krbknp.f.nbw.emd bad source

Endgame analysis using tablebases, EGTB generation, exchange, sharing, discussions, etc..
Post Reply
tdunham
Posts: 2
Joined: Mon May 29, 2006 5:42 pm
Sign-up code: 0

krbknp.f.nbw.emd bad source

Post by tdunham »

This file from td (me) does not match the md5s from other sources. I have been using Wilhelm to verify my downloads, and I have found and corrected a few in this manner. It would appear that I am doing something wrong, though, when I tried using md5sum that comes with cygwin on this file the sum did not match the sum posted on this site. I've disconnected from ed2k, when I have verified my other files with the cmd line utility I'll be back. I will reply to this thread if I find any other bad files. You may wish to recheck any files that you got from me, if one got through there could be more. Sorry for the bad source, and thanks to Watcher who pointed this out.
Tom
TapaniS
Posts: 26
Joined: Tue Mar 28, 2006 4:03 am
Sign-up code: 0
Location: Espoo, Finland

Post by TapaniS »

It is almost impossible to get errors with eMule, especially if we use links from Kirill's Online Project -site.

Even if somebody had defective source, eMule recognizes this by hashing. Wrong file will be ignored if asked hash and source hash doesn't match. The file name is not enough to eMule.

Please correct if I'm wrong.

Regards,
TapaniS
tdunham
Posts: 2
Joined: Mon May 29, 2006 5:42 pm
Sign-up code: 0

Post by tdunham »

Hopefully those factors will have prevented my bad copy from spreading much. My file matches the size from this site, but its md5 is 9423c... I've been using the ed2k links from here for getting tablebases since Hyatt's site stopped hosting the six man. My best theory currently is that some files were damaged while on my disk. I do check new files periodically with Wilhelm, I believe that this file had been checked before, but the md5 does not match now. I did crash hard enough recently to cause a chkdsk on startup, which is not so common on an ntfs partition. I have seen emule calculate the hash for previously shared files, but I don't know if this can happen during normal operation, or if it indicates that a file changed. I do share my file list with anyone who requests it. If you recently got a file from me in this manner, I would suggest checking it if you haven't. I've run md5sum on most of my six men and piped to a file, now I'm checking against the list here. If you have the following files from td, please make sure you have checked them, the md5s don't match: krnkrp.1.nbb.emd, krnkrp.3.nbb.emd, krnknp.e.nbw.emd, krnknp.2.nbw.emd, krnknp.0.nbb.emd, krnkbp.3.nbw.emd. If someone got one of these from me and they do have a good checksum, please reply, then I'd know for certain that they were mangled at some point after you uploaded.
Tom
User avatar
Kirill Kryukov
Site Admin
Posts: 7399
Joined: Sun Dec 18, 2005 9:58 am
Sign-up code: 0
Location: Mishima, Japan
Contact:

Post by Kirill Kryukov »

Hi Tom,

I think you don't need to worry much about spreading of corrupted files. eMule has pretty good corruption detection/recovery mechanism, as TapaniS said. Even if you managed to upload corrupted data to someone, everyone is strongly advised to check *all* incoming files with MD5 anyway, so they will hopefully spot the corruption and re-download. (Never happened to me yet).

So, just re-download the files that are corrupted, and it will be alright! Thanks for sharing, btw.

Just for case: most common reasons for file corruption in my experience (also from communication with other people): 1. Faulty hardware, especially memory. 2. Overclocking. 3. Overheating (strongly correlated with the previous two). 4. "Shared" folder is located on another machine and is accessed via windows network.

Happy sharing!
Post Reply