herbalist

Thanks. Can I ask you when you go here which version of Adobe Shockwave does it report you have currently? http://www.adobe.com/shockwave/welcome/ None. I didn't install shockwave. Don't recall ever using it.

If I remember right, I uninstalled version9, then installed 10. Since then, I've installed 10.1 and 10.2 versions over the existing version and haven't had any problems.

That is a definite possibility. A batch file that renames the hosts file and wipes the browser cache would make it easy to find out. In this thread I've been trying to determine what it is that requires KEX to be disabled on some PCs but not on others for the main executable of SSM. I'm suspecting that it's the PCs chipset and/or processor that ultimately decides this. If you have time, could you list those PCs , the chipset and processor each used, and which ones KEX works properly on? If you remember, could you specify what failed or didn't work on the others? Thanks

I see a few of those unresponsive script messages on some sites, mainly when I bypass Proxomitron. Haven't used SeaMonkey 2.1 enough yet to see if it has the same problem. Proxomitron also enables you to spoof the user agent as well as kill what it calls "nosey javascripts". Unlike browser extensions, it works with all browsers. Revolutions Pack helps with the appearance and performance of your PC, but doesn't affect the actual functioning of the browser. It also won't enable you to run newer browsers. KernelEx is pretty much the only option for using the newer browsers that are designed for NT systems. In spite of how well it's progressed and become stable, IMO it has to be regarded as experimental as it does not always behave the same on different systems. If you want to keep using 98 and not be limited to old browsers, by all means try it, but make a full system backup first. Then be prepared to experiment a bit.

Duffy98. I tried the version of KMplayer you posted the link for. On my primary PC, (98SELite) it did nothing at all. When I switched to an image with IE6 and all the updates for it, SSM and KMplayer conflicted badly. Almost every running process crashed. Media players do seem to be a problem. On mine, I use an old version of WinAmp for audio, WMP for video, and a freestanding flash player. Other than listening to music, I don't get a lot of use out of media players. Fortunately there are quite a few more to try.

Which version of KMPlayer are you using? I'm trying 3.0.0.1439 and can't get it to work, even with KernelEX. The error I get is: The KMPLAYER.EXE file is linked to missing export KERNEL32.DLL: GetVolumeNameForVolumeMountPointA

I just finished the first round of testing SSM compatibility with KernelEX on the hardware I had available. The first PC is a Compaq with Pentium3-866mhz, an i815/E/EP chipset and 128MB RAM. The OS is Lite 98SE (no Internet Explorer) with all the required drivers and the following upgrades: SESP2.1a RP9.7.2 Unofficial Shell32.DLL fix Unofficial 2-4GB Kernel32.dll fix Enable 48bitLBA fix for >137GB drives NUSB 3.3e Copy 2GB fix Ttf pak Additional software: TestRun 2.12 Kerio 2.1.5 7Zip SSM 2.0.8.583 When the configuring was done and the PC was running properly, I installed KernelEX 4.5final, enabled it for all applications, and rebooted. After restarting, attempting to open the Kerio status screen from the tray icon resulted in an illegal operation error message. invalid page fault in module <unknown> 0000:9eff03ae Attempting to open Kerio's help file from the tray menu gave this error: KDSE.DLL Exception Occurred (address 0x9EFF03AE, exception code: 0xC0000005)- firewall driver interface will be closed. If SSM is shut down and the start button is then pressed, the resulting error message is: Explorer caused an invalid page fault in module <unknown>0000:9eff049c These errors are all repeatable. If "disable KernelEx extensions" is selected for SysSafe.exe, the error messages disappear. The 2nd PC is my old workhorse, an HP Pavilion 4463, running 98FE. CPU-Z identifies its processor as an Intel Celeron , 366mhz. System information identifies it as a Pentium 2, MMX. I'd have to completely dismantle the unit to physically verify which it is. The chipset is an i440BX/ZX. The unit has 160MB RAM. The upgrades and software installs are far too numerous to list. It does not have Revolutions Pack and KEX wasn't installed prior to the testing. I installed KEX, selected enabled by default, and restarted the PC. The actions that triggered error messages on the 98SE Compaq unit caused the exact same error messages on the 98FE HP unit. On both of them, disabling KernelEx extensions for SysSafe.exe eliminated the errors. These are the same results I obtained on a virtual 98SE unit, built on Connectix VirtualPC 5.1 build 370. CPU-Z gives no chipset information when run on the virtual system. The processor information it displays appears to be that of the host PC. The only information I can obtain regarding what chipset VPC simulates comes from Belarc Advisor. It identifies the virtual PCI Bus Master IDE Controller as an 82371AB/EB. My present PC, a Dell Optiplex GX260 with a Pentium 4-2.4GHZ and 1GB RAM uses the i845G chipset. It has the same group of upgrades used on the test PCs except that it's had several versions of KEX installed. On this PC, there's no conflict between KEX and SSM. When I can, I'll swap drives in the Dell and build another 98SE test system using the same setup as the test units in order to see if the lack of a conflict on this PC is repeatable. At present, it appears that how well KEX and SSM get along on the default KEX settings is dependent on the chipset, or possibly other installed drivers. Except for drivers and some installed software on the primary PC, these units are equipped very similar. As soon as I can, I'll repeat the tests with all PCI cards removed and as few installed drivers and updates as possible.

In themselves, hooks are not good or malicious. They're a normal system function. Like any other system function, it's how they're used that decides if they're malicious. Linux enthusiasts are used to them. Most Windows users first learned of hooks from the commercial security-ware industry, which hyped all the potential malicious uses for them to scare the hell out of the average user, then promote their wares that can intercept this "new threat". They've been there all along but Windows users weren't aware of them. Compared to 98, XP uses a lot more of them. In over-simplified terms, hooks are a way to send/receive information from one thread to another or for one process to monitor/control another in specific ways. AntiViruses depend on them heavily. The main thing to watch with hooks is what process wants to set it. Malware has to execute before it can set hooks. When it does, it's a way for that malicious process to inject its code into a legitimate one. Explorer.exe is a common target. Iexplore.exe is another. Most of the prompts you'll see are legitimate system components hooking other system components. If the target files are legitimate and the process doing the hooking is too, they're probably necessary.

Such a file would have several problems. Where do you draw the line on what's included? Should it include IE6? WMP9 modified to install on 98? A complete single file would be huge, and very difficult for those who don't have high speed internet to download. Unless a lot of components are optional, a single file could end up installing items others don't need or want, like all the IE6 updates on a 98lite system that doesn't have IE. If users are given the option of selecting the items they want, what real advantage would this have over downloading the individual updates? The user still has to make the decisions. I don't see a single update that contains everything as a viable option. What would be useful IMO would be to combine a lot of the updates into component packages. A core component package could include NUSB, the shell update, the copy 2GB fix, the hard drive size limitation fix, 98kernel update, and non-IE updates. Another could include everything related to IE6, MDAC, etc. IMO, KernelEX and RP9 should remain separate for now as they are both being updated regularly. In the future when they're more or less finished, they could be combined into a 98 enhancement upgrade. To me, the updates look like they could be combined into 3 or 4 independent packages that could be used separately or in combination, depending on the users setup. As for the way it is right now, I definitely agree. It can be confusing. MDGX's site is definitely comprehensive, to the point of being an overload. A lot of the updates overlap and duplicate. It's hard to tell what is and isn't included in the packages, and at times even harder to determine which ones you need.

With KernelEX, newer browsers will run on 98. SeaMonkey 2.0 and 2.1 beta are both working well on 98 here, and both work with the new Yahoo mail. If KernelEX is more than you want to try, you might look into a user agent switching extension for FF. How well it works will depend on the individual sites and whether they're just reading the browsers headers or actually using a function that isn't supported in the older browser version.

SeaMonkey 2.1b1 is working fine here with KernelEX. So is Adobe Flash Player version 10,2,152,26 for non-IE browsers.

Sorry about the delay. Missed your last post. No, there are no special options you need to select as long as you're using the Win98 shell. Not sure about the 95 shell, don't have a copy of 95 to try it. I routinely remove all internet explorer components except for the protocol utilities. Toolbars still work fine. The most common issue that results from removing Internet Explorer is that it breaks some apps that use certain IE files. If an app no longer works, profile it thru Dependency Walker. It will tell you what missing file(s) that app needs. Any Internet Explorer files that an app needs can be put in that apps own folder.

Yes, you can remove Internet Explorer and still keep the quick launch toolbar, and any other toolbars you want to make.

I just installed 2.0.8.585 on the 98 system of my primary PC. This version is not stable on 98. Opening the SSM interface consistently causes a BSOD, after which SSM crashes. This is not KernelEx related. Version 585 will randomly crash on 98 whether KEX is present or not. Version 2.0.8.583 is the last stable version for 98. The links in my post at the beginning of this thread still work. Those are the last reliable versions. Rick

Prozactive, Thanks for posting the setup that you had when you tried SSM and KEX. If 585 crashes, the bug that appeared in 584 wasn't fixed. There's nothing in 585 that makes any significant improvement when used on 9X systems that I can see. On Virtual PC SSM and KEX had issues no matter which was installed first. Disabling KEX for the SSM executable seems to solve it. On a real Pentium 4 PC, SSM and KEX get along fine, at least they do on mine. I'm not certain of it, but I think Connectix Virtual PC simulates a Pentium II, but the speed rating seems to be inherited from the hosts processor. The 2.4 versions have no chance of running on 98, even with KernelEX. They were designed from the ground up for NT architecture. I'm in the process of rebuilding 2 other PCs for testing and will probably also temporarily switch over my old HP for this as well. One is a Compaq P3-866mhz, 128MB RAM The other is a Gateway P2-266mhz, 384MB RAM. The rest of the details I don't know yet. I'm going to put basic but fully updated 98 systems on both and see how well KEX and SSM get along on these units. The P3 is almost ready for an OS. The others will take a bit longer.

On the Virtual PC test units I've been working on, I've tried starting with KEX 4.5RC1, 4.5 final, and updating one over the other. I've also tried installing SSM before, after and in between KEX versions. Once I learned the easy way to reliably trigger the error messages, the results were the same on the virtual systems. Changing the order they're installed had no effect. On my Dell, (P4-2.4ghz, 1GB RAM) SSM was installed first, then KEX, which has been updated several times since. On this PC, they've never conflicted. At this time, I don't have another PC available to try. Based on this limited testing, the order they're installed appears to make no difference. Once they're both installed, the easiest way I've found to check if there's going to be a problem is to start SSM, then shut it down from the tray icon. After SSM shuts down, click on the start button. If they're conflicting, you'll see an error message and won't be able to shut down from the start menu. Will probably require a hard restart. If it does, restart the system, leaving SSM shut off. Go to the System Safety Monitor folder in Program files, and locate SysSafe.exe. Right click, select Propertiies>KernelEX, and select disable KernelEx extensions, then reboot. So far, this has worked on the limited testing I've done.

If the help file you found is ssmhelp-2.0.chm, it's part of the SSM install. Unfortunately, it's very incomplete and does not go into any details about how to use SSM to enforce a security policy. The link to the outpost firewall forum describes global settings for the pro version (2.4.x), which doesn't run on 9X. The 98 compatible version doesn't have the command line parameter checking ability described in that post for RUNDLL32 (or any other executable). It doesn't have options for rule "groups" or the network access permission options. These are some of many details I'm trying to address in the tutorial. Yesterday, I finally managed to reproduce the conflict described here and here on a virtual 98SE unit. The virtual unit uses the same updates and file versions as my primary physical unit, which does not have this problem. I have not yet determined why it's happening, but am suspecting that this could be dependent on the processor and/or chipset the system uses. I don't have the hardware variety I need to explore this possibility. If anyone else has run into this problem, could you post what processor and chipset the system contains? Thanks in advance. Rick

It would seem that I'm getting senile, or at the very least, completely scattered! Version 585 is completely legitimate. It's another one of those items that I wanted to get around to testing, then completely forgot that it existed. I'll test it and see what's changed as soon as I can. I apologize for the my confusion.

I apologize for how long this is taking me to do. It's hard to find any free time for this and several other projects I've wanted to do. If at all possible, I will try to finish it this winter. Regarding version 2.0.8.585, I beta tested SSM all through its development and have never seen that version. The last free version I was aware of was 584, which was withdrawn for stability issues. So far, the download links I've tried to that version are dead. If I find a live link, I'll check into it and see what's changed, and verify that it actually is a newer version of SSM free.

I've run into a bug with RP9 themes and the WinHTTrack website copier. If any skin besides classic is selected in RP9 and the user selects either language preference or the base path browse button, WinHTTrack locks up. With the classic theme selected, it works fine. I've reproduced this with versions 3.40-2 and 3.43.12 on both my 98SE lite system with IE removed and on a 98SE test system with IE6 and most of the unofficial upgrades installed. The results are the same. I don't know how long this problem has existed. It's been a long time since I last used it. While their website says it's for Win 2000 and up, WinHTTrack is working very well on 98 with KEX and RP9, except for this one problem.

Page link in above post fixed. Added direct links to the standalone executables.

Have you tried Adobe's standalone flash player (projector)? Version 9,0,289,0 runs fine on 98 and is 3.27MB. Version 10,1,102,64 is 5.30MB and it appears that it may work with KEX set to 2000SP4 or XP-SP2 compatibility. I just downloaded this version and have only tried it with a couple of local files, which played fine. The browse menu isn't displaying names on my setup yet. Might be due to my font selections or RP9 settings. The files are the projectors/players themselves, not installers. Both are available at http://www.adobe.com/support/flashplayer/downloads.html Might be worth a try. Rick edit. Version 10,1,102,64 will play flash videos from web links on 98 with KEX set to 2000SP4 compatibility. Full screen mode also works, at least it is for me. edit. link above fixed Direct link to standalone version 10 Direct link to standalone version 9

That is completely unexplored territory. The answer will largely depend on what type of malware it is. For kernel rootkits that target NT systems, KernelEX won't make 98 any more vulnerable to them. For more conventional "user mode" adware, nuisance-ware, etc, KernelEX might make a system more vulnerable. KernelEx might also make 98 more vulnerable to malware that targets applications, just by enabling 98 to run these applications. The only way to know for certain would be to collect samples and try them on test units. I wouldn't expect any major increase in the amount of malware that would affect 98 due to KernelEX, but I would expect a fairly small percentage of some types to work. It's just something that we will have to watch and remain aware of the possibility.

Is this what you mean by autostart scheduler? Startup Delayer.

I'm not sure what you mean by "interactive" firewall. On win2K, I'm still using Kerio 2.1.5. It's a straight rule based firewall that's no longer supported but runs quite well on 98 thru XP. On the older operating systems, at times you have to use older software. This older software is often unsupported, but that by no means translates into weak or vulnerable. A lot of "for XP" firewalls should run on 2K, if not the present version, then the previous one. "Firewall" is a very broad term, encompassing everything from simple packet filters to comprehensive security suites. Could you be a bit more specific regarding what you need from a firewall?