pre-OSR2.5 Windows 95 versions, it doesn't need IE to run. I stripped it apart before to ~4MB for my "Running Windows 95 on Thumbdrive" project.... BTW I also discovered that my Memphis build 1525 was not too much newer than the build used my Microsoft @ Comdex...to demonstrate their "Plug n Pray" capabilities... here -> Windows 98 crashes during Gates' Comdex demo - April 20, 1998