jds

Another SP item that should be an optional install : Adaptec Windows ASPI driver (version v4.71.2). If you use a Domex SCSI card such as DTC3181X, DTC3151X, DMX436, DMX-3181LE, DTC-3181LE, DTC-3151 (also known by various OEM numbers such as UDS-IS11) and you install the Adaptec ASPI driver, your Domex card will stop working (it will show OK in Device Manager but will freeze if you try a SCSI data transfer). This family of SCSI cards are ISA and have a somewhat unique feature of being able to operate sans IRQ (for some, use of an IRQ is optional). While this is a disadvantage in regards performance, if you have no free IRQ's (ISA can't share), it's just the ticket. I suspect it's this IRQ-less operation that Adaptec's drivers break (AFAIK, all their SCSI cards use IRQ, so they probably didn't consider this). BTW, Domex state that this card and its associated drivers can only be used with scanners, but that's not actually true, I have successfully used one with an external Sun CR-ROM drive. Joe.

That's just bad news, isn't it? If there were a 9X version of the library, then the offline registry editor tool might be rebuilt and hopefully work OK with the 9X registry files. But as an offline registry editor works directly on the files, not the loaded registry via an API, how can an NT version be used, since 9X has two registry files, whereas NT has about half a dozen? In other words, if I were to (re)build NT registry files from exported 9X "reg" file(s), how would I merge the half dozen NT registry files into the two registry files that 9X requires? Joe.

Correct, this is a modem setting, so it's the modem manual you'll need. For an Ambient chipset, it's set by the "ATX" command. Here's an extract from the manual : Joe.

I'm not entirely sure what configuration you're trying, but the obvious thing would be to get a PCMCIA Ethernet adaptor, then you could network your two computers together (simplest connection is using just a cross-over cable). Joe.

1GB DDR2. Check the thread(s) about running/installing 98 with more than 512MB of RAM, that's probably something to do with it. Joe.

Agreed. See what proportion of scanners at virustotal detect malware and judge how dependable they are likely to be. Also, don't trust any report file XYZ is OK unless there is an MD5 or SHA1 or similar hash given that you can use to see if yours is the really the same file. Joe.

Ummm ... isn't that the infamous version number that almost always meant a 32GB limit bug, beyond which the BIOS would fail to boot? I've never seen (and I've disassembled/reassembled and stripped/recycled quite a number) of that "version" - every one of them went at LEAST up to the 64g-limit and generally had a "patch" for the 132g-limit. It all depends on the MoBo/ChipSet. For example, I have used "patched" on Tyan boards many times (Via Chipsets). The "Base BIOS" (eg 4.51PG) is not necessarily the "problem" - Take note of people actually REQUESTING that "BIOS" for their MoBo's - just THAT is totally... NOT! Pull one of them apart with a Utility ("editor") and you' ll see what I mean. Please note THAT is what the OP has (and supports up to the "highest limit"). @rloew - Thx for details all in one place. Wow! They say YMMV, and they're right! I've had several different MB with version 4.51PG, including my recently decommissioned P2, and every one was completely different and every one had the bug! But I was aware that some MB with this version number are OK. I actually wrote some tools to extract and repack the different code modules within these BIOS (that's why I can say each of these with apparently the same version number were in fact quite different) and on the P2, located and patched the faulty code for 32GB and 64GB - painful!. So that's why your "not known to do weird stuff" comment caught my eye. Joe.

Thank you sooooo much for those interesting links. It's taken a while to follow all the threads/leads. I've downloaded 'offlinereg.zip', 'RawReg13.zip' and even 'regv.zip' and 'regs.zip'. I didn't download the "regfs" stuff, because I have my doubts it's a viable option here. I'll try to find if a 9X version of the "Offline Registry Library" was ever released, since at a file level, 9X and NT registries aren't compatible (they don't even use the same file names or number of files). Then I'll see if there's any practical way to apply any of these tools to this problem. Joe.

Welcome, pulteny. I don't use either app, but suggest you upload 'w9xpopen.exe' to virustotal and check it out. If it looks OK, presumably there's a way you can configure S&D to ignore it. Also report this issue to whoever are the respective app support. Joe.

Thank you for the tip, Mr Loew. I'll keep that in mind for the future. Joe.

Couldn't reach that URL, but this looks like the same thing : http://web.archive.org/web/20030810121208/http://support.wdc.com/download/dlg/dlgmaker.exe There is a similar thingy for WD drives: http://wdc.custhelp.com/app/answers/detail/a_id/936/~/operating-system-and-bios-limitations---137-gb,-32-gb,-8.4-gb jaclaz Thanks, Jaclaz. Cool, another tool for the toolbox. This one's for MS Windows. That's what I thought about. I don't know if Seagate Tools will work for a WD, but HDAT2 should. I'll try it in a while. Any idea if it's better to limit through setting the HPA or DCO (assuming both will affect BIOS detection)? AFAIK, the Seagate tools will work with any brand of HD. The first picture (HPA) looks like the one you need. Setting the User size should do the trick. Ummm ... isn't that the infamous version number that almost always meant a 32GB limit bug, beyond which the BIOS would fail to boot? Joe.

Not quite. My IO.SYS fix resolved a problem that I discovered and fixed with an earlier Patch long before your investigation. The current Patch fixes both. Really? Perhaps you've forgotten our little debate over the merits of Steven Saunderson's patches a few years back, in which (as I recall), you initially said these were only needed on NT systems. At that time, I had already found two of his patches to be required, whereas one was detrimental. Anyway, your current patch is the best because it also addresses another issue (and our collective thanks to you for it!). Joe.

Well, there is a virtual clamp jumper available. Seagate Tools for DOS gives you the ability to set any clamp size you wish. Just temporarily put the drive in a machine that can handle the current/native size, and run the tools from there. Note, this feature is not available in the MS Windows version of the tools. Joe.

Yeah, I can see that. Since you have clearly forgotten, apart from wasting my time testing PTP/MTP drivers for you, rloew's 'io.sys' fix in the SP stems from my investigation of LBA bugs in the MS version and tihiy's patched 'secur32.dll' in the SP is due to my investigation of the missing 128 bit version of this DLL. It means that I need install all components of SP3 except USB component? No what it means is that USB install as a part of main updates. After installation of main updates follow the instructions in the USB 2.0 notice. I presume that if I'm happy with my current USB drivers, I can simply leave them installed and the SP3 won't affect their operation? In other words, if I choose to uninstall my existing USB drivers, then NUSB will take over, otherwise NUSB will only affect new USB devices? I made a perfectly reasonable request, which did not deserve your torrent of abuse. I have nothing more to say to you. Ciao. Joe.

Wow! Mine's 13MB and 1.8MB respectively. I recently deleted dotNet 1.0 and 1.1 frameworks because my system was becoming unstable. I believe that was due to this registry size issue, since I wasn't actually running any dotNet applications, particularly not the 1.0 or 1.1 varieties. Joe.

Thanks, Den! Joe.

At the risk of being misunderstood in saying this, you do have a tendency to interpret what people tell you in a negative light. We all hold you in high regard for your fantastic efforts here, and if you think people are being critical of you work, almost without exception, that is far from their intention. Read what they say from a positive perspective, you'll be closer to interpreting their actual meaning. HTH Joe.

BTW, perhaps our interpretation of the word "remove" is different, however, let's put that aside for the moment ... I can't see why, if NUSB is made optional and the user doesn't select it, they would need to follow the USB instructions regardless. The whole idea is that they can select it, and therefore need to follow the procedure, or not select it, in which case they don't. What am I missing here? This option would be handy for all users that wish to apply the SP to a working system, which already has USB drivers installed. I doubt I'm the only one with such a need. In fact, I would suppose that the vast majority of users of the SP want/need to apply it to working systems, and those users doing a "clean" install would be a minority. Another example of why the present USB procedure may not be suitable - on one machine, I use a Sony digital camera which, although it seems to use the USB Mass Storage class (not that horrid PTP/MTP stuff), only works with the Sony drivers, not generic ones. If I were to follow the SP procedure and delete the Sony driver and install NUSB in its place, I would no longer be able to access the camera. So why should I be forced to install something if it's just going to break my system? Joe.

Thanks for your consideration. I am not asking for the removal of NUSB, I am only asking if it can be made an optional install. If you look back in this extensive thread, you will see that the USB stuff is a regular item that people need help with. For some people, removing all their existing USB drivers to install the SP is no big deal, but for others, it is at least an unnecessary nuisance. For myself, I'm not comfortable upsetting a working driver combination unless I really have to (or as they say, "if it ain't broke, don't fix it"). I also prefer to use official USB drivers for my devices (eg. JMicron, Lexar, Intel), as appropriate, rather than unofficial drivers such as NUSB. I have used NUSB on some machines, so I'm not antagonistic to it, but only where no official drivers are available. So again, this is just something for consideration, which makes life easier for users and also for whoever maintains the SP (due to less queries about the USB requirements). Joe.

Thanks for the input, Mr Loew. I also tried this by exporting the individual hives (HKLM, HKCU, etc.) separately, then importing these REG files one-by-one. I used the User.dat (/L) and user name (/R) command line switches to avoid breaking the working registry. Unfortunately, after a certain size was reached (can't recall exact size, something over 8M), importing further hive REG files failed. Joe. PS. To our respected mod's : I think this thread should be listed as Important/Stickified/Pinned.

NO! SP3.x installs USB 1.1/USB 2.0 drivers. Just because you don't want to follow instuctions doesn'tmean I'm going to complicate things for myself and others. You have the following options. 1. Don't install SP3.x at all 2. Use a older version and replace all updated files manually. 3. If you do install SP3.x follow the instructions 4. Learn how to do it yourself I think what agitates me more than anything is complaints about things that can be avoided. Don't use the SP if you don't like whats in it. One guy said he wouldn't install SP3 because of the changed icons but also said he wanted system stability. He have the same choices above. For something thats FREE, I have never seen so many people complain. DIY I agree 100%. My thing is, I 'm willing to compromise. I can create a package without the updated WinME/2000 icons, but then what about the people who want them. The people who complains the most are the ones who contribute the less.I tested the PTP/MTP stuff for you, didn't I? Why do you insist that NUSB is mandatory? I believe my request was not unreasonable, nor framed as a complaint. Joe.

I think this thread should be noted here : Also, if anyone has any suggestions on how to recreate a large registry from exported REG file(s), that would make this thread even more useful. After all this time, I still haven't found any way to do this. In theory, the most reliable way to make a registry devoid of wasted space is to export it to a REG file (or files) and then import the file(s). However, while the first part of this process works fine (albeit slowly), the second part of this process fails with large registries. Joe.

Last night, I installed SP3.5 on my 98SE laptop which had IE5.01SP2 and KernelEx 4.52. I selected about 2/3 of the options. All went smoothly. One thing I'd request though ... Can the USB drivers in the future SP3.6 be made optional? I won't benefit from USB 2.0 drivers on my USB 1.1 machines, and I've already got Intel's drivers running on my USB 2.0 machine. As I'm happy with the USB drivers I already use, I'd rather not mess with them, if and when I chose to install the SP on my remaining machines. Joe.

Thanks, loblo & tomasz86. This is potentially very useful for running dotNet applications built with VS2008 (or VS2005 with the dotNet 3.0 update) or later. I found the OnePiece packages easier to dissect for required "assemblies" than the MS ones. Don't make my first mistake of downloading the usual dotNet 3.X framework installation packages from MS, these are downloaders only and don't contain anything useful (the "full" packages from MS should be OK, but they are huge and the embedded files need renaming). Use the Remotesoft .NET Explorer to find out what "assemblies" you need. The way this seems to work is that you open your dotNet 3.X application, expand the tree view for the Dependencies branch, then select the properties of anything listed there. If a normal Properties window appears or the error message "This file does not have a program associated with it ..." , then the "assembly" exists. If the error "Can't resolve reference ..." appears, then the "assembly" is missing. You also need to do this for any "assemblies" you bring in, to find their dependent "assemblies". Unfortunately, Windows Presentation Foundation (WPF) applications don't seem to work. Perhaps someone more knowledgeable can figure a way ... Joe.

Same for 8.22 with the following 'core.ini' entries : [DCFG1] contents=Kstub822,std,kexbases,kexbasen [NT2K.names] ComDlg32.PrintDlgExA=Kstub822.0 ComDlg32.PrintDlgExW=Kstub822.0 Joe.