Drugwash

Great news, thanks for the feedback! I'll take a closer look at the versioning system, for now I just bumped the original version up a notch in \common\version.h without taking into account any similar changes made in leyok's builds, mainly because that project failed to deliver a working final version. But I'll probably leave this to jumper once he settles down all tools. Have you tried uninstalling your current version and then installing the whole 4.5.3.1 package? Do you have any entries in KstubXXX.ini that may clash with KernelEx's own functions? Try disabling Kstub temporarily.

And if you get to that, gently remove the cover from the CPU socket and with a magnifying glass look at the contacts; some may have been loosened and need to be stuck back together using a tiny needle or something. I had to do that for a board given by a friend, where the CPU had been forcibly pulled out and half of the contacts were badly tilted. It's now working as a test machine for a KernelEx installation under 98SE.

You may also want to consider checking the hard drive for bad sectors (using HDAT2 or other similar tool). Also try to reseat all cables (at both ends), check out any jumpers on the motherboard making sure they're firmly inserted and also make sure there's no shortcircuit between the motherboard electronic parts/tracks and any metallic parts of the case or peripherals (optical drive, HDD etc.). All that just to rule out possible hardware issues, considering that - as you said - the only change was the case.

Hello, I tested the new kernelex 4.5.3.1 on one virtual machine and on a real machine, and in both cases I have not noticed any problems. Besides kernelex of VM I installed the SP3.36 complete, IE6 and the Revolution Pack 9.7.2. On the real machine in addition to Kernelex I installed the SP.3.36 based, RP 9.7.2, IE6 and 98SE2ME. Thanks for the work! Mcv'93 Thank you very much, first feedback and a positive one! Most thanks go to author Xeno86 and then jumper who added a few needed things (so far) - I just compiled and packed. Well, that's something I'd like to know too, in order to test the build(s). Personally I never was desperate to try such new applications meant for XP+ on my main machine so I wouldn't know about these applications and their requirements.

Well, someone inside M$ should know, not speculate. But then again we, from the outside, might see things as they are, not as someone - even on the inside - would be told to. As for the burial, maybe M$ are afraid of ghosts... because sometimes they rise and haunt the bad people...

Well, how 's the Time market, any changes in price? I've managed to install 98SE on a test machine, plus DirectX 9.0c and drivers. Installed my build of KernelEx with all fixes in (4.5.3.1 mentioned above) and all went fine. Installed and (shortly) tested Firefox 9.0.1 in XP-SP2 compatibility - seems to work for now (I added msvcr70.dll v7.0.9466.0 renamed as msvcrt.dll in Firefox's folder). I've also installed CloudMe 1.8.8 and ran it. It does work without the UPX "fix", which means the GCC fix in KernelEx does its job, but it can't log in. However, the UPXed files from my main working machine don't work either so it must be something related to the test machine. UPDATE: After installing a bunch of other (unrelated?) components (IE6, DSClient etc.) I remembered a Root Certificate update was mandatory for CloudMe to work; as soon as I installed that update, it logged in. So far nobody else offered feedback on my KernelEx package so I'm not sure my tests are conclusive.

Well, that might just work! :lol: Go, Symantec!

Just to be sure, pull the RAM sticks out, take a magnifying glass and look carefully at the contacts on both sides, make sure there's no sign of degradation. I've had such stick that used to throw BSODs every now and then and the only way to "fix" the issue was to put my hand inside the case and move parts, pushing into sockets. When I pulled that thing out and looked at it, some of the golden contacts were worn out. For best results, take a pencil rubber and rub the contacts on both sides until they shine. That way you'll notice more quickly any possible defects and it'll also make better contact in the socket. A MemTest86 boot disk would also come in handy. I'd recommend something that has this among other tools but I'm afraid it's considered *ahem* inappropriate here.

I knew you would. And let me be Cassandra for a second: I bet in - say - twenty-thirty years from now, somebody at M$ will officialy confirm my opinion. Unless, of course, M$ just dissapears long before that.

They always get away with it, virtually nobody makes a case out of this. M$ are really desperate to bury Win9x and everything related to it. Think for a second: why is there no Windows 9? Why did they jump from 8.x to 10? My opinion is, they are afraid people looking for Win9-related stuff might stumble into Win9x-related stuff and be amazed on all the available options 9x had , compared to the new "monsters". What were you saying? Oh, the drivers... OK, I got them saved. Hopefully there won't be another system crash before getting to burn these on DVD. Although last night's repeated power failures did break a PSU in a friend's computer I had here for temporary perusal. Off-topic, I know...

Please use another host. That one doesn't allow download managers and with my 14kB/s flaky connection I can't waste time and nerves. Try a cloud storage such as CloudMe (I'm using it, see my signature).

Too bad they don't provide a change log for the different driver versions. If I were Dave, I'd keep the older (on CD) driver somewhere safe, in case there's any incompatibility between the new drivers and the hardware or other drivers. Anyway, thanks for the heads up, jaclaz!

Well then I'd be interested in those on the CD.

Thanks for the official link - if the drivers on the web are the same as those on the disk (for ST300) then we can just get them online while they're available. I have a fairly old motherboard for my 98SE installation and still got weird issues with the optical drive, also related to networking. At times, when trying to access the optical drive - located on the 98SE machine - from another machine (it's shared on my local network), the 98SE machine just instantly reboots. Not always, just sometimes. Since that is the only (usable) DVD unit in the house, you can imagine it's quite frustrating. Reasons could be too many to count so I have not pursued the issue. Yours could well be similar, especially if you have a similarly updated 98SE system as mine (2000/XP/2003 system files, AutoPatcher/uSP, KernelEx, Revolutions Pack). Maybe we should open a new topic for such optical drive issues sometime. You'd be amused to know that, rarely, my system throws an error messagebox related to... blank screensaver! I mean, in display properties I have selected 'Blank' for the screensaver which should just blank the screen. Well, at times when that blank should come into effect according to the timer, the screen does go blank but I get the crash message box (twice in a row). However, just tapping 'Enter' each time restores the desktop without any side effect. The video driver is extremely old and specific to the videocard (MSI GeForce2 Ti4200 AGP8x). If I try a newer, generic driver from nVidia (even the tweaked one from MDGx) I lose AGP Texture in DirectX diagnostic page (and obviously in applications). So, you see, weird things are happening in this realm. We cope with them as we can.

Well, it was to be expected. Does that mean there's nothing more to add/change/update to the package or is it that you just lost interest in this particular project? Just asking in case someone else would want to take over, if needed. Thank you for all your work on it! Even so, they've still got many pearls, you know... Undoubtedly they do. For how long and how much is still there, that I wouldn't know. I see mostly German files on that page. They could be of use for a German service pack. I wonder if the English ones are already integrated in AutoPatcher or uSP3 and which ones would have been superseded by Win2000/XP/etc versions ported to 9x...

I've released version 4.5.3.1 of the KernelEx installer with the speed optimizations and all other previously omitted flags. There's also a separate package containing the modified files. I've slightly changed some project configurations so that absolutely all of the subprojects (except for sdbcreate, under Win9x) could compile in a single batch operation. Please test only in a virtual machine. Happy New Year everybody! Download here

Happy New Year, Dave! Glad you got everything working! Good news the eSATA card has 9x drivers, maybe you could share them in case we need them for a driver package (there's such an idea somewhere around). Good luck ahead!

OK, thanks, I'll fix those flags then. For the installer I use NSIS 2.46, I hope there's no major problem with it. Haven't got any reply yet in regard to my build, whether it installs and works correctly or not.

Got it, thank you.

Thank you, that's very kind of you but please keep in mind my connection is at 14kB/s currently, as is most of the time after depleting the full-speed (500kB/s) 5GB monthly quota (cheap subscription, GPRS modem only). I've stumbled into so many sites that provide expiring links where my slow connection is simply unable to download anything from unless it's under 1MB of size. That's why I specifically asked for the necessary files only because 160+MB would take forever and may lead to too much frustration than I could take, especially these days. I downloaded the .msu in your first link, will see how it goes with that one. Thanks again. (typos...)

Yes, early in the morning before going to sleep (no mistake here, it was past 6:30 AM) I managed to wrap a couple functions and tested the file I mentioned in relation to that tool I found. GetDeltaInfoA() returned correct data, same as the other tool. Hopefully all other functions would work too. I'll work on the wrapper and a GUI for it whenever possible. However I do need a few delta files (with their source files) to test decompression, since I have no Vista+ around to grab compressed files from. Oh and I got your confirmation in the mail, thank you.

That would take a few changes in the code. I'll consider it. Thanks for the idea!

I'll try my best. Kinda short time now with the holidays and all. Will see how it goes. EDIT: I did find a few changes related to kexbasen in the .dsp project file and in the .rc resource file, apart from newly included files. It's about flags in the Release configuration: /O1 instead of /O2 and an added /FR (in .dsp) and FILEFLAGS 0x21L vs 0x1L and 0x20L vs 0x0L respectively (in .rc). /O1 instead of /O2 was also found in kexbases.dsp along with added SUBTRACT CPP /Fr and same with FILEFLAGS mentioned above in kexbases.rc. Same /O1 found in core.dsp too and also same changes in FILEFLAGS in core.rc. Honestly I'm not familiar with compiler flags and all that so if you could briefly explain what they do and if the changes are intended or not, I'd be grateful. Just making sure everything's as it should be. EDIT2: I've been getting a longstanding warning C4163 at line 239 compiling jemalloc.c in kexbasen: _BitScanForward not available as an intrinsic function Do you get that too or do you know how to fix it? I'm using VC6 on updated 98SE. New one is warning C4273 at line 53 in psapi.c compiling auxiliary: GetProcessId inconsistent dll linkage. dllexport assumed. Also kexcontrol.exe won't build neither in Release nor in Debug mode. Five unresolved external symbols. Can't remember if that happened before. Everything else compiled succesfully after applying your changes (except for the compiler flags and file flags mentioned above, until I get details from you). EDIT3: Nevermind about kexcontrol, I just read COMPILE.TXT and found the reason, now it's OK. I uploaded a newly built installer at my repository in the KernelEx folder. Whoever got the courage and means, please test it. I currently have no working 9x test machine.

Thank you, got it and replied. Curious it wouldn't work by file name, I'm using exactly those routines in my tool since the handle-based ones are still incomplete in the wrapper and would just introduce unnecessary overhead obtaining and discarding handles. Maybe the flags weren't right? I'm using 0x3 (APPLY_OPTION_FAIL_IF_EXACT | APPLY_OPTION_FAIL_IF_CLOSE). In previous tests when I first built the tool, the Test routines and/or flags wouldn't work for me but I haven't rechecked them now after fixing a few bugs. @ den: Thank you for testing. I already fixed the progressbar option (which is disabled in current version due to some bugs) and will probably release a new version after I check a few more things. No worries about the e-mail address.

Can't see any walls... I guess that guy throwing the bricks took care of them. OK, back on topic before some brick hits us both.