Auto-Patcher For Windows 98se (English)

[Drugwash]

Redownload it. Prefferably using a download manager, like FlashGet or similar. Your archive is corrupt, for some reason.

[soporific]

just doing final tests now ... one of the tests is comparing the results of a full patch of a clean windows 98 re-installation using v1.95, and v1.96 and let me tell you, i already hate (or at least greatly dislike) v1.95 !! it's so slow !!! all the bloomin' reboots, how did you ever put up with it ?? I've been working and testing 1.96 for quite a while now and i've gotten used to the changes i've made. I wasn't joking that it's about 33% faster. It could be more ... i'll be uploading any second now ... and i'll post the benchmark results of the 2 versions ... i reckon the figures will be interesting ...

oh, and the next full version of this project will be called Auto-Patcher for Windows 98se (May 2007) and future versions will be known by the month they are released in. It is simply a coincidence that this is exactly how the other Auto-Patchers for Windows OSes are named. Really, it is - there will still be a version number to designate code changes.

Benchmarks: (using Win98 inside MS Virtual PC 2007 on a PentiumD 940 system)

AP version -- 1.95 ... 1.96

Test A ------- 3:02 .... 1:40 ... a 45% reduction

Test B ------- 43:07 ... 30:58 ... a 28% reduction -

test figures in minutes and seconds.

i reckon that averages out at about 33% --- like i predicted in the first post!

the number of reboots for 1.95 for test B was 21

the number of reboots for 1.96 for test B was 15

Test A: running thru default module options on a fully patched computer. Nothing is installed, so this mainly tests code optimization improvements, however, there is some distortion via the differences in the additions (mainly tweaks) and omissions (mainly hotfixes), but there are more additions than omissions, so 1.96 should probably take longer to run if the code was the same.

Test B: running thru the default module options on a just-installed Win98se system. This means everything selected gets installed.

i'll comment on all the results at the end. It looks good so far though, doesn't it ...

Edited May 13, 2007 by soporific

[celtish]

Redownload it. Prefferably using a download manager, like FlashGet or similar. Your archive is corrupt, for some reason.

FlashGet is good (as is a very similar programme called GigaGet) but they do throw up errors of this kind because they break downloads up into chunks and these chunks don't always stick together properly, particularly with large files. When I was downloading DirectX the other day it ground to a halt saying that file size had changed.

[soporific]

Yay, there's a sneak preview of the super duper and shiny new version 1.96 available!

grab it here:

http://soporific.dsleague.com/downloads/

size: 12.8 MBs

instructions:

just run the file, further instructions inside the installer.

[Drugwash]

Redownload it. Prefferably using a download manager, like FlashGet or similar. Your archive is corrupt, for some reason.

FlashGet is good (as is a very similar programme called GigaGet) but they do throw up errors of this kind because they break downloads up into chunks and these chunks don't always stick together properly, particularly with large files. When I was downloading DirectX the other day it ground to a halt saying that file size had changed.

Since version 1.8x, FlashGet tends to throw errors, once in a while, not to mention other bugs/unwanted behaviour (that I reported some time ago in their forum but for some reason it dissapeared) on a Win9x system. But I still use 1.73 on my Win98SE, which seems to work just fine, and was still available from their site last time I checked.

Soporific: I'll try to test the new version these days; currently I got work to do and a pretty tight deadline, so no promises. And I still have to build up that test machine, darn... (what do you think the run time would be, on a 166MHz Pentium I ? )

[briton]

The patcher decided my 48BITLBA was newer (there was a later version on the forum than the one included).

Yo Briton, can you tell me more about this ... which version of ESDI_506.PDR do you have installed?? (to find out, put a copy of the file you have installed into the \bin directory in the AP program directory, open up a DOS box at that location, and type GETVER ESDI_506.PDR and it will tell you the file version) ... i can't find any newer version of this update

OK. I did what you asked and it wouldn't help. So I researched and here is the short answer If you need more, let me know.

TINY answer: I used your GETVER on the new one (my one) and I get version 4.10.0.2226

VERY short answer: If your error check gives the existing file newer message when it discovers that the existing file is identical to the file AP will use to replace it, the problem is solved! (Sorry, I couldn't find where those file message reports are generated, but I am sure you know ) Basically, AP is finding that my file is the same date as the AP file even though the message says that the exisiting file is newer than the AP file.

SOMEWHAT LONGER answer with some details assumes that the message solution above is not the answer:

My version of ESDI_506.PDR is extracted from:

Enable4 8BitLBA | Break the 137Gb barrier!, Who said it couldn't be done? Enjoy your huge disks Post #1

Attached File(s)

4102226F.ZIP ( 14.1k ) Number of downloads: 1172

Manually inserted the extracted ESDI_506.PDR into the CAB file on the Win98SE disk before installation.

VERSIONS/DATES/SIZES

GETVER gives version 4.10.0.2226. So I looked to see where your fix was coming from to compare. I found two files in the Hotfixes directory:

• 48bitlba.exe
  
  
• 48bitlb2.exe
  
  

which from the batchfile tells me are the two versions for normal and IBM laptops. I opened them and extracted the ESDI_506.PDR files to the bin directory and renamed them ESDI_506.LBA and ESDI_506.LB2 resepctively. Then using GETVER ESDI_506.* provides this result:

• 4.10.0.2226 Esdi_506.lb2 - AP file from 48bitlb2.exe
  
  
• 4.10.0.2225 Esdi_506.lba - AP file from 48bitlba.exe
  
  
• 4.10.0.2226 ESDI_506.PDR - extract from 4102226F.ZIP (the one I used)
  
  

Just to check that my version and the lb2 version match and the lba version doesn't in other respects, I checked file dates/sizes:

• Esdi_506.lb2 - Saturday, July 29, 2006, 02:02:18 - 28,526 bytes
  
  
• Esdi_506.lba - Thursday, July 20, 2006, 12:38:32 - 24,431 bytes
  
  
• ESDI_506.PDR - Saturday, July 29, 2006, 02:02:00 - 28,526 bytes
  
  

Same match/mismatch. Conclusion if the newer file message answer above is not correct? Although I am CERTAIN that I disabled the IBM laptop item and enabled only the normal enable48bitlba (after all, these were the default options weren't they? and I checked anyway!), AP is trying to use the laptop version anyway! I doubt this!

Hope this helps - let me know if you want more!

[Drugwash]

I doubt there's any way of checking if one's using/installing this on a laptop or desktop. Unless there's a CPU check... So maybe there should be an additional validation for mobile/desktop installation, which would then make the difference between 2225 and 2226. And if that would be done, then this should be a global check, so any other mobile/laptop upgrade within AP would take this setting into account.

[briton]

Yo soporific! Check this out and see if I am an id***! (If I am, I'll delete it cos it will be in the way!)

Extracts from HFSystem.bat follow:

:48BITLBA
:: ERRORLEVEL 194
:: this is the version for ALL PCs/portables EXCEPT IBM portables with removable disks
SET ABOOT=D
SET TRACK=48BITLBA
SET TITLE=Unofficial Update for Drives over 137gb (not for IBM portables)
SET FILES=48BITLBA.exe
SET CHECK=%windir%\SYSTEM\IOSUBSYS\ESDI_506.PDR
SET CHEC2=4.10.0.2225
SET USECH=FILC
SET USEDP=none
SET SKIP!=OP0901
%COMSPEC% /E:2048 /C "%LOC8%\code\MainCode.bat" DIALOG1

:48BITLB2
:: this is the version for IBM portables with removable disks
SET ABOOT=G
SET TRACK=48BITLB2
SET TITLE=Unofficial Update for Drives over 137gb (for IBM portables)
SET FILES=48BITLB2.exe
SET CHECK=%windir%\SYSTEM\IOSUBSYS\ESDI_506.PDR
SET CHEC2=4.10.0.2226
SET USECH=FILC
SET SKIP!=OP0902
%COMSPEC% /E:2048 /C "%LOC8%\code\MainCode.bat" DIALOG1
IF EXIST "%PATHZ%\@_DoBoot.tra" goto BootNow
IF EXIST "%PATHZ%\@_GoBoot.tra" goto BootNow

Shouldn't the 48BITLBA section SET CHEC2=4.10.0.2226 and the 48BITLB2 section SET CHEC2=4.10.0.2225 ? If you, you have the final digit of the version number wrong in each case. Just swap them and I can test it again (if I haven't changed that and tested myself lol)

Let me know!

[soporific]

Shouldn't the 48BITLBA section SET CHEC2=4.10.0.2226 and the 48BITLB2 section SET CHEC2=4.10.0.2225 ? If you, you have the final digit of the version number wrong in each case. Just swap them and I can test it again (if I haven't changed that and tested myself lol)

Hmmm, i dunno how you missed it ... maybe i've misread the situation, but in all your post text explaining more about which update is installed there's no mention of the determining factor about which update to install:

:48BITLBA

:: this is the version for ALL PCs/portables EXCEPT IBM portables with removable disks

SET CHEC2=4.10.0.2225

:48BITLB2

:: this is the version for IBM portables with removable disks

SET CHEC2=4.10.0.2226

Now, either you have an IBM laptop with a removable disk and as such should have installed the 4.10.0.2226 version, or you don't and thus should have 4.10.0.2225 installed. My guess is you've somehow missed this extremely important piece of info and have assumed that the latest file version is the one to install. Or you actually have a laptop and just have neglected to tell us...

I just re-read your posts and i think this is all to do with the fact you have inserted the wrong file into your CAB files. Unless you actually have an IBM laptop, you should have inserted 2225. This totally explains why AP is giving you the messages you are seeing. I dearly hope you aren't using that on a computer with hard disks over 137 gb! Stop using immediately!

And i did include a little bit of text that explained which one was which in the module options... i think you've had a Homer moment, my friend! D'OH!

about the sneak preview of v1.96: this is not in anyway an official release, a sneak preview in my language is a Beta where the code changes are finished but there still may be update additions to come. What i'm trying to say is there will be an actual v1.96 upgrade official release that you should download and install to replace the sneak preview version, so don't go and think you have v1.96 cuz you don't!

Edited May 14, 2007 by soporific

[briton]

Hmmm, i dunno how you missed it ... maybe i've misread the situation...

OK. The reason I would have missed it is easy.

Years ago, I came to MSFN and checked out a thread which had started in which the idea of a Win98 Service Pack was being mooted. Not developed, mooted. So I went off and developed a purely Microsoft SP for my own uses with no other fixes that weren't available on the Update Site.

This year, I was pointed here by someone on another forum regarding the >137GB problem. I first read the thread in which its development was first mooted and then discussed (and argued especially with regard to a commercial version which already existed). That thread seems to have died when the final result was produced. When I understood THAT thread, I then went to the thread from which I could download the end result. I searched to see why the different versions/version numbers were not explained at the point of the download links. I then took the latest version and used it. It worked. I also compared it with the earlier version whose version number matched my Win98SE version number.

I posted a question on that thread asking which version number I should use (and should I match my Windows version number) as I hadn't found anything very clear. There hasn't been an answer. Maybe I should have reposted, but as the file was working, I didn't chase it. Then when I used auto-patcher, my Windows version number changed so I made the logical assumption that latest version was the one to use and only to revert to an earlier version if there was some reason not to.

Now it seems I was supposed to read some other thread (or this one?) to find out which file to download from the sticky post which held the download links.

So you are saying I should have been using the SECOND TO LAST download version? Hmmm! Wonder why my PC is alive considering my Win98SE partition is entirely above 137GB lol

Thanks for pointing it out. Any chance of someone doing something in that sticky post to point out which file to download for which use e.g. "this latest version is for IBM compatible laptops only and NOT for PCs"? I mean people who want that file don't necessarily read other threads!

Sorry to have reported incorrectly. Still very puzzled but thanks for the pointer!

[briton]

soporific there is something odd going on here.

My Win98SE version starts by installing KERNEL32.DLL version 4.10.0.2222 which is in the WIN98_31.CAB file on the CD and I copied to my windows\options\cabs folder from where I instal Win98SE.

ESDI_506.PDR comes from the same CAB file and was originally presumably the same version as the kernel. That is why I wasn't sure which Enable48bitLBA zipfile to download - the .2222 version zipfile or the latest.

Now according to the acquired wisdom from both this thread and that download thread I should be having startup problems because this is DEFINITELY not any kind of laptop - I ought to know, I built it lol

Now if I understand the (not very apparent) guidance on the other thread, one should ONLY download the 2226 version if your ESDI_506.PDR file has that version number - which the average mortal will only know if they find that their KERNEL32.DLL or Windows 98SE version number ends 2226.

Now come the problems. I manually replace the 2222 version of ESDI_506.PDR with the 2226 version and, luckily, have no problem with startup so I can finish the installation. Later on, I discover that Autopatcher is worth a try so I use it and include the 98->Me upgrade.

Now I find that sitting in my Windows\Options\Cabs directory is a KERNEL32.DLL version 4.10.0.2224 which I didn't put there and couldn't have come from my Win98SE CD. If that were being used, I should still NOT be using the 2226 version of the ESDI_506.DLL file. BUT in my Windows\System directory, the KERNEL32.DLL has version 4.10.0.2226. This of course means that in using the version 2226 version of ESDI_506.PDR, I am KIND OF following LLXX's advice (before she got banned) to use that version of the file .......

Question - how on earth did something somewhere decide to make my Windows 98SE version 4.10.0.2226 instead of 2224 ot 2222 and somewhere maybe even skipping 2225?

And if this is what happens when using the autopatcher and 98->Me, then the advice about IBM laptops is not correct but rather should use the Windows version number as identifier.

Any ideas what is going on here and what I should do? I am prepared to do a clean reinstallation, but I am not going to move the partition which means I need to replace the ESDI_506.PDR file during the windows installation. Presumably I should use 2225 for that even though the system will be 2222. And I can stop at each stage and check versions of kernel32.dll and esdi_506.pdr (the latter using getver.exe). If I need to test like that, should I disable 98->Me when I autopatch the first time so I can check the versions before and after?

Any thoughts!

[Drugwash]

The kernel version and the IDE driver version do not necessarily have to match.

The kernel may be (and most likely has been, through Auto-Patcher) patched/replaced by certain (un)official updates/upgrades, according to some versioning schema that follows MS' style, but that doesn't imply that any and all other files should be upgraded to the very same version number, if available.

In this case, the 2226 version of the pdr driver is clearly only for IBM laptops, so desktop users should stay away from it always.

Also, the official 2225 version of the pdr driver is coming from MS and is an upgraded/fixed version of 2222, meant to officially replace the latter. The patched 2225 version from LLXX is the very same official 2225, patched to support 48bit LBA. So it's safe to install LLXX' 2225 version over a 2222 pdr driver, on a desktop installation.

Someone please correct me if I'm wrong...

[soporific]

soporific there is something odd going on here.

I entirely agree with you, it sounds like crazy stuff ... !

And if this is what happens when using the autopatcher and 98->Me, then the advice about IBM laptops is not correct but rather should use the Windows version number as identifier.

OK, i think the warning to not use 2226 if you have a standard desktop is a theoretical warning - ie its possible something could go wrong, not that it will always go wrong. So just because you are using it without a meltdown of some sort doesn't mean its safe. I wish I knew more precisely but that is how i currently understand it. I remember people reporting that 2226 worked on their machines without a fuss.

Any ideas what is going on here and what I should do? I am prepared to do a clean reinstallation, but ...... Any thoughts!

Why go to the bother of re-installing if this is your only issue currently? It seems like you know how to boot into command prompt safe mode. Put the 2225 file somewhere easy ( eg c:\ ) and copy 2225 to %windir%\SYSTEM\IOSUBSYS\ eg c:\windows\system\iosubsys overwriting the 2222 file. All done. Am i missing something here?

OTHER NEWS

I have had a report of a problem in the Essential 3rd Party Apps module with the latest Flash update (i said i wouldn't include, i lied!) ... has anybody else downloaded the 1.96 preview and tried to run the above module? Did you by-pass Flash? If not, did you get the result below? Thanks in advance for any help tracking this down...

Now searching through the Essential 3rd Party Apps module...

ASET: Warning: syntax error!

*** program code error ***

Track: Flash4ie

Title: Flash 9.0.45 player & plugin for Internet Explorer

TvarB1: 9 0 45 0

TvarA1:

TvarB2: 9

TvarA2:

TvarR1:

TvarR2: ?

You are at the CantDetr label.

Please report this.

This update DOES use a new routine i dreamed up just recently ... it uses the registry to determine the program version. Lots of programs actually have a registry key that is used for advising on the program version --- Flash 9.0.45.0 installs a key called "FlashPlayerVersion" in HKCU\software\Macromedia\FlashPlayer and so the code simply uses that as the comparison version and the normal file version routine is then called to make the decision about whether Flash needs to be installed or not. The strategy is a bit iffy, especially if during a program's historical development a different method is used to store the program version - but it should work for Flash AFAIK ... so please report if you have seen something like the above ...

Edited May 15, 2007 by soporific

[charly]

OTHER NEWS

I have had a report of a problem in the Essential 3rd Party Apps module with the latest Flash update (i said i wouldn't include, i lied!) ... has anybody else downloaded the 1.96 preview and tried to run the above module? Did you by-pass Flash? If not, did you get the result below? Thanks in advance for any help tracking this down...

Now searching through the Essential 3rd Party Apps module...

ASET: Warning: syntax error!

*** program code error ***

Track: Flash4ie

Title: Flash 9.0.45 player & plugin for Internet Explorer

TvarB1: 9 0 45 0

TvarA1:

TvarB2: 9

TvarA2:

TvarR1:

TvarR2: ?

You are at the CantDetr label.

Please report this.

Yes--got the same thing--charlie

[briton]

OK thanks very much Drugwash and soporific. You both made it very clear.

I knew the warning about the IBM laptop version was theoretical - as in "it MAY cause problems".

However, there is still a problem with the way in which the files are provided. The Enable48BitLBA Download Thread simply states:

2006-07-29 20:05 Win98SE version (4.10.2226) released - only for replacing an existing v4.10.2226 file

As that is a sticky post and the thread is only sporadically posted to, this means that there will be users like me who, finding they have version 4.10.2226 will get stuck wondering. Many of them will NOT know how to find out the version of the .pdr file but WILL know how to find their version number.

Can't change the sticky. Not much point burying the advice in a quiet thread. My point is that it would probably be a good idea for a Moderator to amend LLXX's sticky post to explain that the 2226 version should ONLY be used if the PDR file has that version number regardless of what the kernel/windows version is and maybe a pointer to how to use GETVER to find out (I notice that GETVERS doesn't work either). Remember that LLXX cannot amend that post for obvious reasons so could someone else please? (You might also want to make it more obvious which files to download - like put the type of windows against the file link - the number of downloads suggests that many like me downloaded 2226 thinking it was simply the latest version - in fact EVERYONE seems now to be downloading that version and I doubt there are that many IBM laptops around lol)

For me, that's it - I am back on track thanks!

Edited May 15, 2007 by briton