[SOLVED] HFSlip removes KS.SYS when used on Windows 2000 with DirectX9

[DeTard]

Geez... I'm glad at least SOMEONE else has seen the problem that I have with this. For those of you that say to check my hard drive, maybe you didn't read that I saw the exact same problem on two separate computers, on two separate networks (home vs work). If it weren't for that, I would have agreed that it may have been a corrupt file.

As for the EXE, yes you can extract from that no problem. I know that DXNT.CAB is not changed from bi-monthly redist to bi-monthly redist, but I actually believe it was corrupted on Microsoft's hard drive somewhere along the repackaging of the CABs. But it's DXNT.CAB itself that I cannot extract from (try extracting KS.SYS from this file and see if it's ok).

Thanks for everyone's checking on the file though; it's a fairly big download so thanks for the time you all took to try to confirm the problem. I'm not sure, but for those that have no problem with it, possibly try to clear your browser cache prior to downloading again to make sure when you download, it's not just pulling from cache. It's very possible that Microsoft had a good and a bad release as well, but why they would re-release I don't know.

Edited September 10, 2007 by DeTard

[Tomcat76]

johndoe74--

I tried that direct link... Same problem.

I've tested this with both 7-zip and WinRAR. Each program gives an error, even when just extracting the main exe.

[DeTard]

Ok, this is just wierd...

I just downloaded from johndoe's direct link and got the same MD5 hashes as I've always gotten. I really believe they updated the file for some reason, and the second release of it was bad. What I think I'm going to do now is get the June 2007 redist, use that, and then include the August 2007 files.

My file hashes I'm getting are:

DD9488CF38A8028EEF02B47D831A6353

Tomcat or RJARRRPCGP, can you confirm that hash?

[Tomcat76]

I can confirm the hash.

[DeTard]

Well, with hashes confirmed, you can safely rule out corrupted files during downloading. It's the actual file on their server that is an issue apparently. I can't give any explanation for those that it is working for other than cached copies of the file back when it did work ... since I guess they released it a second time?

For those that it does work for, how many August 2007 files are there in your copy? I have 6 in mine; I'm wondering if they "forgot" to include some August 2007 files and re-released it to include those, but managed to corrupt the contained CABs in the process.

[johndoe74]

ok here's one thing we can do, i can extract the installer and upload the required .CABs to my esnips or filefront accounts and u guys can download those and give 'em a shot. lemme know which files u want ? or do u want 'em all ?

[DeTard]

Honestly, I don't really need any of the files - I pulled them from the June 07 redist and added in the August 07 specific files from that redist, so all should be up to date now. Thanks for the offer though.

It was more of a surprise for me to run into finding a corrupt file on Microsoft (I, admittedly wrongly, assume that their files all are without CRC errors internally) and assumed that HFSlip was the problem.

[AgiHammerklau]

This may explain, why I ran into trouble, when I used the latest DX9 for the first time instead of the "ancient" version I´m integrating ever since...

Or maybe it´s just my old GF3ti200proVT (64MB), that is officially labeled DX8 ?

Agi

[Tomcat76]

The current RC for HFSLIP 1.6.3 aborts slipstreaming of DX9 if one of the cabs is corrupt; it does that by checking for the existence of KS.SYS after extraction. For Win2K, if there is a *DX9* hotfix present in the HF folder, HFSLIP will even abort completely because at that point the hotfix is already processed, which can't be undone (at least not easily).