• Content count

  • Joined

  • Last visited

Community Reputation

0 Neutral


  • Rank
    Windows installer, chief cook and bottlewasher
  • Birthday

Contact Methods

  • Website URL

Profile Information

  • Country
  1. Been here before, but not recently. I develop a freeware package that has to run identically in XP and up, and I was adding a small feature while testing on a Vista64 system all the way up, but the bug presents on a clean install of Vista 64 SP2. I have a work-around but I am stymed I have to do so. I have to check this out in all the other supported systems, not sure if Win 7 is affected, but please help me out! I have a shortcut to a BATCH file meant to be stored with similar ones in an open window as a control center for a bunch of related activities. The idea is to have a shortcut on the desktop and users click on that to open this command-center window. Most of the icons/shortcuts in that window point to one of many .BAT files in another directory, etc. I happened to add another one and the icon winds up with the UAC elevation overlay on it. Worse still, making changes to the icon particulars can lead to the UAC actually being required! Here is the ONE change that makes it go away: If I change the name of the .BAT file and a corresponding change in the shortcut info the bug totally goes away! There seems to be a reserved keyword problem that you have to avoid the word SETUP in the file name. In particular If the name of the batch file is Setup Log File.BAT then the problem happens. However, if I change the fille name to Set Log File.BAT Then the entire problem goes away. I have isolated it down to merely what I am posting here, totally reproducible: The problem requires that you change to a different icon, any change will do, but predictably change the icon and the problem pops out! [The package has about two-dozen icons already, all with renamed labels on them to make sense to users, but the added frill is the only one that the BATCH FILE name pointed to starts with SETUP, and I happen to have a bunch of coherent icons, etc.] Not sure of other variations, but why can't the name of your own BATCH file not start with S-e-t-u-p? Any ideas. [The renamed shortcut still says Setup Log File so it cosmetically is fine, but I've never heard of seemingly ordinary reserved filenames in Windows! Any help? The BATCH does some mundane things to implement a new frill feature, and that always works, just this surreal-looking problem also shows up. Thanks in advance if anyone has ever encountered something like this. cjl
  2. Hi folks! I haven't been here in quite awhile [because I got all my problems solved here!]. But of course there is always one more: I am attempting to install Win7 on a Lenovo W540 w/8GB and the 2.70 CPU. I can run 8.1 on it fine, and I have also installed Vista [although many driver problems, but I still got it up enough to have it proclaim "genuine software". I can install the product key [that matches the Windows version] fine. But I cannot install the cert file LENOVO_V2.1_Cert.XRM-MS and get the error code 0x00000BBA The spool file was not found. Is this somehow related to the fact that the Win8/8.1 key is built into the BIOS? Do I need a[n even] new[er] Cert file? All else is fine, but only got about 3 weeks more to activate. Any help appreciated.
  3. ...To this day we still have a paradox: This is 2009 and there are certainly many things discovered of W98 (patches for newly discovered problems exist). There are new ways to integrate drivers (software like Driver Magician show basics of such if nLite doesn't to you). 4.7 maybe the last but if we consider adding current available patches (USBz, INFs, default VGA drvs etc) there can still be 4.8 of current date Can you elaborate about 4.8? All I have is a provisional 4.8 that literally has bugs in it if you want to just not address the SLEEK V2 situation. Unless you are attempting to do that [which doesn't even fully work] it is a step backwards for all of the other situations that work perfectly fine, as long as you are willing to apply all of my variant steps, etc. Or are you merely suggesting there could theoretically a 4.8 version that could work [based on a lot of hand-waving]?cjl ps: Another SLEEK paradox: The Recycle Bin empties really, really fast underl SLEEK. All other ways to run 98 up are really slow, especially if you have multiple partitions.
  4. I am the chief instigator of [among other things] getting 98lite SLEEK support put into the 98SEUSP2.1a. [You'll see an acknowledgment to me in the installation pages as you start the SP up.]I still use 98lite, and was one of the most prolific people attempting to get him to make a stable version 4.8 that would actually work, i.e., a "V2" implementation of the SLEEK shell. To this day we still have a paradox: 1) The SLEEK shell works perfectly fine in terms of file maintenance tasks. Without SLEEK, even with all of the kludges using files from IE55SP2, once you install IE6.0 or SP1, you never get stability. This is not a 98lite issue, just that IE60 was never truly compatible with any Win9x system, period. So, you have to stop at IE55SP2 if you want stability, or use IE60 and suffer from always just barely stable [as long as you don't do too many operations that slowly kill memory. Tip: A band-aid is to turn off display windows contents while dragging]. Of course there are minor issues of the Win95 shell having little silly problems such as the inability to properly size large directories [over 4 GB] but that's purely cosmetic, and of course the total lack of such things as a property-sheet-oriented "My Documents" icon, but I would gladly use sleek over CHUBBY or OVERWEIGHT because it runs rock-solid. I did all of the valiant effort to attempt to patch literally dozens of apps to be sleek-compatible with the shell32.w98, only eventually to be thwarted as newer versions came out that just couldn't be satisfied. [And that includes all sorts of stuff that could only be fixed after they were installed and crashed!] I was instrumental in showing shane all of the gyrations of how to shoehorn in IE60 [any version] using the original patched LOADWC.EXE I created, and he was working on integrating it in. I still use my own procedure to get IE60 installed on SLEEK; I have some systems I cannot use anything else on due to meager resources, etc. On faster systems, I still use CHUBBY because there is no downside; once you are using the 98SE shell, the IE-related damage is done, mitigate it as best as possible, and move on; all the apps will just install, etc. However, all the above said, if anyone is interested, I can post the procedure to get 98lite to install IE60 [sP1 preferably]. It is not straight-forward; a major user-level disappointment of the product, but it *does* work. Here's in outline form what to do: 1) The one most people would want to use: a ) Install 98SE straight. If you want, you can update the drivers, etc. b ) Repair install over the freshly installed 98SE, this time starting up the 98lite install instead. c ) Install IE60 SP1; fix it the best you can, a topic for all win9x that does not use the 95shell [sLEEK "V1" as I call it]. d ) Make a sizeable bunch of small patches, TweakUI stuff, and some .reg files I wrote to do the stuff that cannot be done without regedit, etc. Done that way, you get full app compatibility, and some 98lite improvements, just not all. Most notably, you get the IE desktop icon in an unbroken manner. [in WinME, it's often broken, again an issue beyond 98lite.] A small few desktop folder settings are modestly unstable, and can easily be put back; this could be an issue beyond 98lite, and could be driver-dependent [especially video]. I of course use a lot of the Me files that improve stuff, most notably scanfrag stuff. For the more "adventurous" types, a far-more involved install is to first install SLEEK on the first and only install. This involves a lot of the tricks mentioned above, but is only the start of things. You then need to shell swap to CHUBBY. Then you have to reinstall IE60SP1 repeatedly until you get it to issue a proper message; this usually takes a total of three passes at an install! Once done, you have to apply a registry patch I derived to get most of the cosmetics and property sheet for the desktop IE icon to work, and of course eventually do all of those IE-related shell problems since SLEEK is not in the picture. However, it is apparent that the cleanest install of IE60 paradoxically is when the SLEEK shell is in effect, despite MS's prevention of installing IE60 on Win95. Once this is done, the long-way-around to CHUBBY can commence, along with a somewhat larger laundry list of patches, etc. And one over-reaching point: Other than the designated shell swaps done exactly once if at all, NEVER swap again! It's all stable, albeit messy to perform, and I have profusely documented the procedure. The only remaining issues are: SLEEK V2 never worked. For unknown reasons, it is incompatible with some seemingly unrelated aspects of explorer.exe that get confused and either crash or hang. It accomplishes all of the major goals, but he was never able to devote enough time to solving the other problems, whatever is causing them. I have the patched files themselves, and they are quite easy to apply from DOS, but it makes random other things crash frustrating our efforts to get SLEEK to also be compatible, etc. ME compatibility was never finished. I created some batch files to fix what was overlooked; how to patch the relevant files was straight-forward. The only implication to the above is of course the lack of MS-DOS being available, so you need to dual-boot to change files, and if you use the SLEEK-then-CHUBBY method, you may have to install IE60 as many as SIX TIMES! And as I said, only the recommended method gets you a "pristine" desktop icon, wierdly enough. cjl [back to msfn after a serious and long illness, largely overcome]
  5. Hello erp!I'm back from the [almost] dead, and I mean that literally. I'm finally able to catch up on e-mail, forums, etc. and noticed this recent thread, so I assume you and others are well. [i've been out of the hospital for 6 weeks now, so it appears I'll live.] Can you give me the short version: What is the reasonable consensus on what is the most effective SP-related procedure to follow today? Are we using SP2.1a, or is there something stable newer? Is something *almost* stable awaiting something with an end in sight? Is there a work-around to make useful something beyond SP2.1a? What? cjl [enquiring near-death-experience survivors want to know]
  6. Good idea, and yes, it does work for all folders. But it does NOT work for the dialogue box in the run command which just stays reversed. [it does come up in list mode and you can temporarily goto details mode, get the little arrows and temporarily re-reverse it back to normal, then return to list mode now un-reversed, but it won't stay un-reversed when invoked later, etc.] Are there dialogue box-related separate settings unrelated to the "normal" folders? [it would appear to be at least partially true; If I make the rest of the folders all tiles, the dialogue box is still in list mode by default, etc., so clearly there is some "separation" etc.] cjl
  7. You can also use XXCOPY to "flatten" directories, and overall it's a lot better than ROBOCOPY. cjl
  8. There is nothing particularly AMD or Intel-centric about Win98 [sE being the variant of choice in the 9x world; see the 98SE unofficial SP sub-forum here on MSFN]. That said, drivers are crucial. Stability is ONLY determined by how well the hardware-specific drivers are correctly written for the hardware, not anything else. Irrelevantly, there is more built-in ANCIENT Intel drivers, thus this hasn't been an issue in years; today you have to add on everything, etc. However, 98 is designed to run in much less than 1.0 GB of memory and will crash with "too much" memory unless you set specific limits within system.ini which can be reached in safe mode once the system is up enough. Failing that, access the file from DOS and hand-edit in the changes. [Ask on the forum for safe starting values to get the system up.] Unless these changes are made, it is likely that the installation will stall after the first reboot [saying something to the effect of "running windows for the first time"]. If you can make your machine temporarily have .5 GB memory, it shouldn't have this problem, and then you can add the fixes later [a standard feature of the unofficial service pack]. The usual killer preventing 98 SE install is the lack of a video driver that can work properly, and that in turn often requires the underlying chipset support that also may not be available. With regard to the hard disks, you can use SCSI disks on any supported host adaptor without problem, or you can use up to the first 127 GB of any IDE disk as long as it closely resembles the PATA IDE disks usually used [and a patch is available to use the rest of the disk as well, but only if partitioned to 127 GB or less per partitiion]. The standard files usually work in this regard, albeit not necessarily optimally. There is a last-ditch way to get the video up to avoid using "compatibility mode" which in turn forces you to have 640x480 in 16 colors. It's a little bit better than that, but slow: There exists from MS a driver that was intended for Windows 3.1 to run in VBE mode. This is EXACTLY what XP does when it lacks a driver for the video card. It runs slow, but is still capable of at least 256 colors, perhaps even 16-bit color. This takes advantage of the fact that for many years, all video cards are VESA VBE compatible. The only problem is that the driver didn't quite get its VESA act together and only ran on a small set of then-current ATI cards. However, there is a patch for supervga.drv that makes it truly VBE-compliant compatible. As such you can select 640x480x256, 800x600x256, and 1024x768x256. However, since this is a driver for Windows 3.1, and not Windows 9x, you lose a minor convenience point: We are all too accustomed to the slider that changes the resolution. In this driver set, there is only one setting per driver, so to change the screen resolution, you have to select one of the three each time. Other than that, Windows 3.1 drivers are compatible with Windows 9x! I have built Win 98SE for a T43p ThinkPad using this method. There can be a glitch: You may not be able to select full-screen text mode for the MS-DOS prompt and instead use a window of any size you wish, and just never hit alt-enter to flip it, etc. Yes, it's as slow as XP running without a driver, which isn't all that bad on a modern machine. However, no driver is running in MS-DOS compatibility mode, so all else is fine. You can limit all your partitions to FAT32 and 98 can access all of the drives, and can also defrag all of them a whole lot better than that built-in Diskeeper garbage. [And better yet, use the specific one from Windows ME, which does it better and faster still.] Note: In some cases, you may get 98SE up if you can limit memory temporarily to 1.0 GB depending on the machine, but don't count on it; be prepared to make those system.ini changes so it cannot screw up regardless of actual memory size. Note: there are no problems whatsoever using 98SE on AMD boards with VIA chipsets and CPU's such as Athlon XP 3200+ on down. Again, assuming the memory problem is contained either before or during the install, once the proper VIA drivers are in, all is totally stable. Of course, if NVIDIA you are totally at the mercy of their [lack of] support, etc. P3 and P4 drivers from Intel vary in support of 9x, but those that not only do but also support the so-called "application accellerator" Intel stuff gain something else: Full 48-bit LBA support, thus no upper limit on drive size per se. [Note: You MUST partition the drives into chunks each no larger than 127 true GB, but there is no part of the disk that would be unaddressable even if you have a 1 TB disk partitioned into 8 logical drives.] There are also third-party drivers available to augment/patch the standard driver in 98SE [for a nominal fee] to accomplish this as well; again how well your disk performance goes varies, but generally just about any disk is compatible enough with the standard driver thus getting adequate performance and full access to the first 127 GB just after installation, etc. And when the hard disks are connected to BIOS-ROM-enhancing controllers from the likes of Promise, Vantec, etc., all of the drives are accessible from DOS and generally there are MPD drivers to allow full-speed 32-bit access to the disks as well. I understand it might seem silly to hook a SATA drive to a plug-in card when you have 4-6 built-in interfaces, but for example, the VANTEC card allows me to use a full-length external E-SATA cable to a drive in an external enclosure. All of those kludges to give rear-panel access that starts at the motherboard don't work unless the connecting cable is impractically short, etc. If the interface itself is on a plug-in card, you can get a 6-foot cable to the external disk AND have 98SE 32-bit access [as well as XP]. There are variants of this card for PCMCIA, PCI, and PCI-E available, etc. The PC-card supports two drives while the other two support one internal and/or one external whether 98 or not, etc. The best answer is TO TRY it. Remember the memory problem and learn to solve it, otherwise it can't work. You may be quite happy otherwise due to all of the backward compatibility all manufacturers have to deal with. [Note: Invariably, running some utility on ANY board involves DOS, not even anything called Windows!] cjl (A small part of the team that created the 98SE unofficial service pack)
  9. Just to not waste time, I did try changing to view order by name and selected something, but it didn't become permanent, so next time still reversed. cjl
  10. Somehow when browsing in the run command [defaults to list view, correct by default], I can browse to a volume and all of the folders and files are displayed in reverse order. If this were details view, I know how to re-reverse it back to normal, but how do you reverse the list view? [i'm assuming that if it gets reversed back, it would be permanent until again changed... but then again, I don't know how this got reversed in the first place!] Any ideas, people? tia, cjl
  11. I am attempting to include this update [KB917275] in an unattended batch file. Using no switch options it does install fine. Doesn't even demand a reboot. However, I have to answer all of the questions. Notably, this update seems not to support /? to find out what the other options are. I did try /Q which made all of the questions and answers go away, but, it seems to make it not install anything, just complete as if I cancelled out on the answers, etc. Could this update possibly need some sort of response file [unattend.txt or whatever] or invoke some other option switches? [i don't know of any other update with this particular collection of traits, but some other new updates are appearing here and there breaking with their pasts as well]. Thanks guys, cjl
  12. I use the QFECHECK-type program available in KB282784 (Qfecheck.exe Verifies the Installation of Windows 2000 and Windows XP Hotfixes) and haven't ever had this problem before. For all other updates, either no complaint or at most a reboot and in an extremely rare instance actually reinstall the relevant update. But for the recent KB923689 (and no other old or recent update!) Qfecheck always gives KB923689: This hotfix should be reinstalled. Of course I have a sea of other fixes installed (and all of them are "Current on system." as they should be), but this is the only one that ever failed to correct for obvious reasons, if applicable. Also, KB923689 cannot be uninstalled, so a reinstall is probably not doing much. My system is XP SP2 with Media Player 9 then upgraded to 10 then 100% current as of two weeks ago with all patches from WU, except I hid IE 7 and didn't install it, then apply the KB923689 from either Windows Update or a download [doesn't seem to matter; tried this on several systems already both ways to install, etc.). I have three similar machines and they all get the same end result. I also have a GHOST image of where I was two weeks ago back before "patch tuesday" if that is helpful, etc. Note that Windows Update itself is perfectly fine with the update being installed, i.e., after you install it, they don't ask you to install it, etc. I know this could be a bug of the Qfecheck, but maybe they miswrote something in the update? [Note: MS screwed up exactly a a parallel thing in Win98 about a year ago. Some MSFN forum members fixed it by providing a patch for the registry info that was the culprit; qfecheck WAS correct to complain in that case! Unfortunately, this NT-family Qfecheck gives out a whole lot less info than the 9x version.) Can someone provide a list of what registry changes this update is supposed to perform? It's possible the registry is erroneusly reporting WMP 9-related info (I use WMP 10) or somesuch. [To my knowledge, the only thing the update does is update WMVCORE.dll to an appropriate rev for a WMP 9 situation or a WMP 10 situation; perhaps the registry doesn't match what a WMP 10 situation actually is?) Thanks for the help, cjl
  13. Well, we went over this before:There really isn't a need for superfluously separate packages to update 98SE. The objections only came up if there were unacceptable options. However, options different from the current scope of the SP as of 2.1a could become: 1) Do less - Don't even bother to check for IE 5.00 and update it, just ignore IE completely. Makes the package only insignificently less bloated than otherwise. 2) Do check for IE 5.00 and update if present - Baseline size for bloat of SP currently 3) Do check for other IE versions and update if they are present, NOT just for IE 5.00 - Adds insignificent bloat IF accomplished in conjunction with an optional .CAB module that contains all of the relevant updates to the version that might be updates. Clearly the "bloat" of checking for the option to do this is much smaller than the current "bloat" of actually updating IE 5.00 should it exist, etc. Thus, there are no options that bloat anything, and everyone gets what they want. It could become a quaint tiny "bloat" to maintain the update for IE 5.00 if present, but the point is that no case can be made for not having the SP update all relevant versions of IE without providing ANY bloat. Just package it accordingly. You can have a base add-on .CAB file for IE 5.00, and anyone could chose not to want it. The net effect is that the SP is slightly smaller when used for most of us who could care less about IE 5.00. For those who care, the total size is the same; they just have to also load a tiny file that would become static contents for the IE 5.00 that at this point will never have another update at all. Net effect is a small positive - future versions of the SP stay permanently otherwise slightly smaller and the few who want to have the IE 5.00 updated won't even need to load the tiny file again, since it would be a static optional small .cab or whatever file, etc. But for the larger group of us that want an added IE option for their pet version, whichever that one is, making the SP be the master installer of all relevant updates is a far bigger win. Having to necessarily have an additional installation package or (URP) as you call it [are you also the URPman? ] becomes the "bloat" for some of us. [And if insisted upon, why not then create a stand-alone URP installer compatible with the add-on module that benefits the SP anyway? This way, all of us get what we want, etc.] cjl
  14. Does he know, or are these just slapped up on his website? I don't see how anything past what I report can change anything.It's apparently true that the Q832353 version doesn't like my configuration. Perhaps it wasn't tested in 98SE or in conjunction with IE 6.0SP1 and patches, even though usually obtained along with IE6, etc. There really isn't much to explain about an explicit message "You must install Windows Media Player 6.4 first" is there? Upon examining the innards of both patches, it's clear that Q832353 was written to expressly try to solve the 6.4 problem, and clearly MP832353 handles all of the other cases for 7.1 and 9.0, thus they would be complementary if they both completely worked. I can only assume the author of the former is aware that the latter doesn't work for 6.4, or else why produce a patch for the only case needed? [The original, let's call that one KB832353 for clarity, is intended for all three cases, but won't work directly under 98SE, but does work manually using files within, but only for updating 7.1 or 9.0; the 6.4 case does not work. I would like to know if the original even works for ME in the specific case that applies: 6.4 and 7.0 and no upgrades would be the relevant ME configuration, since strictly the package only updates the highest version you have and only from 6.4, 7.1 and 9.0 while that ME configuration doesn't qualify as 7.1, only 7.0, thus would qualify for the 6.4 update, etc] So, once more for the record: KB832353 from MS hates 98SE, was written expressly for ME, and I want to know if it ever updates WMP 6.4 should the highest version from the set of only 6.4, 7.1, and 9.0 be 6.4, essentially an initial install and never updating 7.0 to 7.1 [which clearly wouldn't get updates because 7.1 is a prerequisite for the relevant part of the update]. KB832353 when unpacked can manually be used to update 7.1 and 9.0, but not 6.4, all of this from 98SE. MP832353 was apparently written to at least allow usage from 98SE whether unpacked or not. Although it claims it handles all three WMP versions, it clearly doesn't work for 6.4. However it works perfectly fine for 7.1 and 9.0 thus eliminating the need for manual updates from the KB832353 version in 98SE, but still no solution for users of only 6.4. When you attempt to use it, there is some activity, but the file doesn't update Q832353 clearly was written only to upgrade 6.4 only. This would make sense where both the KB and MP versions fail to accomplish this, at least from 98SE. When unpacked, it allows a successful install of the upgrade to 832353 level, but only manually. When used as packaged and executed normally, it issues a complaint message telling me I need to first install WMP6.4 even though I already have when I earlier installed IE6.0 SP1 with 6.4. I have also tried installing MPfull.exe over this to no avail. I can only conclude that it wasn't tested in 98SE or that it interacts badly there with IE6.0 SP1. If it can be made to work in 98SE at all, I would have to install MP6.4 before installing IE6.0 SP1, which I can try in my next overall install, but it is contrary to my typical install order, and I believe also anyone else's recommendations along these lines. Especially is this quirky in light of the manual usage that functions fine. The need to fix this stems from the fact that there are further updates for the newer Media Player versions; you need not even apply 832353 for 9.0 since it is superseded by I believe several other updates, one even better than the others. However, for 6.4, this is the highest revision for this particular file, replacing 828026 which was known to be problematic thus MS wrote this update at all, etc. cjl