dencorso

In principle, any key retrieving software, like Magical Jelly Bean Key Finder should be enough to solve your problem, if the OS asks for the key after changing just the HDD, which I doubt it will.

Err... @xper: AFAIK. he's right. So, would you please change SearanoX ==> Tommy ?

That's what Murphy's Law is all about. I was able to unpack both installers I tried (just two of them)... But the one above resists the WinRAR 4.20 / UE strategy I used with those others. It seems you're right: it says it's a "Compaq Builder Generated Installation Executable", and seems to be a customized InstallShield. It'd be necessary to run them through a debugger to get them to brute-force unpack case-by-case... And that's more work than I can do right now. Sorry for raising unwarranted hopes.

WinRAR 4.20 can unpack the drivers, at least those I tested, then the Universal Extractor can take care of the cabs.

One example is KB2724197: MS blabing mumbo-jumbo about 'security', but said nothing that it would impact EMS availabilty on NTVDM. Well, the previous "security" update, KB2707511 already caused NTVDM to crash on opening a pipe, an issue that hasn't been fixed. Since security updates are cumulative, KB2724197 must have both issues. Now, ain't those latest updates really awesome? Just to yet again point out that maybe two rollbacks may be in order, not just one... KB2633171 being the last non-flawed krnl set. I can't believe that Microsoft hasn't been very helpful remedying the issue of EMS not working for 16-bit apps under Windows XP as I personally believe that the important of the 16-bit Windows subsystem is increasingly deprecated. True enough. BTW, if you really need EMS, EMS Magic seems to be the way to go... That's a very good idea! Good news! At least the NTVDM pipe and UMB issues seem to be solved by KB2785487. Not sure about EMS availability, but EMS Magic is known to solve that.

They rock!

The Gavotte Ramdisk will accept up to 4 GiB on the default FAT-32 formatting flawlessly, so I guess a 3 GB ramdisk is in order, leaving 8 GB to the system. Sounds nice. BTW, the latest Gavotte version is in post #36 of the pointed thead.

BlackWingCat's modified kernel plus the Gavotte Ramdisk should do nicely, then.

Did you try the kernel moded by BlackWingCat?

@LoneCrusader: would you agree to this? I think it's a nice solution.

My own copy of o2ksr1adl.exe was downloaded directly from MS, too, way back when. It does, as expected, have the properties reported by submix8c, that is: By appending CR,LF (0x0D,0x0A) bytes to the end of o2ksr1adl.exe, with a hexeditor, it grows by two bytes and then presents the following properties (exactly as reported by Peter): So the file Peter downloaded is bona-fide and can be used safely. It's not that uncommon to get files with some bytes appended to the end of the, usually either CR,LF or some zeroes. They cause no harm, nor modify the functioning of the installer in any way. Sometimes, some trailing zeroes get deleted also... And then it's more complicated, because they must be restored by trial and error... Both are part of those mysteries of the internet that are better left alone...

Thank you very much for the feedback, Leolo! We sure are interested. Attached is a good quality image of the ACARD adapter with ARC772 chipset, taken from the site you offered a link to, just to ensure we'll keep on having its image even when the original site takes it down, since it's quite useful for reference:

Hence, I stand corrected: the last free version of XXCOPY guaranteed to work with 9x/ME is really v.2.96.8 - xxfw2968.zip. From that zip one can extract XXCOPY.EXE and confirm it has the following characteristics: Size: 385,024 bytes PE Timestamp: 0x4BEF6642 or Sun May 16, 2010 03:28:02 GMT CRC32: FDE5CCB3 MD5: E8DDA25DBDB7FDC93FAC471FBEBD9B84 SHA-1: 6F166273EA975148C3C0477EA9CDD2ACDD9F561A Bear in mind that the new patch pattern is named "xxc2968.pat", not "xxcopy.pat"... So, the patch command must now be: patch -p XXCOPY.ORI xxc2968.pat XXCOPY.EXE And, after patching using the patch pattern attached to this post, all having gone well, the new, nagless version of XXCOPY.EXE shall have: CRC32: 731C8F2F MD5: 721AF3B37412BAC880E528A341FFC14A SHA-1: A0B0C766CB8B010C50EC860613C9444DA6F7C028 xxc2968.7z

If you prepare this on an NT-type system (2K and up), HashMyFiles is excellent (stumbled across on MSFN - worth re-mentioning...).

Why would anyone ever consider creating an *UNrestorable* backup of anything? Or, more to the point, can an *UNrestorable* backup of something actually be described as a backup at all?

Did you actualy even try to seach the forum?

Yes to both questions. Search around here in the Office forum and you'll find what you seek.

No, you didn't need to activate for that reason, since there's a grace period (30 days, I think, it's been a long time since I used that for anything). So you could have set everything up and had it run in the office for a couple of weeks legally, whithout activating.

My POV is similar to bphlpt's... Since IDE is natively recognized by 9x/ME, there are no drivers involved when the conversion is SATA => IDE (i.e.: connecting a SATA HDD to an IDE motherboard connector)... now, the other way round is another beast entirely, because most SATA boards require RLoew's patch to work, unless they're those old SATA I (which use RAID drivers even with no RAIDs present). Hence. IMHO, it's hard to define this matter as Win 9x/ME related at all. On the other hand, I agree the thread was started with 9x/ME in mind, but it seems to me to have wandered very far away from that. Be as it may, I'll let it remain here some more, but I feel we may be loosing interesting contributions to this thread precisely because it's "hidden" in the 9x/ME forums. That's just my 2¢, of course.

Windows is behaving as expected. You think it's the same hardware. It knows it's not (different serial numbers and the like). If you had cloned without first activating, it might work. By cloning after activating you're trying to use the same license on 3 further machines and windows is refusing to comply. That's by design. BTW, read the Rules, especially #1a.

AFAIK, you're right, that's the one. In any case, here's a link you might want to explore: Ted's Blog

I don't think Leolo is interested in 9x/ME, really... And I do think this thread may be in the wrong forum, in fact. Please advise.

I'd go with the Marvell 88SA8040 based board, but warn the seller that the Marvell 88SA8040 is mandatory.