cc333

Is it possible that driver signing could be an issue? I've found that some drivers won't run if not signed properly. Of course, this is on 64-bit Windows, which i believe has a somewhat stricter driver signing policy than 32-bit. c

How so? c

I forgot about this thread! Last November, I was riding in a taxi in midtown Manhattan, NY (I don't live there; I was visiting from California), and came across this peculiar sight on the video screen: I'm pretty sure that whatever's driving the display is running XP, judging by the styling of the dialog box (all later versions look different). What makes it puzzling to me is that the taxi couldn't have been more than a couple years old, so why didn't they go with a newer system? c

I have a couple thoughts, if I might: Firstly: could it be that your trackpad is somehow intermittently faulty? Drivers which, by all accounts, should support your hardware and run on 8.1 aren't, and sometimes this can indicate that there's a problem in the hardware (or perhaps firmware) somewhere, in my experience. Of course, this isn't necessarily so, and is consequently rather unlikely, but when all else fails, it's something to consider nevertheless. Secondly: I know you're reluctant to redo your XP install, but have you considered upgrading to XP 64-bit (hereafter XP-64)? API-wise, it shares a bit more in common with Vista than XP 32-bit (XP-32), and thus also seems to support some hardware that doesn't ordinarily support NT 5.x-based Windows. The reason for this is that to some limited degree, XP-64 can support some 64-bit Vista drivers, provided said Vista drivers don't use APIs that don't exist on XP-64 (which, while not as troublesome as with XP-32 and it's somewhat older codebase, can still be a problem). Indeed, the reason this is possible is because the kernel of Windows Server 2003 (upon which XP-64 is based) was used as the basis of Windows Vista. c

OK, got back to my 98 machine, and the last version of CTR that would install on Firefox 31.x is 1.5.3. To fix the rendering artifacts in the tab bar, I've found that setting the appearance (first drop down list on the "Tabs (1)" screen) to Australized V2 works well. Still no Javascript, but at least we're at a point now where the UI chrome is 99% functional (the popups for installing addons, saving passwords and such have big black lines around them, likely due to the fact that the transparency effects they use are broken and missing on Win9x). c

An old version of Classic Theme Restorer will also work. I don't recall which version, but it works fine with 31.x, even on Windows 98SE. c

I have standard Firefox 31.8.0 ESR more or less working with only standard updates + KernelEx 4.5.2 + latest KernelEx updates. The only issue (and it's a big one) is that javascript is almost completely broken. I have to turn it off in order to browse anywhere, or else the browser will lock up hard and then crash, occasionally taking Windows out with it. No one seems to have devised a fix as yet, but it stands to reason that the problem must be a manifestation of an incomplete or nonexistent implementation within KernelEx of something the javascript engine depends on. I would think that if that were so, it would throw up an error, but who knows? Maybe it's failing before it can? As the community researches the problem, it can hopefully be better understood so a fix can be made. c

I've been following this interesting adventure, and have a radical thought: is the trackpad the same physical shape/size between the new and old machines? Perhaps you can swap trackpads between them or, if they're incompatible, find another variant of this model whose trackpad is 1) similar to tthe one on your new machine, and 2) known to work on Windows versions earlier than 8.x or 7 and swap that into your machine? If it requires too much hacking (it might), it's probably not worth the effort, but it's nevertheless something to consider. c

I've heard of DVD-RAM, but never CD-RAM! c

@Nojus2001 Interesting. I guess 2003 somehow leverages 2010's MSO.DLL to facilitate its ability to open docx files, and the non-XP compatible version likewise breaks it? c

Hmm, OK. Maybe you can try 2000 with BWC's extended core? With some careful tweaking, it can run much of the software XP-32 can, but with somewhat less overhead. c

Perfect! That's exactly what I wanted to hear! Armed with this information, I shall now go install and update Office worry-free c

You can do that?! Can you install any 64-bit OSes? c

This is a potentially cumbersome solution, but what about getting an older PC, such as something from the Core2 Duo era (or even something like a Raspberry Pi or some other similar embedded-like device), and installing a pair of VMs on it, one for WRP+Chrome, and one for iSH? Then, in theory, all you'd have to do is chain them together (which, as I said, could potentially be rather cumbersome) and set it up so the computer running WFW3.xx and/or Win9x accesses it as it does your iPad now, except it'd be a one-stop-shop that incorporates both solutions into one package. Thoughts? c

OK, I'm planning to install/upgrade my old copy of Office 2010 on this old XP install I've resurrected, but I want to know which updates have the bad MSO.dll file so I can avoid them. Without having to read the past 15 pages, is there a definitive list of every to-be-avoided update released thus far? c

I'm updating an older XP Home installation (2011), and it's gone fairly smoothly for the most part, but I seem to have hit a snag with the .Net Framework 4.0 updates. They all fail with error 0x800B010B. I'm sure this has been encountered and solved several times over throughout the course of this thread, but I have very little desire to scan through 200+ pages at the moment EDIT: I should clarify that I've also used the dotnetfx cleanup tool and reinstalled the whole of dotnetfx 4.0 + updates, and I've applied the POSReady hack before doing so. I should also note that all the updates that are failing are post-2014 and are designated as being for Windows XP Embedded, which could be the problem, maybe. I note also that every post-2014, non dotnetfx40-related update has installed successfully as far as I can tell. c

That makes sense. However, JS on these same browsers (such as it is) works fine when they're run on a natively supported OS (XP), so I'm inclined to think its depends on some sort of API that KernelEx hasn't fully implemented yet. c

Hmm, OK. I guess the only two options, then, would be to either re-implement from scratch a modern JS engine/Gecko in its entirety (very nontrivial, I would think, but probably ideal since it could be highly optimized specifically with 9x in mind for maximum speed and efficiency), or tweak/extend KernelEx so that Gecko 3x works more or less completely as is (probably not as fast or efficient, but should be much more possible with the tools at hand, and a slow, inefficient browser that is stable and usable is far better than a fast one that crashes constantly). We're already 80% there, due to the fact that, JS notwithstanding, Gecko 3x otherwise starts up and renders sites with little or no adjustment. The key is to make the necessary adjustments to KernelEx, Windows, Gecko, or some combination thereof, such that JS will work. If I knew more about coding and such I'd give this a try myself, but alas I don't, so all I can do is churn out ideas. c

Huh, OK. It worked on 3.6, at least. I guess built in support still works in these more recent versions. By the way, I have 2.29.1 up and running, of course without JS (leaving it on resulted in the same crash as it's Firefox cousin). Is it possible there's some APIs that Gecko's JS engine uses which KernelEx isn't representing fully? This requires extensive investigation, I think, because it would be relatively easy, I would think, to extend KernelEx to encompass the missing API functions, if any, even if they're only minimally functional stubs (which would be better than nothing, and may allow JS to fail more gracefully rather than crashing the browser, at least). Nevertheless, even with the JS problem, this is a super excellent advancement in the state of web browsing on Windows 98! Here's a thought: what about taking an earlier, known working JS engine (say, Firefox 3.6) and grafting it onto SM 2.29's Gecko in place of it's original engine (in effect, a transplant)? The old JS engines are inferior to the new ones, of course, but at least they run, and a running JS engine would provide a usable base onto which 9x-friendly implementations of newer JS features can be built (or at least give a good idea of how the newer JS engine could be rewritten). c

Tahoma is the default system font for Windows 2000 and XP; it gets installed with the unofficial Service Pack. Interesting that 2.29 works! What is the comparable Firefox version? Twenty-something? To fix the SVG font problem, you can do the following (these instructions are for Opera, but with a few adaptations, they work on Mozilla browsers too): Download the two fonts mentioned in this post: Go to about:config and make sure gfx.downloadable_fonts.enabled is set to false. Drop the fonts into %WINDIR%\Fonts Restart Firefox Enjoy the newly available font glyphs! 2.3x must be based on some version between 31 and 38. EDIT: Confirmed! SeaMonkey 2.3x is based on Gecko 33 and newer, so now we know that the effective limit of usability (so far) is Gecko 32, upon which SM 2.29 is based. We might be able to get Gecko 33 or even 34 to work with some careful tweaking, however. c

I noticed that as well, but I switching the system font to Tahoma (Western) seemed to fix it. This happened to me as well, but oddly, deleting the profiles folder, recreating a fresh one with FF 3.6 and going into 3.6's about:config to disable js solved this problem for me. c

I will! I shall report back here soon with my findings. c

Yeah, the point isn't necessarily that it's usable as is, but that it shows progress, and eventually, a working solution (plus maybe a fork of 24.x, such as, say Pale Moon 25 or 26, modified to be somewhat more 9x friendly) can be possible. I agree that broken js is a dealbreaker, but if the bulk of the UI and rendering engine works (as it appears to), then there's some hope that js can be made to work or mostly work, eventually. c

OK, let's see if I can help, since I'm already in this mode myself a bit :) It depends, but I think a Pentium 4 would offer somewhat better forward compatibility, with SSE2 and such. The AMD Athlon was a popular substitute during 1999-2004 because it tended to be a bit faster than most early P4s, but earlier versions lack SSE2, and later P4s eventually outsped them. Of the two, the Pentium 4 will be much more compatible with Windows 9x, particularly early ones from 2001-2003 or 2004. Stay away from Core2 Duos unless you plan to dual boot with Windows 2000 or XP, as some devices may not work properly due to a lack of driver support or a hardwired 9x-hostile configuration. You can do that, though you should install the RAM patch even with 512 MB to ensure stability. 9x should handle a DVD-ROM drive just fine. Worst case is it'll treat it as a CD-ROM. A 3.5 inch floppy is good, but make sure you have a motherboard with the proper interface (virtually all motherboards I know of from the P4 and C2D eras do (aside from some "legacy free" designs), so this shouldn't be an issue). This is where my knowledge gets iffy, as I've never tried installing/booting 9x from an SATA disk, but you should check out Rloew's SATA and AHCI patches, as I believe they are meant to make that possible. I can't say for sure, but I just read in another thread that someone recommended a GeForce 6600 GT, so that might be a good way to go? I'm currently using a Radeon 9550 in my P4 system, and it works OK too. I think the 6600 GT is supposed to be among the fastest-performing 9x-compatible cards, which I'd guess is good particularly for later games which can take advantage of it (I'm not a gamer, so I can't say for sure due to a lack of experience). Slot type will depend on your choice of motherboard, but I think you should try for an AGP card for maximum 9x compatibility (as far as I know, PCI express is a bit hit-or-miss). Plain PCI is okay too, but probably too slow for most newer games. Something like a Sound Blaster will probably offer the best compatibility sound-wise, and any standard ATX PSU ought to work as long as you make sure it's able to supply enough power for your chosen components. Setup /p i disables ACPI support, which can help ease some of the problems that can occur when trying to install 9x on newer hardware, such as a Core2 Duo-based system. I hope you find at least some of this useful! And anyone else please correct me if I've made any errors! c

I have devised a cheat, which I recall has been mentioned before. I am "running" roytam1's build of Firefox 45 ESR on Windows 98 by accessing a Windows XP VM on my MacBook Pro via Remote Desktop. Fonts are weird looking, and it's kinda slow and choppy, but it works! c