C# rhash generates hashes different than the rhash.exe and utorrent

Question

I am using C# with RHash in order to calculate the btih hashes of of file. Currently I'm using 3 tools in order to generate the btih hash:

1. rhash-1.2.9-src\bindings\mono with librhash-1.2.9-win dll
2. rhash-1.2.9-win32 command line tool
3. uTorrent

The problem is that every tool generates different btih signatures for the same file (the photo was taken by me, it's royal free):

1:  2FF7858CC0A0B216C3676A807D619FA30101E45F
2:  E6F07BB3C3B3B67531C84E3452980698AC1B0DAA  A:\IMG_0400.JPG
3:  D0B96839A14A8C45BB81AD157805AE73425998E5

For the C# hash generation I use Hasher.GetHashForFile(f.Name, HashType.BTIH); and rhash --bith in the cmd tool.

What am I doing wrong? Is there another way to calculate the btih?

Aftermost

are you sure you use the right encoding? show the code by that you read the file and encrypt it – Katha

Encoding? those are binary files... the rhash binding uses simple FileStream (see line 163 and 77 of the sorcecode github.com/rhash/RHash/blob/master/bindings/mono/Hasher.cs) – Aftermost

thought you might have read the file it into a string – Katha

Answer 1

The difference between the first two is that, according to the RHash source code, BTIH hashes requires additional data to be correctly computed.

The init_btih_data function in calc_sums.c is documented with:

Initialize BTIH hash function. Unlike other algorithms BTIH requires more data for correct computation.

In test_hashes.c, the BTIH examples are actually treated differently depending on whether USE_BTIH_WITH_TEST_FILENAME has been defined.

That init_btih_data function (which only seems to get called when running the command line application) in turn calls the rhash_trasmit function a number of times depending on various parameters. At a minimum, it will call it twice, which probably explains the difference between the first two. It can, however, call it a number of other times, which I think explains the difference we see with uTorrent.

The difficulty is that while the unmanaged DLL exposes the rhash_trasmit function, the .NET bindings do not, which means that it is not possible to supply the additional data that is expected.

Mary