jumper

I suggest downloading from the above links Lenovo's Win98/SE drivers for the 2100 3B and installing it first. You will probably have to select it manually for your card and confirm that you want to use it even though it wasn't written specifically for your hardware. If you can get it to work and are satisfied with 'B'-only speed, you can stop here. The next step is to download Lenovo's Win2K drivers from the above links for the 2100 3B and 2200 BG. Compare the Win2K files (especially inside the *.inf) for the two products to see what drivers changes were made to upgrade from 2100 3B to 2200 BG. Then try to make those same changes to the Win98/SE 2100 3B drivers to create a new Win98/SE 2200 BG driver. If you can create a working driver that did not previous exist, it would be a giant leap forward in our efforts to learn how to create Win9x driver for many modern devices that we can't currently use in Win9x systems!

Check the cabling from the card to the front panel--you could have a floating ground. Is it connected to the correct output (line-out or speaker) ? Try disconnecting the front panel from the card and testing the line-out and speaker outputs directly off the card. Headphones are usually higher impedance than speakers. Check the manual to see what the card supports (without the front panel headphone amplifier). Make sure you test all outputs with appropriate impedance-matched playback devices. The on-card and/or front-panel amps for the headphone and speaker outputs could be blown. The line-outs from the card to self-powered (amplified) speakers or a hi-fi system should provide the best sound quality. The external amplification can then deliver all the volume you desire.

Have you checked the mixer settings for individual levels on each different audio source? You've only mentioned setting the overall volume via various controls.

According to MSDN, GetPerAdapterInfo doesn't set the last error via SetLastError(). So without the "e50" that should be: [IpHlpApi.dll] GetPerAdapterInfo=r50s3 ; Error_Not_Supported, 3 Params

Apologies for not posting these 6-day-old results sooner.... I downloaded Opera_1210_en_Setup.exe and followed these steps: Used 7-zip to extract all files into E:\Opera_1210_en_Setup Set KernelEx mode on opera.exe to Win2K Added GetAdaptersAddresses definition to kstub822.ini: [iphlpapi.dll] GetAdaptersAddresses=t5 [*]Confirmed registry had been backed-up that day. [*]Ran opera.exe What happened: Opera 12.10 loaded successfully and began to self-install A new profile was created in E:\Opera\Opera_1210_en_Setup: 12 folders, 21 empty or stub files, 45KB total (82KB used) After about 30 seconds, Opera hung without displaying a window No error dialog or log file, even after forced termination Side effects: Opera 10.63 would not work until saved registry was restored System now back to normal After researching GetAdaptersAddresses, I'm not convinced it is really needed for Internet browsing; P2P and VOIP, probably (I don't know what all features/protocols Opera 12 has tried to build-in support for). But I do have a ReactOS-based solution in the works for this function. NOTE: Do not try to use iphlpapi.dll from 2K or XP--that doesn't work! Uxtheme support might have been the real show-stopper for me, but we may already have a solution for it that jds discovered: win2000-uxthemedll Conclusions: Use Kexstubs definition for GetAdaptersAddresses instead of replacing iphlpapi.dll Use uxtheme.dll from win2000-uxthemedll.zip instead of KernelEx version @Giant2011: please edit your long posts with

Here's an updated link to Delta Compression Application Programming Interfaces

In this ossbuild respository, I did find some old win32 binaries in the deprecated GStreamerWinBuild-0.10.4.exe. Some patching would be needed to see if they could work without KernelEx.

As usual, I'm browsing in Firefox 2.0.0.20 right now. I occasionally use Opera 10.63, IE5.0, or Off By One. Your CD-ROM drive could be dead. Test it in another system, use a different cable, use a different power connector. If the drive tests good, get it working in DOS, then we'll tackle Windows.

Yes, I think so. On my Win98se machine, Mspatcha.dll exists in the System folder. It is an XP version, but Depends reports no problems. It should work with RUNDLL32.

>Interesting that the subsystem change helps only partially. It might be possible to replace to an older version of gstreamer, but on a quick search I couldn't find binaries. It seems to have been Linux-only until recently. Source code is available going back ten years, so we might be able to port it to (or recompile it for) Win9x. >Can KernelEx report what functions of it were used at a given time? The standard version cannot, but Kexstubs can. I'm working a way to redirect more KernelEx support through Kexstubs for easier debugging without requiring a special version of KernelEx. >What does msimg32.dll help with, transparency? Is it a simple DLL replace or does it require KernelEx? I think so. Opera usually prompts Win9x users to upgrade msimg32.dll to the WinMe version. Yes, it's just a simple DLL replace; or KernelEx will provide an updated version. Either way will work. @Foxbat: Win98se comes with version 5.00.1693.1 of Msimg32.dll. In my experience, the WinMe version fixes nothing but the warning message--transparencies still don't work. Perhaps my various video adapters have been the limiting factor....

Open Watcom looks like the winner! I produced a "final" at about the same time this afternoon using VC6. It's functionally identical to yours, but I was unhappy with the 19KB of bloat VC6 added. Your "final" is actually 160 bytes smaller than two days ago and looks great! Well done. @PC: I count 14 dialogs with "New Hardware" in the caption. Which are you referring to and how can we reproduce it?

>opera.dll directly dependent on uxtheme.dll KernelEx already contains a uxtheme.dll stub module, but if Opera 12.10 needs it to be functioning in order to display the UI, ...hmm.... >Time to install Kexstub and see if it can do something but I fear that the missing getadapteraddress function in iphlpapi.dll has to actually exist to be able to use the network. I'll look for a portable version in the morning and see if I can get it past the system loader. I've had mixed success getting modules from other projects to work with Kex via Kexstubs. I not overly optimistic about iphlpapi.dll in the short term, either, but we'll see what we can do.... http://www.opera.com/browser/download/requirements

Temporary attachment: - SYSDM.CPL I went back to your previous version, rebuilt the problem icon so the resolutions were in the correct order, then changed the flag and the version strings. (I didn't use your last version because excessive movement in the Troubleshooting tab made it further rather than closer to fully adjusted.) Drugwash was right about only needing to remove just the WS_VISIBLE style. We have Microsoft to thank for the non-standard icon that was incompatible with MS tools, and for neglecting to extend the sheet sizes for the first four tabs when they extended the group boxes--it wasn't my fault after all! . . .

Not quite enough, I'd say! I see what you mean about the groupboxes in the other four tabs needing to match. I've never worked with tabbed controls before and didn't realize that changing the size of one sheet could possibly affect the other tabbed sheets. The editor must know to tie them together. Some of the drop-downs also need to be shortened. And I see the odd cd-rom icon, too--it looks fine in the Developer Studio preview, however.... I'll look in the exported rc source and compare parameters to what the other icons use. I'll think I'll start over and just do the minimum as you wizely suggest.

This very MSFN page we are reading includes: "http://connect.facebook.net/en_US/all.js". So, YES links to Facebook use javascript!

MSVC++ 5 was able to edit the NE file. I learned the hard way that Save crashes, but Save As works. I was able to disable the Visible and Tabbed options on "Disable System &Restore." and shorten the group box and dialog box to match the 98se version. I also bumped the version to .3002 (in two places). The only down-side seems to be that VC5 can only edit by adding a modded resource without removing the original data, so the file is 7KB larger. Temporary attachment: - SYSDM.CPL Edit:I was also able to save as .rc and .res: Sysdm.rc.res.7z - sysdm.rc, sysdm.res Here's the most relevant part of the rc code: 1742 DIALOG DISCARDABLE 0, 0, 256, 133 STYLE DS_MODALFRAME | DS_3DLOOK | WS_POPUP | WS_VISIBLE | WS_CAPTION | WS_SYSMENU CAPTION "Troubleshooting" FONT 8, "MS Sans Serif" BEGIN ICON 605,1736,7,7,21,20,WS_GROUP LTEXT "It is recommended that only advanced users and system administrators change these settings.", 1737,36,7,193,19 GROUPBOX "Settings",1738,7,33,241,92 CONTROL "Disable new file &sharing and locking semantics.",1750, "Button",BS_AUTOCHECKBOX | WS_TABSTOP,24,48,204,10 CONTROL "Disable long &name preservation for old programs.",1749, "Button",BS_AUTOCHECKBOX | WS_TABSTOP,24,60,204,10 CONTROL "Disable &protected-mode hard disk interrupt handling.", 1752,"Button",BS_AUTOCHECKBOX | WS_TABSTOP,24,72,204,10 CONTROL "Disable synchronous buffer &commits.",1758,"Button", BS_AUTOCHECKBOX | WS_TABSTOP,24,84,204,10 CONTROL "Disable all 32-&bit protected-mode disk drivers.",1751, "Button",BS_AUTOCHECKBOX | WS_TABSTOP,24,96,204,10 CONTROL "Disable &write-behind caching for all drives.",1753, "Button",BS_AUTOCHECKBOX | WS_TABSTOP,24,108,204,10 CONTROL "Disable System &Restore.",1814,"Button",BS_AUTOCHECKBOX | NOT WS_VISIBLE,24,120,204,10 END I also tried eXeScope, but eXeScope seems to do a poor job parsing controls inside the dialog box.

Three ideas: Blank out the text: "Disable System &Restore." -> " " Reposition the text box outside of the dialog box / property sheet: (999, 999) Resize the text box to something very small like: (1, 1) The more I think about it, removing the checkbox is probably the issue. Repositioning (method 2) should work visually, but will probably cause a "hole" in the tabbing sequence that will need addressing....

gstreamer.dll will load if the subsystem is patched, but the rest of the plugins won't load and Opera hangs as a result. There must be missing delay-load functions. KernelEx is required for Opera 10.6x HTML5 video and audio at this time.

Yes, although not intentionally. Opera runs in the default compatibility mode which I have set to 0/DCFG1/"Default mode". When I choose "Disable KernelEx extensions" for Opera, support for the <source> element is lost (the audio controls still appear). Any KernelEx compatibility mode enables it. The ogg clip plays as expected; the mp3 source fails in a way that correctly falls through to the ogg source (when followed by the ogg source), but does not display the error message (when no alternate source is provided). I also have Kexstubs and some other non-standard KernelEx extensions installed. I've been attempting to narrow down the exact dependency. Opera also works best with the Msimg32.dll from WinME. Edit: level of support Edit2: The problem seems to be gstreamer. gstreamer.dll and its plugins are linked for subsystem=5.0, so won't load on Win9x without the help of KernelEx. The plugin dlls are also in cascading subfolders instead of in the app directory, so perhaps can't even be found. Later tonight I will try patching the subsystem values of these dlls and placing them directly in the Opera folder.

10.63 works correctly

Try this (the "if exist" is incorporated into the file set): @echo off for %%b in (300????.jpg) do copy 0001o.jpg %%b for %%b in (301????.jpg) do copy 0001o.jpg %%b for %%b in (302????.jpg) do copy 0001o.jpg %%b for %%b in (30300??.jpg) do copy 0001o.jpg %%b for %%b in (30301??.jpg) do copy 0001o.jpg %%b for %%b in (30302??.jpg) do copy 0001o.jpg %%b for %%b in (30303??.jpg) do copy 0001o.jpg %%b for %%b in (30304??.jpg) do copy 0001o.jpg %%b for %%b in (30305??.jpg) do copy 0001o.jpg %%b for %%b in (30306??.jpg) do copy 0001o.jpg %%b for %%b in (30307??.jpg) do copy 0001o.jpg %%b If we note that the copy is unconditional, and presume that because of previous action, the files already exist, then we can reuse the previous plan of attack: @echo off for %%b in (30308??.jpg) do copy 0001o.jpg %%b for %%b in (30309??.jpg) do copy 0001o.jpg %%b Combining the two: @echo off for %%b in (300????.jpg) do copy 0001o.jpg %%b for %%b in (301????.jpg) do copy 0001o.jpg %%b for %%b in (302????.jpg) do copy 0001o.jpg %%b for %%b in (3030???.jpg) do copy 0001o.jpg %%b If we cannot presume that any files have previously been created, this second batch file can be done recursively: @echo off if %1*==* for %%a in (8,9) do call %0 3030%%a if not %1*==* if %2*==* for %%b in (0,1,2,3,4,5,6,7,8,9) do call %0 %1 %1%%b if not %2*==* for %%c in (0,1,2,3,4,5,6,7,8,9) do copy 0001o.jpg %2%%c.jpg

I just googled "4.10.2222B" and learned that it has caused installation problems for a number of apps, so that registry change is to be avoided.

If you can modify the iframe onload to be onload=tb_showIframe(this.name), the tb_showIframe function will receive the iframe's name as its parameter. Otherwise the script needs to know something as a starting point. If that is "id=TB_iframeContent" then the vbscript solution would be the syntactic equivalent of this javascript: <script>var iframeName = document.getElementById("TB_iframeContent").name</script> Note: if you want this to work in any browser other than IE, you should be scripting in javascript, not vbscript.

SWF files could be animations (FWS signature) or flash videos (VLF sig., sometimes with FLV extension). QuickTime for Windows should be able to play either and can be used stand-alone or as a plug-in both to play and to "export." If the files are flash videos, any FFmpeg-based converter should work.

@HoppaLong: If your ThinkPad's CPU is a P3 P4 or later, http://mirror.palemoon.org/archived/8.x/Portable/Palemoon-Portable-8.0.exe is an optimized, portable build of FireFox 8 (7zip self-extractor). Edit: Last build of Palemoon to support the P3 (and other SSE processors without SSE2 support) was roughly 3.6.9. (reference)