Damaged Files Alert

[Kirill Kryukov]

Hi all,

In this thread, following Dan's idea, we will post warnings about damaged files found on eMule search. It's easy to see them - damaged file has same name and size with normal file, but different hash. Corrupted files usually have just one source, while the correct file should have more.

The first one: krppkr.2.nbw.emd, shared by "Auryn" and "French-Phil-;-)". If you are one of those two guys, please check that file, and download the correct one.

[French-Phil-;-)]

Hi,
corrupted file has been deleted from my computer.
Thank's.
Get all my best whishes for 2006.
French-Phil-;-)

[Kirill Kryukov]

Thanks, French-Phil-;-)!

More broken files:
krppkq.1.nbb.emd shared by Alb.
krppkq.0.nbb.emd shared by Cato the Younger.

Please take care!

[Kirill Kryukov]

Two more:
kbpkbp.1.nbw.emb and kbpkbp.5.nbw.emb, both shared by Alb.

[Alberto Gueci]

Thanks for warning Kirill, i'm checking everything again with md5sum.

My best whishes to all

Alberto

[guyhaw]

A suggested assurance-protocol for this project might be that:

1) Sharers indicate whether or not their file has been MD5sum-checked just before being shared

2) Downloaders (obviously) check whether the downloaded file MD5sum-checks.

I've certainly had a file fail this MD5sum-check after downloading: I think it was the 1 file out of 6 downloads, so I was still impressed with the ability of the infrastructure to download correctly.

g

[Kirill Kryukov]

I agree, and I can only add that everyone releasing new files should check md5 first. All files that I share are checked. One problem is that we don't have all original md5sum information, but someday we will fix it.

I've certainly had a file fail this MD5sum-check after downloading: I think it was the 1 file out of 6 downloads, so I was still impressed with the ability of the infrastructure to download correctly.

I downloaded some hundreds of EGTB files already, and all passed md5sum check, so far, so I'm pretty confident in infrastructure too..

[guyhaw]

I have all the MD5sum data for Eugene Nalimov's 3-to-5-man and 6-man pawnless EGTs.

Provenance: I think it came direct from Eugene via Rob Hyatt's site.

Action_1: I have asked Rob if he can make the MD5sum data he has available on his ftp site. I have also asked Rafael Andrist, who I presume has made it all part of WILHELM's MD5sum-checking ability.

Action_2: I will check my 6-man P-less MD5sums against what KK files as the 'MD5sums' when all the 6-man P-less data is indexed on the home page of the project.

g

[Kirill Kryukov]

Hi Guy,

I have all MD5 for all pawnless 6-men Nalimov bases. They are already available at the EGTB sharing page, at "md5" link near the "All together" for pawnless bases.

Direct links:
http://kd.lab.nig.ac.jp/chess/tablebase ... all-33.md5
http://kd.lab.nig.ac.jp/chess/tablebase ... all-42.md5

I have many md5 files for 33p and 42p, but many are still missing. The ones that I have are all available as well:
http://kd.lab.nig.ac.jp/chess/tablebase ... ll-33p.md5
http://kd.lab.nig.ac.jp/chess/tablebase ... ll-42p.md5

If you will have any md5 that I miss, by all means please send them to me and I will add them to the index page.

[guyhaw]

I have cross-checked 'my' files of MD5sum file-signatures ...
... received via Rob Hyatt's ftp site from Eugene Nalimov

with the files of MD5sum file-signatures made available on KK's site.

All MD5sum file-signatures agree for these files:
all 3-, 4- and 5-man, and 6-man pawnless (i.e., 3-3 and 4-2).

I don't have a second-source yet with which to x-check the MD5sum file-signatures as available at KK's site to date for 3-3p and 4-2p. Working on it.

Files of MD5sum file-signatures ... need their own MD5sum file-signature

g

[Kirill Kryukov]

Thanks for checking, Guy! I feel much safer now. I also hope you will find md5 for 'pawnful' bases. (those with pawns).

Files of MD5sum file-signatures ... need their own MD5sum file-signature

A zip-archive with CRC will do just fine.

[guyhaw]

Rafael Andrist sent me his file of Nalimov DTM EGT MD5sum data. This, or similar, is probably what he sent to Kirill before.

Not surprisingly, it agreed 100% with what I had from Kirill earlier, although MD5sums for some 6 4-2p EGTs were not included in it:
kbbbkp, kbbnkp, knnnkp, knnpkq, kqbpkr and krrrkp

KK has on file I believe MD5sums for all 3-3p EGTs that have been published via Hyatt. The main issue seems to be the 4-2p EGTs.

Rafael thinks the MD5sums originated with Hyatt rather than with Nalimov, i.e. one stop down the line. So it looks like Hyatt, and then Eiko Bleicher, are the primary and secondary sources of MD5sum data for the 4-2p EGTs.

g

[guyhaw]

Whether this has now been corrected or not, it is said that on a Chessbase DVD, the file krppkr.2.nbw.emd was found to be corrupt.

All other files checked out against MD5sums held in Andrist's WILHELM.

Does anyone know what the MD4 and MD5sums of this corrupt file were, and what they should be for the correct file?

g

[Bryan Hofmann]

I have downloaded the KBPKBP & KNPKNP tables from the emule project and both Wilhem and Crafty ID them as bad. Yet when I run a MD5 check sum they match the posted sums given on the web site. Any ideas here?


Bryan

[guyhaw]

Kirill's posted MD5sums against KBPKBP and KNPKNP agree with Rafael 'WILHELM' Andrist's, and he got them direct from Rob 'CRAFTY' Hyatt.

Both KBPKBP and KNPKNP EGTs are (> 10) well shared, so I wonder what MD5sum-check/WILHELM/CRAFTY experiences others have had.

The 'EGTs Online' homepage recommends you get the latest 1.49 version of WILHELM as earlier versions may incorrectly report corrupted files. Sounds like an incorrect MD5sum may have previously crept into (CRAFTY and) WILHELM. I have no memory of this.


Btw, Chessbase issued a corrupted krppkr.2.nbw file at one time. It may be 'out there' and offered for sharing.

g

[Wulff]

Hi!

There appears to be a damaged version of the file kqpkrr.3.nbw.emd in "our system"

The size of the file in question is approx. 66 MB.

If you have that version of the file, please delete it!