rloew

Not quite. Long File Name entries hold up to 13 Unicode characters. Your LFNs are 46 characters (the dot is included). They require 4 Entries. The Short File Names use one for a total of 5 Entries per name.

A Folder can only hold 65536 Directory Entries. Long File Names take up 2 or more entries depending upon length. Apparently your html Filenames are taking up an average of 5 Directory Entries each.

Well, the 128 byte and 1024 byte sectors in the above KB are from the time of MS-DOS 1.0 and 2.0, which was pretty much (if not completely) the pre-hard-disk era. So the BIOSes of the time (especially the original IBM one) must have been able to boot with these now-obscure/defunct sector sizes. Joe. I assume so. Modern BIOSes do support these formats. They can't be booted because the Boot code does not scan for them. It only tries 512 Byte Sectors. I did create a Bootable 1K Sector Floppy by slipping in one 512 Bytes Sector.

I agree. Maybe if KernelEx comes more stable in the future, I'll consider it. ATM is not going to happen.It seems to me, guys, your posts are extremely unfair with KernelEx and give a completely wrong impression of its compatibility with programs running on stock 98 systems and stability. KernelEx is rock solid and there are currently no known pending issues with it and programs also running on stock systems. I personally did not comment on it's stability. The Package has been stable because development has stopped. There are many things missing or only stubbed. The latest versions of SumatraPDF cannot print because KernelEX stubs PrintDlgEx. Some MSFN members worked out a workaround, and I have a DLLHOOK based fix in my KEXEX Package, but it belongs within KernelEx. The fact that it is sometimes necessary to disable KernelEx for individual Programs is a sign of the issues I pointed out. If a Program is not listed in a compatability database, the user has to debug his setup by trying different combinations. This can be particularly diffcult if the Package has multiple Executables especially if they are not well documented. In addition, if KernelEx cannot completely hide it's existence when disabled, then a Program that runs on a Stock system will not function properly. My original DLLHOOK Program and Packages based on it such as my 64-Bit File Emulator were incompatable with KernelEx regardless of setting. I had to redesign DLLHOOK to work with and without KernelEx. The second part of my comment relates to the subtle and often undocumented changes to Functions that can cause a Program developed on a KernelEx based Computer to not work on a Stock system, even if no new Functions are Called. I have been working on a Project, I call WDMEX, which is a superset of WDMSTUB. I am currently attempting to load the generic USB 3 Stack Drivers for Windows XP that appear to be in common use. I already found a case where a Driver calls a Function that exists in both Windows 98 and XP, but they are not the same even though the documentation is the same. The XP version zeroes out a returned buffer while Windows 98 does not. Clearly the Programmer who wrote the Driver assumed that the buffer was zeroed and of course it worked in XP, but it crashes in Windows 98.

Seems to me that if you go down that road, and I would like to see that, that KernelEx and ALL the files that are "For KernelEx Systems only" should be ONE single option. That way the folks that want a more "pure" solution can have that, and the folks that want a more expanded solution could have that with a group of files that have been tested to all work together correctly with the basic 98 SE SP 3.x files. Any more fine tuning of which files to include and which to leave out puts too much of a burden and workload on PROBLEMCHYLD. A single all or nothing kind of option only. For anything else the user can choose not to select the option and then add in the individual files that they want. Cheers and Regards How many options is up to him, but including it in the basic package is not a good idea. I use a "pure" system in all but one of my Windows 98SE Instances. KernelEx alters the nature of Windows 98 in ways that can cause problems with Software not written to expect it, or cause developed Software to work that will then fail on non-KernelEx Systems.

They would not only have to be OPTIONAL but they would have to be identified as "For KernelEx Systems only". Of course, at that point, you might as well include KernelEx as another OPTION.

Yes, we are talking about file-based cloning. Beyond that, I fail to see how this is relevant to my question. I'm not asking anyone to do any tests, I'm just wondering if anyone actually knows. When I get some time, I'll put some files with long names (that have conflicting short names) on my ramdisk and clone it to a flash drive (or floppy or another ramdisk) using this drag-and-drop, icon-to-icon method. If all names match (long and short), I'll do some renaming and retesting in an effort to get it to fail. Slow method? Yes. Easy method? Yes (if it works). Third-party app required? No--just what's already there. Not only is it necessary to create conflicting short names to get ~2, ~3 etc. but they have to be out of order in the Directory so that xxxxxx~2 comes before xxxxxx~1 in an unsorted Directory. The following should do the job: 1. Create a new Folder x 2. Copy a file to x\aaaaaaaaa1. 3. Copy a file to x\aaaaaaaaa2. 4. Copy a file to x\aaaaaaaaa3. 5. Delete file x\aaaaaaaaa1. 6. Delete file x\aaaaaaaaa2. 7. Copy a file to x\bbbbbbbbb. 8. Copy a file to x\aaaaaaaaa2. 9. Copy a file to x\aaaaaaaaa1. The source file can be anything, preferably small.

In this case the Drivers need to be cleared from the Registry. The Driver Files and INF Files do not need to be removed, except those being directly replaced by NUSB. The Drivers should be rediscovered. I already have a REG File that can clear ALL Drivers, but not one that is USB specific. It comes in handy when porting to a new machine. Some .INF Files may ask for Disks over again but they can be skipped or read from the System and Driver Folders. WARNING! do not try this on an NT, XP or later system as the Driver Files will be truncated.

of course target must be partitioned/formatted/SYSed, cannot remember if I used some additional tools/steps, exception made for temporarily disabling pagefile (no pagefile) and running in safe mode, original source: http://www.duxcw.com/digest/Howto/hd/cpyhd/cpyhd2.htm The directory dates are usually of no relevance (keeping 'em or having them changed won't make a difference in normal operation). jaclaz Many years ago. I did an XCOPY of my Windows 98 System. The basic system worked OK but some *.LNK Files were jumbled.

You would need to copy everything including attributes and Short File Names to avoid problems. Most copiers do not match up the Short File Names under all circumstances. Also Directory Timestamps will typically be lost with most copiers. My XFILE program preserves Short File Names and can Copy Directory Timestamps if run from Windows 98. @Dencorso: Does XXCOPY preserve Short File names. The documentation does not mention if it does. Another consideration is time. I have a standalone duplicator that can copy an entire 2TB Hard Drive in about 7 hours. It would take at least 24 Hours to copy the Partitions, even longer to do a File by File Copy. Afterwards I use RFDISK to eliminate all of the duplicate Checksums and Signatures.

Which "signature"? You mean volume serial number? I.e. the data that is @27h in a FAT16 and @43h in a FAT32 PBR ? jaclaz Correct.

The "mystery bytes" do not need to be different but something does. The Checksums of the MBRs must be different or confusion can occur. My RFDISK program puts it's timestamp in a different location. Amigas use a different Format entirely. Connecting an Amiga Hard drive to a Windows 9x machine will corrupt it, as Windows will set the "mystery bytes" and break the Checksum. In addition to the "mystery bytes", the Signatures in each Partitions Boot Sector must be changed so that they are all unique. Otherwise you may be accessing the IDE Drive when you think you are accessing the SATA Driver or vice versa. I added a feature to RFDISK to check and repair this issue.

I remember having gotten a similar value. Maybe old routers use old FTP specifications http://tools.ietf.org/html/rfc5797 or the MS FTP server doesn't get along with old routers. My Amiga Software is more than 15 years old, so I doubt that their server needs newer protocols.

I downloaded the entire set twice on my Amiga. The results were exactly the same, byte for byte. Other than the three retries I see no issues with the Server.

Because NUSB 3.5 (and by extension the Service Pack) replaces all of the USB Driver stack files. That includes ALL USB devices, not just storage. If one installs either of these packages without first removing devices, it can lead to problems, such as those described by Drugwash in the NUSB thread. Clearly removing Storage Devices is necessary as the drivers have been changed. In the case of non-storage Devices, it would be necessary only if different Drivers would be chosen and/or the Registry Interface has been altered. Otherwise they would be found after the Controller Drivers are installed. Someone should test different USB Devices to see which ones need to be removed or not. Some, like the Keyboard and Mouse, probably do not need to be removed.

You will need my SATA Patch. This motherboard has Gigabit Ethernet. This may lead to other issues at a later date.

In the Win98 clone I have replaced in SYSTEM.DAT and USER.DAT (and all relevant .INI and .LNK files) all occurrences of "C:\Windows" with "F:\Windows". So when I am running the Win98 clone it only does business on F drive. Of course I would stop doing that if I set up the HDD with one of the methods you guys suggested, and simply make another self-referencing Win98 clone elsewhere on the disk. There is also C:\Program Files and a few C:\ references in the Registry.

There wouldn't be any confusion if people followed instructions. The instructions specifically says REMOVE ALL DEVICES.Not some, not a couple, but ALL. Why are peolpe following NUSB instructions? The heck if I know. They should be following the Service Pack instructions. If my one year old daughter can understand All when asked to pick up her toys, then any adult should be able to do it also. Finding ALL Devices may not be so easy, especially if using USB Keyboards and/or Mice. Some Drivers are hidden under "Human Interface Devices" for example. Also removing ALL Devices may be difficult. On my MA785 Computer either my Keyboard or Mouse must be USB. Some Computers require both to be USB. Others may have internally mounted USB Devices. The !README.TXT in NUSB35E.TXT fails to mention using Safe Mode. It also tells you to remove the Controllers, which is impractical. It should not be necessary to remove Devices if installing from Safe Mode as they are not detected.

The clone on F contains references to C:. If you run it using a boot floppy, you are probably still accessing a lot of things on the C: drive. Changes that you think are being made to the F clone may actually be made to the C Drive. Yes. My RFDISK Advanced Partitioner can reconfigure the Partitions without losing data.

The y2k Directory was included in the totals. I had downloaded the different OS sets and did a Count on the WIN98 Directory. I did not see any corruption or differences. I downloaded the entire set twice. There were three failures, one on the first download and twice on the second, but selecting Retry resumed the download. The two copies were identical. I transferred the files to my Windows 98 Computer over a LAN. I got the same count, so no questionable File Names were present.

Windows 98 does not support setting the date on a Directory. If you download or copy a Directory tree, all of the created Directories will have the date the new Directory was created, not the date of the Source Directory. Cut and Paste does not Create any Directories, only renames them, so the original date is preserved. I was able to get around this limitation in my XFILE Program by bypassing KERNEL32.DLL but Explorer does not. Incidentally, by downloading the Reskit using the FTP: Driver and XFILE on my Amiga, I got 2389 Files in 95 Directores.

The SATA Patch is on my website http://rloew1.no-ip.com

Considering that many files have large blocks of zeroes in them, Disk readers should replace bad sectors with something else rather than zeroes. This would make it easier to identify the damaged files.

I already have many of the missing NT Functions. The 3 remaining NT Functions are not too hard to implement. The missing USB and TDI Functions, I'm not so sure about. There are some Bluetooth devices that support Windows 98. They may have useable Drivers.

Hi rloew,With one wireless router in Europe the 3rd digit is 2, and I wasn't able to change it to 1, like for my other routers. That means whenever I am there I have to change the Windows IP settings for the gateway, of the laptop I take there, and to change it again when I am back in the US. I am surprised. I have never seen a router that did not allow an arbitrary setting for the Router's IP number. Thanks for your hint, VERY helpful. Your Buffalo has apparently only 100 Mb Ethernet, like my current spare router. My son does games etc. on the internet, and he complained already that the spare router is too slow, with a 100Mb cable connection, so I probably will get a faster/Gigabit router. Your Buffalo router is on the dd-wrt compatibility list, have you installed the dd-wrt firmware in your router? dd-wrt firmware is open-source, so no government spyware. When I looked at their page about "Static DHCP" http://www.dd-wrt.com/wiki/index.php/Static_DHCP I saw the mule, my favorite animal, and my router-phobia was forgotten. I use the router mainly to access the Internet, which is a lot slower than 100Mb/Sec, rather than as a local switch. I have a Gigabit Switch to handle Gigabit capable Computers. I haven't tried to share IP numbers among multiple MAC Addresses. Another option is to override the MAC Addresses of the Cards, so they can use the same MAC Address.