Sfor

Indeed. Switching back to APM does remove the problem of duplicated USB drive letters. However, the ACPI buttons are a very useful function. I do not want to loose the button function by switching to APM.

The CardWizzard is necesary to use PCMCIA cards in DOS. The Windows 95 OSR2 is able to use PCMCIA without additional software. However the correct drivers could be necesary. The CardWizzard integrates DOS and Windows PCMCIA support. Without need for DOS support the Windows 95 should do on it's own. I do have a Satellite Pro 460CDT. It does not support Card Bus. The Cards Bus cards are not detected, simply. In case of plin PCMCIA none Card Bus cards Windows 95 detects them and prompts for drivers, when necesary.

I've decided to play a bit with the Letter Assigner. The result is it is unable to remove a drive letter. All it can do is to replace it with some other drive letter. So, it is not possible to get rid of a duplicated drive letter problem with the Letter Assigner.

I've been thinking about removing the duplicated USB MS-DOS compatible drive letters from Windows by creating a DOS utility. It should be possible to remove a USB drive letter from the DOS. But, there are a few problems: - Where the Windows takes the DOS drive informations from? (DOS or BIOS) Is it necesary to remove the drive from both BIOS and DOS disk drive devices list? - The DOS support for the USB drives comes with loading IO.SYS, as BIOS does the whole job. So, there is no possibility to remove a driver or something. It would be necesary to undo quite a few things, probably. - Perhaps a boot manager could do the trick of removing the unwanted USB drives. ------------------------------------------- I've tried to hide the partition on the USB drive using XOSL boot manager. It did hide the drive in DOS, and the Windows showed just one drive letter. Unfortunatelly, Windows detected the SD card as an unformatted one. So, this system is useless. Clearly XOSL is hiding partitions through a change in the partition table. The Windows can not access the drive in such a case. It could be possible to enable the partition back, after DOS booted. So, the windows gets a correct partition information, that way. But, it would require to create an application to change the partition of the flash drive. It should be relatively simple, when compared to the task of forcing the DOS to loose the USB drive letter support.

Well, I was able to get rid of the Fiolog.vxd through disabling the Task Monitor in the Msconfig. But, it changed nothing. So, Fiolog.vxd is just a side effect, not the cause. As for the ACPI. I do believe the applications closing and disk buffers emptying are done before using the ACPI functions. So, since the disk operations were not concluded, it seems the ACPI is not the problem.

It appears the system does not shut down correctly when Fiolog.vxd is still present after shutting down the application. Fiolog.vxd File I/O Logging VxD for Application Defrag 4.10.1998 I do not know if a newer version of this file is available. ----------------------- A further investigation revealed the Fiolog.vxd is a part of the Windows Defrag system. It is enough to disable Task Monitor to get rid of the Fiolog.vxd. Now, there is no difference in the VXD list before and after running the application. Since the Fiolog.vxd monitors the file operations it is possible to form a theory, the Windows does not shut down because of unfinished disk operations. Further evidence is the fact the Windows does the Scandisk check after switching the power off, when it stucks at the shutdown screen. The good question is, why there is no problem with the system reboot? What is the difference with unfinished disk operations between power down and reboot procedures?

The BIOS USB drive support can cause the multiple drive letter problem, as well. My Asus EEE PC 900 is a perfect example. One drive letter comes from BIOS and is recognized by the Windows 98 as a MS-DOS mode drive, the other is through the Windows drivers. Switching the BIOS USB support off solves the problem in most cases. But, the Asus EEE PC 900 BIOS does not let to do it.

I've compared all the files versions from the Microft's article with files on my computer. All of them are the same version or newer. So, it looks like the Gape's Unofficial Service Pack 2.1 did install the shutdown suplement, already. I do have the Unofficial Explorer Lockups SHELL32.DLL Fix and browser .DLL fixes installed, as well.

Windows 98 does suffer from many issues related to the system shutdown procedures. This time I would like to gather some informations about possible explanations and/or solutions related to application caused shutdown issues. I've encoutered a few applications causing the system to shut down incorrectly. In the same time there are no problems with system reboots. After running and then closing one of the problem causing applications, the system works correctly. But when powering down, it stops at the shutdown screen. After a forced reeboot the Scandisk is started. The pesky applications do not show any error messages or unhandled exceptions. Without running them, the system powers down correctly. I can replicate the effect on many Windows 98 running computers. All do have polish Windows 98SE USP 2.1, IE 6SP1 installed. One of the pesky applications is the Gothic game. It was localized to the polish language, so I can not be sure other language versions will cause the same symptoms.

Bearwindows helped me again. It is possible to enable the ACPI support in Windows 98. To do so it is necesary to import a following registry settings: REGEDIT4 [HKEY_LOCAL_MACHINE\Software\Microsoft\Windows\CurrentVersion\Detect] "ACPIOption"=dword:00000001 Then the "new hardware wizard" will detect the ACPI BIOS. However, there is a catch in it. In my case, enabling ACPI BIOS made the windows detect the built in USB card reader. Now SD cards are both detected by the BIOS USB drive support and the Windows giving two drive letters for a single SD card. It does not seem to be a very important issue, but it is a bit irritating.

I'm working on Windows 98 on a Asus EEE PC 900. It works, but a few features are not available. Graphics: VBEMP9x driver - The native 1024x600 resolution is available. - Some minor DOS window application related issues - No Direct Draw support Chipset: Intel chipset utility. LAN card: Atheros NDIS DOS driver. Wireless: I have not tried this one. Web Camera: Did not tried to use it. Sound: No driver available. Power management: It is possible to enable ACPI - Limited selection of the ACPI button events (no suspend option).

The partition image to restore a Windows 9x system is like using a cannon to kill a fly. To make a copy of the whole system it is enough to copy just two folders WINDOWS and Program Files. Also it is possible to do it in an automated way with XCOPY command. To switch between the main system and the copy it is enough to change the folder names, in such a case. It is also possible to keep multiple system copies on the same partition and to switch between them. http://www.msfn.org/board/lofiversion/index.php/t121287.html

Oooops! The problem with DOS applications was not solved, yet. On the other hand, the system does not hang, as it was with the older driver. When starting a DOS application in a windowed mode the screen contents gets destroyed completely. In order to get back to normal, it is enough to switch the DOS application to full screen, and then back to windowed mode (ALT+ENTER). When starting a DOS application in full screen mode, everything works correctly. So, it is a good idea to keep the _DEFAULT.PIF file forcing DOS application to use full screen as the default mode.

I have abandoned the GO OO testing. I've discovered a significant problem with Calc. Everytime I'm trying to copy/paste a range of cells a protection fault error happens. So, this version of Calc is not good for any advanced use. As for the DOCX support testing, I've oppened a document or two. Thats, all. The GO OO works on a Windows 98 without KernelEx, just as OOo 2 does. I know nothing about any system registry entries related to OO, I'm afraid.

Thanks to Bearwindows there is a new version of VBEMP 9x driver available. The problem with DOS applications have been solved. It is possible to run them in a window, now. The 1024x600 screen resolution is available as well. To make it work it is necesary to add a VBIOS TSR utility in the autoexec.bat. Both driver and utility are available at http://www.geocities.com/bearwindows/vbe9x.htm.

It is true for the NT based systems. Since the Windows 98 does not have server type of services (as NT based windows do), all you really have to do is to patch the browser and E-Mail client. The simplest solution is to use Firefox or Opera. If you do want to use IE, then yes. You have to patch a lot. The hardware router as a firewall could be a nice addition, but it is not necesary.

Asus EEE PC does have just USB ports. So, no RS232 available.

I have just tested it. The OO 3 installs and runs with the just released KernelEx 0.3.6. Looks like the problem has been solved.

The Windows 98 Explorer does have a problem related to the integration with IE. Operations on large files, or large amount of small files are locking system resources. This problem does not happen in Windows 95. There are several work arounds available. - The easiest one is to keep IE at version 4. With IE 5.5 it works quite well. With IE 6, the Explorer bug is very annoying. - Windows 98 lite without IE and with Explorer from Windows 95 works very well. - An alternative shell like LiteStep or some other not using the Explorer for file operations. - There is a work around by keeping two sets of browser DLL files. One for the Explorer from IE 5.5 and the other for IE 6. - There is an Unofficial Explorer Lockups SHELL32.DLL fix, as well.

Good news is the DOS sessions are working in windowed mode with the new driver. So, it is significantly better, than the regular one. The bad news is, I can not load Windows with the VBIOS loaded, still. The InfoVBE tool works correctly without the VBIOS. But, with the VBIOS it does not exit as it should to. A lot of informations scrolls through the screen, then stops at some point. It is necesary to reboot the system then. I've tried to redirect the screen output to a file. But, the contents of the output file gets lost after the reboot. So, I can not check if the new modes are present in the VBIOS. ----- Edit The new VBEMP 9x comes with a properly working VBIOS TSR included. After loading the /1024x768/alv_1411.exe in the autoexec.bat the 1024x600 mode becomes available. ----- I found another strange behavior connected with disabled wireless network card. The Windows 98 started to crash at booting when the unknown ethernet adapter device was present in the device manager and disabled in BIOS. To make everything working correctly with both enabled and disabled wireless adapter, I had to let the windows detect both Atheros network cards as unknown network cards, and then I have disabled both in the device manager. Now everything works fine, no matter if the wireless card is enabled or disabled.

I've been observing items disappearing from the transfers window in the Opera 9.6, when the concerned file remained in the download folder. I'm aware of the function removing deleted or moved files from the list. In my case the files were removed without apparent reason.

I was able to build a VBIOS. It seems to be working fine in DOS. But, Windows 98 fails to load with the VBEMP 9x, if the VBIOS module was loaded to the memory. I found no indication how the VBIOS should be loaded, so I've added it to the AUTOEXEC.BAT. I was unable to switch the SNAP driver to the VESA mode. The "gaoption.exe vesa on" command brings an iformation the nga_w32.dll could not be loaded. The IEGD package suggests it is possible to build own EFI module. The EEE supports EFI, as far as I know. The linux version of the EEE PC 900 does have the EFI partition by default. Still, I do not know what to do with it.

No, it does not seem to be the case. So, the BIOS reports more than enough of memory available.

Is there a VNC solution capable of transferring screen data from applications running DOS full screen mode? The Real VNC 4.1.2 works qute well, with exception of the DOS applications running full screen mode.

The conclusion is the Opera 9.5 introduced some serious speed related core changes. The current 9.6 version is the result of trying to make the new design stable again. The old 9.2 design is slower but more reliable. It seems to be logical to wait for higher than 9.6 version, before upgrading from 9.2.