Prozactive

From the lack of responses, I'm obviously not the only one baffled by this mysterious problem. I wanted to report that after a lot of effort and work, I finally was able to successfully resolve this error. Hopefully this will help anyone who happens to encounter a similar error in the future. I did extensive Googling on the specific page fault error I referenced in my original post, and while I did not find any association with WinImage, I did find numerous instances where a wide range of other programs and applications caused an identical page fault error in KRNL386.EXE at 0002:00005c83, with the exact same Bytes at CS:EIP. Unfortunately the vast majority of these were just reports of the problem without any specific solution. I did, however, come upon two cases where the error was successfully resolved - one had to do with some specific registry value and the other involved the location of an .ini file. WinImage did not have any associated .ini file, but I did install and start using RegCompact to compact my system registry (in the period between the last successful use of WinImage and the first occurrence of the page fault error) after reading the excellent thorough discussion of that utility by CharlotteTheHarlot. I happen to maintain an extensive weekly archive of system registry .cab backups, so I restored a registry backup made just before my first use of RegCompact and lo and behold, WinImage loaded successfully again! My first thought, naturally, was that RegCompact somehow caused some subtle error(s) in my system registry, but I decided to do some more extensive detailed testing. It turned out that RegCompact was not the cause of this error, as subsequent compacted registries with RegCompact also allowed successful loading of WinImage. I finally was able to isolate the error to a specific time period (between registry backups), but I did not make any OS system changes then and all of the registries test 100% error-free. So, I'm still at a loss as to the specific cause of this error. I could do more specific "before and after" testing of the registries using a file compare utility and I probably eventually will. One other thing I did note was that now the year in Date/Time Properties has reverted back to its former 4-digit format (i.e., "2013" vs. "13"). The year format had changed to just 2 digits sometime recently and I had no idea how/why it happened nor how to change it back, as Googling it drew a complete blank. I don't know if it's related to the KRNL386 page fault error or not.

I have WinImage 8.1 installed on my system and periodically use it to update and create various boot disk images. Recently I suddenly started getting a consistent page fault error in KRNL386.EXE upon attempting to load WinImage. This baffled me because I had made no OS changes since the last time I successfully loaded and used WinImage a few months earlier. The only system changes were routine updates of my antivirus and anti-malware definitions. I performed numerous hardware diagnostics and have confirmed that my memory and HDD are error-free and working properly. I also decided to try disabling KernelEx extensions for WinImage system files without any effect (not surprising, as WinImage worked perfectly previously with the default KernelEx settings enabled). The last thing I tried was reinstalling then uninstalling WinImage but I found I still got the exact same page fault error during those operations. Web searches draw a complete blank.. there doesn't seem to be any information about WinImage and KRNL386.EXE errors. I'd appreciate any thoughts and advice on this puzzling problem. I should note that I do have the updated KRNL386.EXE v.4.10.2000 installed by 98krnlup.exe (Unofficial Krnl386.exe stack corruption fix) from MDGx. As reference, here is the error that I get: (the Registers: and Stack dump: data change but the specific memory address and Bytes at CS:EIP: are exactly the same every time) WINIMAGE caused an invalid page fault in module KRNL386.EXE at 0002:00005c83. Registers: EAX=12af0000 CS=015f EIP=00005c83 EFLGS=00000202 EBX=00000004 SS=3317 ESP=00007f66 EBP=00007f6a ECX=00000000 DS=016f ESI=000047ff FS=0000 EDX=00000000 ES=475f EDI=00000000 GS=0000 Bytes at CS:EIP: 26 88 05 f7 d9 03 cf 58 5f c9 c3 55 8b ec 57 33 Stack dump: 00000000 59627f78 017f47ff 00000000 7f940000 0002516f 47770000 475f0000 47c70000 12af0000 017f14de 7fec47ff 01af0028 00000002 00004777 0000475f

I also switched to Opera 12.02 a while back courtesy of Dave-H via another forum. So far it's been working well overall despite some formatting bugs, especially a black background on the navigation/URL and some other tabs that make it extremely difficult or impossible to view the buttons/options (using the Windows Native skin). Searches of the Opera forums show this is a common complaint without any solution, unfortunately. In recent weeks I have noticed progressively increased load times for Opera, which is becoming extremely annoying. It now takes around 3 mins for the Opera task to show up on the Windows taskbar and around 5+ mins before it fully loads with the bookmarks toolbar, etc. I have no idea why this is occurring. Also, occasionally the Opera window will freeze and become unresponsive, and then not redisplay after minimizing/restoring. The only way to solve this problem is to exit and reload Opera. Not to be the Thread Police, but it would be nice if these recent Opera-related posts could be moved to the long extensive Opera/KernelEx thread that dencorso and I spent a great deal of effort a while back creating.

I have not posted in a long time but I still regularly check and keep up with this forum. I just wanted to say I totally agree with bphlpt, and I appreciate his long thoughtful assessment of the situation. I am shocked and distressed to find this thread devolving into a flame war. While I still have not had the opportunity to use the SP, I am extremely impressed by PROBLEMCHYLD's initiative in taking up this project, and all the hard work he's put into it ever since. Like bphlpt, jaclaz, et al., I also thought that jds made a very reasonable request but it is certainly PROBLEMCHYLD's prerogative whether to implement it in the SP. As I said, I agree with all of the other points made by bphlpt and I hope we can stop this flame war and continue to support PROBLEMCHYLD in his work. The Win9x community is small (and getting smaller) and we need all the help and support we can get.

Thanks for the great info as always, dencorso. I could not locate a changelog for XXCOPY but I knew you had some valid justification for your statement that v.2.96.5 was the last version compatible with Win9x/ME. The main XXCOPY website is very sparse in terms of information and apparently most of the valuable data is contained within those numerous technical bulletins, which are not very user-friendly and convenient to navigate. I would've been happy with v.2.96.5 until I found that it didn't support the "Wild-Wildcard" exclusion feature. That's when I decided to try the latest v.3.11.2 since as I said, the website states "Supports all Windows versions (except Windows CE)". So far in my limited usage it seems to work fine, but the lack of support for short (8.3) filename aliases is quite annoying since it's a real pain to have to type the entire LFN when specifying files. I may revert back to v.2.96.5 in Win98 SE and use v.3.11.2 in WinXP like you.

Resurrecting this thread from the dusty vaults... I recently installed and started using XXCOPY per dencorso's strong recommendation, and as he stated, I am very impressed with the utility and its powerful capabilities. I initially installed v.2.96.5 since he indicated it was the last version that still works in Win9x/ME, but later I decided to give the latest version v.3.11.2 a try since the (meager) documentation indicated it works in all versions of Windows. I haven't used it that extensively, but so far it seems to work fine in Win98 SE. The only quirk I've noticed so far is that it doesn't seem to handle the OS-generated short filename (8.3) alias... you have to specify the full LFN (long filename) in v.3.11.2 whereas v.2.96.5 seems to handle the short name alias without any problems.

It is extremely unlikely that newer versions of Avast! will work with Win9x/ME and KernelEx. I believe that even Win2000 isn't supported anymore. Of course, as a definitive test, you could always try it and see. Let us know your results. And yes, support for Win9x/ME has actually been extended considerably from the original deadline for Avast! 4.8, for which I am grateful.

This belongs in the main Opera/KernelEx thread. Maybe dencorso could move it there. Unfortunately I don't have any solutions to your problems but hopefully someone else will.

Thanks. It certainly wasn't clear when I browsed their website as they gave contradictory information on different webpages. And I didn't see any mention of the ANSI version.

@jds: That is indeed strange. In all the years I've used Ghost I've never had to use the -FNI switch. My first guess was that you have one of those new Advanced Format HDDs but a quick search showed that on the contrary, you have an older WDC Caviar HDD. It would be interesting to see if switching the HDD to the primary IDE port makes any difference. I also wonder if there's some sort of quirky disk geometry setup misconfiguration or related type thing going on. Web searches on the -FNI switch show that it's often associated with SATA HDDs but I'm pretty sure that doesn't apply in this case.

It's unclear whether Audacity 2.0 still works in Win98/ME. They give conflicting information on their website. As I reported earlier, 1.3.14 beta resumed Win98/ME support. Anyone know if this is still the case? I probably will test it out soon and I'll report back on the results. Always appreciate your heads-ups on software updates, halohalo.

Well, the setup installer overwrote and replaced several files I didn't want to be overwritten, eliminated my Windows Native skin preference, and also somehow changed my default system font. After a lot of work, I subsequently figured out how to restore the Windows Native skin but it was extremely aggravating and frustrating. I will always use the manual extract and replace method of upgrading Opera in the future as it gives me full control over which files I want updated. And yes I did select Win2000 SP4 compatibility mode in KernelEx.

Yeah Opera still has a fair number of annoying bugs. I never did figure out a solution to the missing "Find in Page" bug I reported earlier and finally submitted a bug report. Unfortunately it's still not fixed in 11.61. I just installed 11.61 and so far it seems to be working well. I tried running the setup .exe for the first time and that was a disaster. I finally went back to my tried and true method of manually extracting and replacing all files with 7-Zip.

Opera 11.61 was just released.

Like Dave-H, I was saddened to read about your personal situation in another thread and I could definitely relate, having gone through many similar events in my life (like so many of us). I too am deeply appreciative of your work with the Service Pack and commend you for it, even though I haven't personally had a need for it yet, having piecemeal patched my Win98 SE systems over the years. That may change soon if I decide to do a new fresh install of Win98 SE in the future. Keep up the great work and best wishes for everything.

I finally was able to get a screenshot of the duplicate CD drives in Windows Explorer on my laptop. I thought the duplicate drives error/bug mostly occurred on my desktop PC but it obviously also occurs on my laptop. Clicking either entry for the Microsoft01 E: drive allows access to the CD-ROM but one of the entries still remains after ejecting the disc.

I just found out that Audacity resumed support for Win98/ME starting with v.1.3.12 beta. Previously, 1.3.7 beta was the last version that supported Win98/ME and I was (pleasantly) shocked to find that they resumed support again. This may be the first time I've ever seen this happen. The latest version is 1.3.14 beta. EDIT: I think it would be nice if we expressed our appreciation to the developers for resuming Win98/ME support as this so rarely happens. I never expected it but was surprised to see that it resumed as I was browsing the Audacity site. BTW I don't know how to update the master list at the beginning of this thread to indicate that development is now ongoing.

Thanks rloew. I just checked and I was wrong. Auto Insert Notification is turned on for both of my optical drives (DVD and CD). I thought I had disabled AIN a long time ago but I was mistaken.

My duplicate CD/DVD drives bug looks identical to LoneCrusader's. I have Auto Insert Notification turned off on my drives. I believe clicking either entry will allow access to the drive but one entry still remains after ejecting the disc, and that "phantom" drive gives an error when clicking it (obviously).

Thanks dencorso! Of course you can remove my previous post.. actually I was kinda kidding after your massive thread surgery on the new Opera/KernelEx thread. And thanks as well, LoneCrusader. I think I understand what you said but I need to go back and reread the previous discussions in the IO.SYS thread as my first impression was that perhaps a few additional posts should be moved here. I think this new EXPLORER.EXE thread is great, as it consolidates a good detailed separate technical discussion of a new topic.

@dw2108: I'm using Opera 11.60 FINAL. I presume CharlesF is also. I still have not resolved this Find in Page search issue after more troubleshooting.

Wow. Has this thread ever gone WAYYY off-topic. Sorry for mentioning my problems with Windows Explorer, dencorso. Feel like doing some more thread surgery?

Thanks CharlesF. Yes I'm familiar with the many alternative ways of searching in Opera. It's somewhat difficult to explain this bug but Opera 11.60 no longer shows my previous (saved) Find in Page searches in the address field as I type the string 'f <space>'. Opera 10.63 would automatically display a list of previous (saved) Find in Page searches (in the format "f xxxxxx" where xxxxxx is the search term) in the pulldown list whereupon I could select the one I wanted. I finally was able to restore this functionality to Opera 11.5x by fiddling around with the search.ini file but I am unable to do so with Opera 11.60. I'm starting to suspect it's a deliberate redesign feature of their new address field. I've searched the Opera forums without much luck although there are tons of complaints about similar search issues.

Thanks for the info. I updated to Opera 11.60 and so far everything seems to work well, except I have the bug again where my past "Find in Page" searches no longer show up in the address field (in the format "f xxxxxx" where xxxxxx is the search term). This also occurred after I updated from Opera 10.63 to 11.51. After a lot of effort, I finally was able to restore the "Find in Page" searches by fiddling around with search.ini. It had somehow gotten extensively modified and I replaced it with the former full version which somehow fixed the problem. Unfortunately, that no longer works for 11.60 and I'm not sure why or what to do now. Also, the PgDn/PgUp scroll bug I previously mentioned is still there. I thought at first it was fixed but alas, that's not the case. They did finally fix the Wand password missing cursor bug though.

Sorry I've been extremely busy and sidetracked by lots of other issues. Thanks for all of the replies and info. Yes, I have Windows Explorer 4.72.3612.1710 with 256 colors system tray icons that was installed by MDGx's Explor98.exe. It replaced Windows Explorer 4.72.3612.1700 that was installed by NUSB33, which in turn replaced the original Windows 98 SE Windows Explorer 4.72.3110.1. And yes, I've also installed MDGx's Iosys98.exe update.