Drugwash

Oh, sorry... Sleep well!

People can build simple scripts to check for specific text on web pages. Having the exact file size displayed on the page together with keeping at least the top structure of the page identical, would help people know when there is a new alpha/beta/release published, without the hassle of forum login/logout/search/etc. More so when lots of intermediary versions are being released within short periods of time, when one may not be able to keep count whether they downloaded a particular release or not; the exact size would help avoiding useless redownloads, thus alleviating server's load. So the way I see it it may not be just a whim but an automation attempt. On another note, I haven't been able to download current release, neither from main link (download doesn't start at all) nor from the mirror (download freezes somewhere after 8MB). In Firefox I see no notice, but in SlimBrowser it says my number of downloads has exceeded 6 in 24 hours. So using FlashGet to make sure the package comes integer (due to this bad modem connection - max 10kB/s download at the time of typing this - and large package size) practically kills any chance for me to get anything from the main site. Wonderful! Why do I even bother...

Oh, no need to apologyze, it's a fairly common confusion, happened to me too, a few times. Don't worry, I know my place in the world is somewhere down the line. But let's not make it off-topic. Anyway, it's good info and someone might find it useful. Muchas gracias! B)

I know the flag here is similar but I'm not in Belgium (unfortunately). I'm not even sure we're allowed on eBay. You don't know what 'poor' means (and I hope you never find out). <sigh>Nevertheless, thank you for the suggestions - it's good to know the available options, in case the opportunity comes up. For now, I'll stick with what works and leave drastic measures for later, when time comes.

One shallow mention I found on the Lunchbox page you linked to above, but it may be related to the mod. My hunch is that it may be the one that enables lower voltage range but in this domain we work with certainties, not hunches. That's what I got on the Vectra. That used to be a good machine but the HDD crapped out, developed 4000+ bad sectors and not even multiple repair sessions with HDD Regenerator and HDAT2 couldn't fix it back. Freakin' ironically, after years of usage, few days ago I just found out that the HDD is actually 40GB not 20GB as the additional label said! And this appears to be (or have been) common practice for many OEMs when they had a large order and no direct match at hand. Not in this country, amigo. Not in this country...

Thanks, I found that japanese site soon after posting here. Apparently the jumpers match but there's still no mention about J2, whether it should be open or else - they only say "reserved", but nothing about its default state. I'm not trying anything fancy because that's the only adaptor I got (and I read a lot about those Tualatin mods these days, I know about cutting pins, nail polish, drilling through, etc) - all I want is a working setup for a potential replacement of my Slot1 667MHz CPU. However, al I got here is a 1000MHz Celeron with 100MHz FSB which is not worth the bother since the 667 has a 133 FSB and it's a PIIIE Coppermine not a Celeron. I do have a Tualatin Celeron (Tualeron, Celertin as they used to call it as overclockers.com and elsewhere) but that one is also a FSB 100 (@ 1300MHz speed) and I wouldn't want to destroy my adapter for it. The only worthy replacement would be the Coppermine 1000/133 in my HP Vectra, but I've not yet decided whether I'll downgrade that machine. Besides, my board is a Soyo SY-6VBA133 which is VIA-based and I read the guys didn't have much luck with modded Tualatins on the VIA chipsets anyway. But being my main machine I'd rather play it as safe as possible. (just fixed a minor typo) (somebody should fix the board - it's inserting plain HTML tags when Quick Edit is used under IE).

Depending on the settings in the SP3 package, some existing files may have to be removed prior to installing those in the package. I'm referring here to driver files with version numbers higher than the ones in the service pack, which may not be downgraded at install time and whose presence could be the real root of the problems in certain situations (such as the Intel/VIA drivers discussed somewhere above). Nevertheless, there must be a safe procedure to remove those USB drivers only in certain conditions, as we don't want anybody to inadvertently clean up their registry and then come back here with guns loaded - figurately speaking, of course.

An USB driver remover could be custom built, specifically for Win9x, but I'm afraid it could also be used for malicious purposes. Although at most it would just remove all USB drivers, this could be a major inconvenience for people that don't have a copy of the driver kit, those that don't have the knowledge to (re)install them or when device driver installation requires a certain procedure that is not known to the user. Besides, such tool should also account for any connected devices, check if there's any running or pending file operation involving such device, ask the user to stop such operation(s) and then physically remove the device(s). After all this, there may still be devices whose specific applications come bundled with the drivers so the tool should warn that some applications may have to be manually uninstalled and reinstalled. As you see, it may not be such an easy task to build a reliable tool. Just plain deleting some registry keys - as risky as it is by itself - and some .inf files would not be enough. It's been quite a while since I've been planning to build such thing but always got sidetracked by so many other things. Dunno if I ever get to do it, at least not very soon.

Your modem http://www.alibaba.c..._modem.html?s=p seems to be recent. Perhaps changing your FTP settings (bigger local cache, I increased mine from 4096KB to 40960KB) could help if the files not downloaded were big files, as I would speculate. Or repeat the download for just the missing files. Or try another modem, maybe the FTP of your modem doesn't work with the hardware of your ISP... The modem is under lease from the ISP, on subscription and alternatives are much more expensive and less advantageous, something I can't afford. FlashGet never had trouble downloading big files over HTTP or other protocols, but one never knows. Not sure if there's any such FTP cache setting in FG. Well, the tool of my choice is TC and what you do with a dozen separate apps I can do with only one, provided any necessary plug-ins (which are mostly freeware/open-source) are installed and properly configured. To the point: the list you provided has the same format as my .lst, only it has the extra MD5/CRC that the TC plug-in DiskDirExtended doesn't provide. Even so, changing file extension from .txt to .lst allows me to browse the folder tree in your list - difference is that date and time will be zero. The point of this "browsing" is to have a direct visual of the folder tree structure, as opposed to a long list of files and folders in a hard-to-define order. Other than that, to each their own. Please note I am not directly interested in the Reskit itself - as I can build my own tools when necessary - but more in the quirks of FTP downloading and other side-effects of this exercise. If anything, we just share personal experience. Thanks for your input so far.

Yeah, and most probably of the Pebkac species, there seems to be a few of them who have abducted this thread. Since I'm in a (relatively) good mood, let's just blame it on the state of exhaustion I was in, at the time of posting. Maybe I forgot to turn the page. I usually recognize that "alien species" you're talking about when I see them. I'm used to a slightly different approach: there's a packer plug-in for Total Commander that allows one to create and browse files of type .lst, which are exact replicas of the folder tree used as source when creating them. With the only difference that any .lst file neested within such .lst will appear in the form of a folder, allowing to browse its contents.If you have the means to browse .lst files you'll see what I got. download EDIT: This topic goes much faster than my Internet connection. I don't have a (working) router in my house, connection goes through a GPRS modem (ZTE MF-190 or something). No clue on the hardware used by my ISP along the way.

Do you happen to have a manual/datasheet for that MSI slocket? Mine is v2.3 and the only manual I could find does not match the actual jumpers on the PCB. I think I might have fried a couple CPUs while testing it, because of wrong jumper connections (unless it was already defective when I got it).

The correct path regarding capitalization of folders appears to be: ftp://ftp.microsoft....s/Win98/Reskit/ I had previously downloaded the whole bunch of files and they all got mixed up, some main folders were not created (has been previously reported). There appears to be a healthy crop of FPT paths, all with slightly different casing, such as services -> Services or technet -> TechNet. All this could be a deterring attempt (with my paranoia hat on) or simply a clash of paths between multiple servers in different locations that have not (yet) been synchronized. I'm still in the process of downloading but so far the folder structure seems to be preserved, using FlashGet 1.9.6.1073 under XP. Since there's been quite a talk on FlashGet earlier, I should say I've been using it ever since version 0.96a back in 2001, mainly because of its browser hook which offered the 'Download all with FlashGet' option in the page context menu. What I wanted to add is that the security concern that came up earlier may be based on the fact that after 1.7x they built ed2k/Kad and t0rr3nt capabilities into FlashGet, which require multiple port usage (both TCP ad UDP). Anyhow, I personally had/have no fear in using it. It still has some bugs though, which may well depend on the current version. The one I'm using chokes at some random file and needs manual stop and restart that file to resume correct operation. Other times it just freezes for a few seconds, when download speed is too high and/or too many streams are running, but that may also depend on settings (including OS TCP/IP settings, TTL timings, etc) As it just finished downloading, Total Commander reports 2375 files in 91 folders for a total of 169,970,943 bytes. According to the numbers throughout the topic, it seems FG missed a few files but comparing to submix8c's index list would take too long and I just noticed the sun is up already and I haven't slept. Bottom line is, there may be more paths to different (incomplete?) versions of the Reskit folder or there may be bugs in the software that performs the path aliasing. Or whatever. What was this Reskit good for, actually...? EDIT: Interestingly enough, I quick-posted in one thread and suddenly find myself in another one. Aliens?!

Nice to see you back! Got one small issue: some of us have a customized Device Manager applet (mine dates back 2006 when I installed the system). There's no warning and/or backup option in the package about its replacement, if I'm not mistaken. Those who get the package from third-party sources (copied from a friend, whatever) will definitely miss the tiny mention in the changelog posted here at MSFN. Therefore, could you please add to the package (at install time) a visible warning that the Device Manager applet will be replaced, so anyone installing the package could get a chance to backup the old one? Other than that, thank you for the update!

Actually I've had a catalogue application for years and have stored each CD/DVD's contents through it, so it shouldn't be a problem to retrieve the exact location of certain files. It can catalog/retrieve files in common archive types but I'm not sure if it can open .iso, .bin/.cue, .mdf, .nrg or other image formats. I may build my own catalogue application at some point if I find the proper information and the spare time required. For the iso images, you may only need the ISO plug-in which can read - according to its readme - iso, bin, nrg, img, mdf file types. The latest (at the time of typing this) 1.7.7 beta4 seems to work in 9x but I haven't tested it extensively - just checked if it's loaded and opens an iso, which it did.

Please be very careful with those plug-ins - it's been a long time since I navigated the totalcmd.net repositiory and it seems many/most of the plug-ins are now Unicode-only, while their older 9x-compatible versions are nowhere to be found. I do have a few of the older versions scattered on CD/DVDs and who-knows-where else, if you need them. Here's the list of all the plug-ins I have installed currently; a few of them are new, haven't yet checked whether they work correctly, any bugs, etc but most of them have been there for years. I think Multiarc might be one of the most useful, together with Install Explorer and the other packers. For 7-zip, an old version by Adam Streleczki (0.5.5 here) is the only one I got to work and have had it installed - all the others at the repository are Unicode-only and will throw an error upon instalation. PACKER (WCX): iso TreeCopyPlus 7zip HA Multiarc bzip2dll MHTep DiskDirExtended ResExtract wipe InstExpl IShield FILE SYSTEM (WFX): httpbrowser (HTTP browser) ex2fs (Linux drives) PROCFS (Task manager) MirandaFS (Miranda database explorer) Back2Life (Erased files) Registry (Registry) wmdmplug (Media audio) LISTER (WLX): ICL View TTFViewer pdfview FlashView IEWebLister listdoc anytag DSView ? Imagine fileinfo Ansi CONTENT (WDX): super_wdx AudioInfo anytag Shortcut RarInfo FileDesc

There definitely is a Bluetooth driver that works in 9x. I had one driver package from a BT dongle installation mini-CD, then I found (maybe mentioned here in these forums?) two other rather huge packages (35MB and 62MB, respectively) that were distributed by D-Link and Emtec. All of these packages contain an inf with a long list of vendor and product IDs and software produced by Widcomm. But there's a catch: while the driver (namely btcusb.sys - do not mistake with btwusb.sys which does not work in 9x!) does work in 9x, the software - depending on version and package it's been installed from - may throw an "Evaluation mode" notice and limit transfer to 5MB!!! I wouldn't risk being banned by posting my blunt opinion about this practice but you get the idea. Thing is, the driver is useless without a working application for it, so this should be our main concern: finding a proper application that'd work with the respective driver and all supported devices. All of those packages bundle the BlueSoleil application by IVT, but not all versions work in 9x. I did find one that does, with the above-mentioned limitation. There's more to dig in related to this, maybe parts of the packages can be put together for a more complete set of options. On another note, I just stumbled into another interesting matter: Windows Easy Transfer. It's (among other functions) a driver for a special cable that allows USB-to-USB direct transfer between computers. The MS driver ony works with XP but maybe it can be made to work in 9x. The package is rather weird, many .man files (containing xml-formatted manifests). I found a link to it here.

Regarding searching into .iso (and other type of archives): jaclaz might not have tried Total Commander with a truckload of Packer (and Lister, when needed) plug-ins. When I have an iso and need to extract files from it, I open TC, double-click the iso, select the files by dragging with right-button down (as in old Norton Commander, for those who remember it) and then drag the selection to the folder open in the other panel. If there can be something easier than that...

Regarding that install order, one thing may be worth mentioning: at times, I've had trouble installing certain drivers (for video cards, usually but also sound and others) due to their dependancy on newer DirectX libraries. Therefore, to be on the safe side, I'd stretch to first installing the OS, then DirectX (latest supported version depending on OS - that is 8.0a for Win95 and 9.0c/2007 for 98SE/ME) and only then the drivers. But that's to be established on a per-hardware basis - just keep this in mind if anything doesn't seem to work quite right.

Dunno about Aida64, I've got an old version of Everest here (4.20 something). File > Preferences > Hardware monitoring > Sensor icons --> check all of them. And.. here we are:

That's precisely why I mentioned Ultimate Edition, not the Free edition. It can do: Motherboard temp, CPU temp, GPU temp, GPU memory temp, HDD temp, CPU fan speed, GPU fan speed, CPU core voltage and probably others, depending on the sensors installed in the hardware. It's been years since I last had Aida32 on my system and that's why I said I didn't know (remember) if or which of these functions it had. Talking about SpeedFan, I suddenly remembered the icon on the desktop and I vaguely recall having had some problems with it in the past; the version I have is 4.35, released sometime around 2008 and it doesn't recognize anything but the videocard and shows nothing at all.

Lavalys' Everest Ultimate Edition has all that and then some (i.e. GPU cooler speed), however that is a commercial application. I don't know if it's previous freeware version - the well known at the time Aida32 - has these functions or whether it can be found anywhere on the web for download. I have no knowledge of other freeware application that encompasses all the functions required. Something relatively close but not without quirks may be the old SmartGuardian by ITE, which has been shipped on drivers' CD as utility for some motherboards. It has the nasty habit of setting itself to autorun on startup even when that option is disabled in its Settings panel.

billyb, is that the same system that reads VEN_8086 (that is, Intel, not VIA) in the USB controller string? In post #1318, billyb stated his USB controller on the P4G533-LA motherboard in Compaq Presario 6350us appeared in registry as follows: HKEY_LOCAL_MACHINE\Enum\USB\ROOT_HUB20\PCI&VEN_8086&DEV_24CD&SUBSYS_80891043&REV_01&BUS_00&DEV_1D&FUNC_07 That is the hardware detection's result and cannot be fooled through drivers. I went to the HP/Compaq site, downloaded the supposedly VIA USB driver and upon opening it, I noticed two configuration files inside: - usb2x.inf - VIAEHCXP.INF The first one contains references to PCI\VEN_1033&DEV_00E0&REV_01 and PCI\VEN_1033&DEV_00E0&REV_02, both of them being manufactured by NEC and PCI\VEN_8086&DEV_24CD being manufactured by Intel. Looking a few rows above, we notice the latter matches exactly billyb's registry report. The Strings section at the bottom of the file states PCI\VEN_8086&DEV_24CD.DeviceDesc="Intel PCI to USB Enhanced Host Controller" Therefore, I conclude there is no doubt that billyb's controller is definitely manufactured by Intel, not VIA. The fact that the said driver can be found in a VIA package is probably more than just a coincidence - I suspect Compaq just had to put them together one way or another and they chose the VIA name as being more appropriate, that's all - however it is clear that the driver has nothing to do with VIA. The setup does not contain or manipulate usbhub.sys; actually, the configuration file mentions in the comments that they leave it on the system and actually use it. The setup does install usbehci.sys though, which is described as EHCI eUSB Miniport Driver manufactured by Microsoft. On the other hand, the second configuration file (VIAEHCXP.INF) refers to PCI\VEN_1106&DEV_3104 which by vendor ID 1106 and the line in the Strings section PCI\VEN_1106&DEV_3104.DeviceDesc="VIA USB 2.0 Enhanced Host Controller" we clearly see it belongs to VIA. However, since billyb's registry doesn't mention that vendor ID anywhere in the USB controller string, we shouldn't waste any more time on it. Now, are we clear that the so-much-debated USB controller on the P4G533-LA motherboard in Compaq Presario 6350us is manufactured by Intel and therefore should work with Intel/Microsoft drivers?

Dunno what to say... if I had $1, I'd buy myself something to eat. Seriously now, I do have something like a hub in form of a Lexmark X11xx combo (scanner + printer) which appears as a Generic hub in DM (which I did not mention in the list above). Just plugged it out and back in - nothing unusual. I also charge my phone's battery from this computer's USB (a Huawei U1220s) but that thing is not recognized since the "good guys" from the Far East didn't bother to build 9x drivers for it; still, it doesn't disturb the system in any way when un/pluggin it. As for billyb: man, the string says you got an Intel controller. That 8086 is their unmistakable brand (remember the 8086 CPU?). However, the SUBSYS section says it's being distributed by ASUSTeK (1043 seems to be their vendor ID). That particular SUBSYS combination does not appear in my database, but the main one clearly shows "USB EHCI controller by Intel Corporation". I have no idea how come the VIA driver works fine with that controller and the MS one doesn't...

Actually it is unusual. As you see in my previous post, all USB controllers are listed as VIA, whether it's 3038 or VT6202. Open RegEdit, go to HKEY_LOCAL_MACHINE > Enum > USB > ROOT_HUB (or ROOT_HUB20 if you only have a 2.0 controller and not a 1.0-1.1) and look at the strings that make the name(s), they should be in the form PCI&VEN_xxxx&DEV_xxxx&SUBSYS_xxxxxxxx&REV_xx and so on. What we're interested in is the VEN_ and DEV_ part, which make up the vendor ID and the device ID. If vendor ID is 1106, then you have a VIA controller. If it's 8086, it's an Intel. Can't remember the others right now but you may search the web for details if string has other identifiers.

OK guys, now that we stirred the pot, let me add some confusion in! Below you'll see my current configuration for the on-board VIA Tech 3038 USB 1.1 controller, to which a VIA VT6202 USB 2.0 PCI to USB Adapter adds. As you will notice, there is no single driver pertaining to VIA in this configuration, yet it works. Truth be told, I can't vouch for no BSOD ever or no other problems, just that I can't remember any, other than some composite devices such as the Bluetooth dongle not being recognized. For whoever has spare time and is willing to test such configuration, here you are: (3x) VIA Tech 3038 PCI to USB Universal Host Controller: - SYSTEM32\DRIVERS\UHCD.SYS 4.90.3000.1 (Microsoft Corp.) - SYSTEM32\DRIVERS\USBD.SYS 4.90.3000.1 (Microsoft Corp.) - SYSTEM32\DRIVERS\USBHUB.SYS 4.90.3002.1 (Microsoft Corp.) - SYSTEM\usbui.dll 4.10.2222 (Microsoft Corp.) - SYSTEM\hotplug.dll 4.10.2224 (Microsoft Corp.) - SYSTEM\VMM32.VXD (NTKERN.VXD) (3x) USB Root Hub: - SYSTEM\DRIVERS\USBHUB.SYS 5.00.2195.6689 (Microsoft Corp.) - SYSTEM32\DRIVERS\USBCCGP.SYS 5.1.2600.5585 (xpsp_sp3_qfe.080422-1455) (Microsoft Corp.) - SYSTEM\VMM32.VXD (NTKERN.VXD) VIA VT6202 USB 2.0 Enhanced Host Controller: - SYSTEM32\DRIVERS\USBPORT.SYS 5.00.2195.5652 (Microsoft Corp.) - SYSTEM32\DRIVERS\USBEHCI.SYS 5.00.2195.6882 (Microsoft Corp.) - SYSTEM\VMM32.VXD (NTKERN.VXD) USB 2.0 Root Hub: - SYSTEM32\DRIVERS\USBHUB20.SYS 5.00.2195.6891 (Microsoft Corp.) - SYSTEM\VMM32.VXD (NTKERN.VXD) All mass storage devices report as 'upperfilters': - SYSTEM32\DRIVERS\USBNTMAP.SYS 4.90.3000 (Microsoft Corp.) Probably used when mass storage devices are plugged in: - SYSTEM32\DRIVERS\USBSTOR.SYS 4.90.3000.1 (Microsoft Corp.)