jds

I did not try to Import anything, but I did create a new Registry successfully. It just takes a long time. As I noted, make sure you use SMARTDRV or you will be waiting days. If it fails, there may be other issues, not just size. Hmmm ... the problem is with the import/create step of the procedure. So perhaps 'regedit' doesn't work any different for you as it does (doesn't) for me. And yes, it took over half an hour (estimate) to export the registry. Joe.

Well, I've never managed to find a utility to set the CPU clock. I know Intel tend to lock the upper clock limit, so perhaps nobody has bothered with a utility for this reason. However, I've now installed Flash Player 10.1.82.76 (I also have KernelEx 4.5.1 installed) and the sync problem is gone. Depending on CPU needs, the video either plays as a series of still frames, or plays normally. No more of the "sound plays fully, then the video" effect. Flash Player 10.2.161.23_Beta_2 may also be OK (haven't tested), but 10.3.181.14 is no good (crashes). Joe.

Hey, great idea! Unfortunately however, it didn't work. RegCon failed with the same error. Haven't tried SCANREG, but REGEDIT (DOS) does work. Extracting and rebuilding the Registry took two hours with SMARTDRV enabled. Without SMARTDRV plan on two DAYS. Interesting. So with 'regedit' it seems, YMMV. I can export the registry OK, but not create/import it. Joe.

Well, I've finally tracked down a copy of 'mtppk11.exe' at this part-time web server : http://dizzyonion.dyndns.org/uploads/Microsoft%20Tools/Multimedia/MTP%20Porting%20Kit/ Comparison to mtppk10 : Apart from a few updated DLLs, code samples and utilities, it includes the above 'MTP9x-2k' and also 'wpdmtp_sample.inf' (kinda useful!). On the negative side, it omits 'wmdmcmd.exe' and 'readme_wmdmcmd.txt', so you still need 'mtppk10' for these (actually, the latest version of 'wmdmcmd.exe' I can find is in the huge 'p4sdevicekit.exe' package, but MS didn't have any room there to fit the "readme" file). So, with 'MTP9x-2k.exe', 'wpdmtp_sample.inf', 'wmdmcmd.exe' (and 'readme_wmdmcmd.txt'), you have rudimentary support for PTP/MTP devices in Win98/SE/ME. Install 'MTP9x-2k', make an INF file for your device using 'wpdmtp_sample.inf' as your template (use 'usbview' to find the appropriate PID and VID values), then use this to install the device driver. Finally, use the (rather crude) 'wmdmcmd' utility to copy files to/from your device, etc. (hint, 'wmdmcmd' can operate in command-line mode, so you can take advantage of the command-line editing features of 'doskey'). Joe.

Do the MS 'scanreg' and 'regedit' DOS tools work properly for you with such large registries? I've found (per above) that 'regedit' does not. As for 'scanreg', it seems happy to backup, restore and "vanilla test" large registries, although I haven't been brave enough to try "scanreg /fix" since my registry has grown beyond the 8M mark that is mentioned earlier in this thread. Joe.

If you need a copy of this "lock.exe" utility (to reproduce this problem), DR-DOS 7.03 can be obtained ... Well, a new clue : This errant behaviour is not present in KernelEx 4.0-Final2, but is present in 4.5-Beta2 and 4.5-RC1 and all subsequent "4.5" versions. Joe. PS. Hey, today this problem is gone! I applied the Unofficial Q932590 hotfix yesterday, so that must have done the trick!

Well, P3CPU and FineSSE only come into play when an invalid instruction exception occurs. However, just for completeness, I temporarily removed these and tried VLC 1.1.9 again on the P2. Unfortunately, same result. The main GUI appears and seems functional, but when a video is loaded, the player window appears for a fraction of a second, before VLC comes crashing down without any sort of error message (on restart, VLC says that it just crashed and offers to send a report). So with KernelEx, I can get VLC 1.0.3 to work fine on my P2 (except that two colours are swapped, yuk!), but not VLC 1.1.9. Yet on my P3, VLC 1.1.9 works great. It may not be a CPU issue, but I wonder if anyone else has managed to get VLC 1.1.9 to work on a P2 (with KernelEx)? Joe. PS. Well, perhaps it is a CPU thing after all. All versions 1.1.* seem to have the same problem on my P2. And occasionally, FineSSE does report an illegal instruction : 63881DCD:0F AE F8 0F 77 8B 44 24:SFENCE =>90 90 90 0F 77 8B 44 24

Thanks, Den. Unfortunately, time has been rather scarce lately. However, I did check that this thread hadn't been updated before posting. I'll be downloading those files shortly. Also, the caution about the "regedit /c" technique still stands. I tried a refinement of this as follows : 1. Export the hives of the registry as separate files. 2. In DOS, start a new registry with "regedit /c hkcr.reg" 3. Add the other hives in turn, ie. "regedit hkcu.reg", "regedit hklm.reg", etc. However, at "regedit hklm.reg", this error was produced : "Cannot import hklm.reg: Error accessing the registry." Joe. PS. I've now tried the English-patched 'RegCon' utility and unfortunately, this crashes with the error "SYSTEM.DAT: make sfx failure(32809)". I also tried the associated 'regstrip' utility, and this loads 'user.dat' OK but fails on 'system.dat' (gets part way through, so it looks like a size issue). So, no joy with these utilities by 'haltz'. I also downloaded but can't use the "TestRun by BB" utility, since that is hardcoded for Windows installations on drive C: (whereas mine is on drive E:). BTW 'scanreg' and Norton WinDoctor (1998 vintage) both say my registry is OK.

My 'system.dat' is about 12M, and I just tried the above. Unfortunately, the "regedit" step stopped at 74% complete (no progress for 30 minutes) and I was forced to reboot. In DOS mode, I observed that 'system.dat' was now about 9M. I then did a plain vanilla 'scanreg', which reported the registry corrupted and restored a backup copy. So it looks like the above procedure cannot handle a "large" registry either. BTW, while I was able to find the Japanese 'regcon' package (version 3.11, file 'rgcon311.lzh') via the Internet Archive, the links to the English translations here are defunct, so I was unable to try this. Joe.

Just in the past couple of days, I've installed Opera 11.11 (using W2K compatibility mode) on two machines, both W98SE with KernelEx 4.5.1. Generally, I'm impressed with it, compared to FF (2.0.0.22pre). On one machine however, no icon appears on the taskbar when Opera is started. I've looked around for an explanation/solution, but the only similar thing I can find relates to Linux. Ideas anyone? Also I've noticed, on both machines, Opera doesn't like to open multiple instances of itself. Minor issue, although more annoying when there's no icon on the taskbar to remind you it's already running. Joe. PS. Just solved the mystery of the missing taskbar icon for Opera. On a hunch, I installed the Unofficial Q932590 hotfix for 'msvcrt.dll' from MDGX's site. (A similar problem existed for Wireshark 0.99.4, now also fixed.)

After installing Opera 11.11, I also tried printing. The printer driver pop-up appeared (BTW, I use a network printer), but after clicking "OK", no actual printing occurred. Then after a reboot (more on this later), I tried printing from Notepad and Wordpad, with and without various KernelEx compatibility modes. Basically, if I use any NT mode (I tried NT 4.0 SP6, 2000 SP4 and XP SP2), the same occurs as for Opera (except that for Notepad, the driver pop-up window only flashes briefly, which is normal behaviour). Furthermore, after resetting Notepad and/or Wordpad to KernelEx default mode, I must either log-off/log-on or reboot before printing will work again. So, this problem is not specific to Opera. Joe.

Encouraged by this report about VLC, I downloaded and installed version 1.1.9 and it worked just fine on my P3 machine with KernelEx default mode. So, nothing particular seems to be required, other than having KernelEx 4.5.1 installed. Then I tried to do the same on my P2 machine (with P3CPU.VXD and FineSSE23.exe installed) but although it installs and initializes its font cache OK, it crashes when trying to display any video (no error message). Joe.

Thanks for the tip! Just the sort of detail that makes it worthwhile to try this; I've just installed and am trying Opera for the first time in many, many years. Looks good so far, and there's a huge improvement in speed of JavaScript-laden pages (such as here) compared to FF (v2.0.0.22pre). Joe.

Unless you got it already, here is the hard to find MTP9x-2k: http://www.mediafire.com/?r155zetst89543n MTPPK10 or 11 are apparently only for XP from what i gather Version 12 is available from MS but requires WGA for dowloading it and version 11 nowhere to be found but version 10 can be downloaded freely here in case you think something might be compatible in it: http://am.net/lib/tools/Microsoft/WinXP/mtppk10.exe Have you got an older porting kit? Hmmm ... I've searched high and low, I've previously found the above 'mtp9x-2k' and 'mtppk10', but not 'mtppk11'. The MS link to 'mtppk11' redirects to 'mtppk12' (BTW, I can use GenuineCheck 1.7.0069.2 to complete this download) and the two "rapidshare" links I've found for 'mtppk11' are defunct. As for versions 10 and 11 only supporting XP or greater, this is incorrect. Version 10 definitely supports W98 (but is dated) and I've seen many reports of people installing version 11 on W98. Joe.

Does anyone know where to download 'mtppk11.exe' (Media Transfer Protocol Porting Kit, version 11)? This was apparently the last W9X compatible version ... Joe.

An update : I now have partial success ... 1. Basically, all versions give that nonsensical error (per first post, above) when running in plain W98SE. 2. Versions 1.0.7.4880 and 1.0.8.4990 seem to install successfully when run in "XP Compatibility mode" under KernelEx 4.5.1. 3. Version 1.1.0.5790 is a mystery, since all versions I've found to download, have actually been a later version, misidentified. 4. Version 1.2.* to 1.4.* are nowhere to be found (1.2.0.3 supposedly did exist). 5. Version 1.5.0.7220 just crashes with an Access Violation. 6. Version 1.5.1.8210 and above, give that nonsensical error, regardless of KernelEx. So, if anyone knows of genuine versions 1.1.* through 1.4.*, I'd like to give those a try (the two versions that do install for me, aren't adequate for the Netgear software I'm trying to run). Joe.

Are these the last 9X compatible versions, or were they the current versions at the time of this posting? I initially downloaded version 3.1.15.0 and this threw an Access Violation Exception during installation. However, I then figured the download URL for version 3.1.9.0 and installed this. During installation, it offered to download and update to version 3.1.15.0, which I accepted. So now version 3.1.15.0 is installed successfully. Perhaps if I didn't accept the update offer, and installed the previously-downloaded version 3.1.15.0 on top of version 3.1.9.0, this may also have been successful. A word of warning: According to Avast, the optional "Ask toolbar" that comes with version 3.1.9.0 is infected by a Trojan or Virus or whatever. I had deselected this optional install anyway, but this seemed to put Avast into an infinite loop when I selected the "Delete" option, so I eventually had to tell Avast to take no action (the Trillion installation was progressing a bit at a time, but I lost patience with the recurring pop-up warning from Avast). As for Trillian Astra v4.0.0.78, I couldn't find a copy of this anywhere (I presume this is also known as Trillian Astra Alpha Build 78), nor any version close to this. Joe.

It might be the same problem I encountered a few months back : http://forums.seagate.com/t5/Barracuda-XT-Barracuda-and/Silent-write-failures-with-500G-SATA-in-JM20337-USB-enclosure/m-p/72767 Joe.

Version 8.1.0.421 also seems to work fine. Joe.

Eh?! What strange world do you live in? Anyway, there's the specific requirement for handling BD content (assuming you have the appropriate rights or subject to the fair-use provisions of the copyright act for commercial content, as may be appropriate) and then there's the more general requirement. For the former, perhaps this software may be some use (no, I've never used this, so caveat emptor, etc.) : For the more general requirement, FAT32+ support seems to make the most sense. Joe.

KernelEx-4.5-Final causes a GPF when the DR-DOS 7.03 "lock.exe" utility exits : If you need a copy of this "lock.exe" utility (to reproduce this problem), DR-DOS 7.03 can be obtained from any of the following archive locations (Caldera, Inc. at the time, released DR-DOS 7.03 as a free download for non-commercial and evaluation use, as described in the 'LICENSE.TXT' file) : ftp://ftp.uni-bayreuth.de/pub/pc/caldera/DR-DOS.703/ ftp://ftp.fsn.hu/pub/OpenDOS/DR-DOS.703/ http://replay.waybackmachine.org/*/http://www.au.horde.org/pub/drdos/DR-DOS.703/ While the entire package can be downloaded as 'dr703.exe', for the purposes of this exercise, the following two files are sufficient : images/disk01.144 images/disk02.144 Also useful is a free utility to extract individual files from the above images, available via : http://www.winimage.com/extract.htm Once you have all three files (and extracted 'extract.exe' from 'extrac21.zip'), you can extract the 'lock.exe' utility as follows : extract -e disk01.144 pnunpack.exe extract -e disk02.144 lock.ex_ pnunpack lock.ex_ Joe.

Actually ... there's a 2G bug in there to deal with, before worrying about 4G! : Joe. Technically it is not a bug. It is "By Design". To work with files larger than 2GB, you need to set a flag on an "Extended Open" Call. A small TSR probably could be used to force the flag. Great catch Mr Loew! From Ralph Brown's Interrupt List : So Int 21/6C00 does provide a clue. However, since I also use DR-DOS 6.0, which does not support Int 21/6C00 (BTW, DR-DOS 7.XX does), I know most 16 bit applications/utilities I have don't actually use this particular function call, but the older Int 21/3C or Int 21/3D (I haven't verified 'chksum.com' yet). So it would be necessary to intercept these older function calls too and redirect them to Int 21/6C00 with a suitable tweak to BH. I dunno, this seems like a bug that became "by design", so as not to admit it was a bug in the first place. Why on Earth would you deliberately constrain file sizes to 2G? Joe. Update: I've now tested Charles Dye's 'chksum.com' utility (which was where I initially discovered the 2G file size limitation) on DR-DOS 6.0 (which doesn't support Int 21/6C00), and since it works there, can confirm this utility uses only the older Int 21/3C or Int 21/3D function calls.

Actually ... there's a 2G bug in there to deal with, before worrying about 4G! : Joe.

AFAIK, the maximum size of VOB files on DVD is just under 2G, although the normal practice is for them to be just under 1G. However, that's not the point. The point is that DVD can already exceed the 4G content barrier and has this scheme of handling such content, with file sizes that are suitable for many O/S including W9X. So : (1) Does BluRay use a monolithic file for content, or does (can) it support the same scheme as DVD? (2) If BluRay content is monolithic, why not have an application/utility that translates this content into the "DVD-style" multiple-VOB-file content that can be readily handled and is already supported? Perhaps some existing DVD-copying/ISO-burning software can already do this? Joe.

Well, so far, P3CPU has been perfectly stable. FineSSE is also installed. With Flash 9r289, I have reproduced the "sound plays fully, then the video" effect reported by Larry some time ago (for testing, I use the Kate Bush "Army Dreamers" clip he gave at post 44). I want to see if this may be a CPU throughput issue, by running the same test on a genuine P3 with a reduced clock speed, say 200MHz. However, my P3 BIOS doesn't give me any options for the CPU clock multiplier, so I'll have to search for a utility that may do this. Joe.