HASH suggestion for v27

[Charles Dye]

HASH /S. Because sometimes you want to verify an entire tree.

 

[vefatica]

I don't think I knew about HASH. But I do think the following has been discussed before. The following statement (from HASH's help) is not true.

The hash value is a unique value corresponding to the content of a file. If two files have the same hash value and they're using SHA256, SHA384, or SHA512, then they have the same content (regardless of their file names or locations).

 

[Charles Dye]

Should be "are very likely to have".

 

[vefatica]

Should be "are very likely to have".

Probably should be left unsaid. Consider only 65-byte files (256^65 = 2^520 of them). There are 256^64 = 2^512 possible values for SHA512. That's an average of 256 files for each possible value of SHA512. ... hardly a good bet that two with the same SHA512 will be the same!

 

[Charles Dye]

Probably should be left unsaid. Consider only 65-byte files (256^65 = 2^520 of them). There are 256^64 = 2^512 possible values for SHA512. That's an average of 256 files for each possible value of SHA512. ... hardly a good bet that two with the same SHA512 will be the same!

And 255 times as many possible files with different hashes? Seems like pretty good odds to me.

 

[vefatica]

And 255 times as many possible files with different hashes? Seems like pretty good odds to me.

Huh? You were presuming equal hashes.

 

[AnrDaemon]

It is known that hashes could collide. The thing is that probability of deliberately crafting such a hash is insignificant for a given purpose.