fao Greg Turgeon re SHA-1

[Doug Gamble]

There is a third possibility, I tested it wrong. I went back and
carefully examined the program I was using to test the code and
find I was looking at the same test file twice! The code Semen
posted works fine!

Sorry for the scare.

------------------

[Greg Turgeon]

 [quote]The problem with SHA-1 is if you take 2 different strings that
are only slightly different, they will return the same hash.[/quote]
   
Only two possibilities exist. 1) The implementation is buggy.  
2) You've found a problem that has escaped everyone from the NSA on 
down, including every reputable security expert worldwide. 
   
I'd say that #1 is more likely.

------------------
-- Greg
[email protected]

[Adrian C Aitken]

How similar do you mean? using the ASM code (posted in source code forum for a nice person to translate into PB compatible ASM ;-] whilst I jet off to New Zealand for 3 weeks) I can use two text files with only 1 character different. The results are completely and utterly different. This is what is supposed to happen with hashes. I don't know which version of SHA-1 you used.

Regards

Adrian Aitken


------------------

[Semen Matusovski]

Doug --
The problem with SHA-1 is if you take 2 different strings that
are only slightly different, they will return the same hash.

Can you give a link with results of investigations ?
If this is true, even CRC-32 works better.



------------------
E-MAIL: [email protected]

[Doug Gamble]

The problem with SHA-1 is if you take 2 different strings that
are only slightly different, they will return the same hash.
Try using the RIPE-160 hash posted to this board previously, it
will return a differnt hash for two 20kb strings that only have
one character different.

I'm not sure if the same problem exists in the SHA-256 code.

------------------

[Semen Matusovski]

This code is based on original document.
Of course, there are serious reasons to rewrite it using inline ASM.

   #Compile Exe
   #Register None
   #Dim All

   Function f(t As Long, B As Long, C As Long, D As Long) As Long
      Select Case t
         Case <= 19: Function = (B And C) Or ((Not B) And D)
         Case <= 39: Function = (B Xor C Xor D)
         Case <= 59: Function = (B And C) Or (B And D) Or (C And D)
         Case <= 79: Function = (B Xor C Xor D)
      End Select
   End Function

   Function K(t As Long) As Long
      Select Case t
         Case <= 19: Function = &H5A827999&
         Case <= 39: Function = &H6ED9EBA1&
         Case <= 59: Function = &H8F1BBCDC&
         Case <= 79: Function = &HCA62C1D6&
      End Select
   End Function

   Function S(n As Long, X As Long) As Long
      Dim t1 As Long, t2 As Long
      t1 = x: Shift Left t1, n
      t2 = x: Shift Right t2, 32 - n
      Function = t1 Or t2
   End Function

   Function SHA(Str As String) As String
      Dim i As Long, j As Long, t As Long, kg As Long, nb As Long
      Dim H(0 : 4) As Long, w(0 : 79) As Long
      Dim A As Long, B As Long, C As Long, D As Long, E As Long, Temp As Long
      Dim AA(0 To 3) As Byte
      
      kg = Fix((Len(Str) + 8) / 64) + 1: nb = 16 * kg
     
      ReDim ww(0 To nb - 1) As Long
      ReDim bww(0) As Byte At VarPtr(ww(0))
      ReDim bwc(0) As Byte At StrPtr(Str)

      For i = 0 To Len(Str) - 1: bww(i) = bwc(i): Next: bww(Len(Str)) = &H80
      ww(nb - 1) = Len(Str) * 8
      
      For i = 0 To nb - 3
         For j = 0 To 3: aa(3 - j) = bww(i * 4 + j): Next
         For j = 0 To 3: bww(i * 4 + j) = aa(j): Next
      Next

      H(0) = &H67452301&
      H(1) = &HEFCDAB89&
      H(2) = &H98BADCFE&
      H(3) = &H10325476&
      H(4) = &HC3D2E1F0&

      For i = 0 To kg - 1
         For j = 0 To 15: w(j) = ww(i * 16 + j): Next
         For t = 16 To 79
            W(t) = S(1, W(t-3) Xor W(t-8) Xor W(t- 14) Xor W(t-16))
         Next
         A = H(0): B = H(1): C = H(2): D = H(3): E = H(4)
         For t = 0 To 79
            Temp = S(5, A) + f(t, B,C,D) + E + W(t) + K(t)
            E = D: D = C: C = S(30, B): B = A: A = Temp
         Next
         H(0) = H(0) + A: H(1) = H(1) + B: H(2) = H(2) + C: H(3) = H(3) + D: H(4) = H(4) + E
      Next
      Function = Hex$(H(0))  + Hex$(H(1)) + Hex$(H(2)) + Hex$(H(3)) + Hex$(H(4))

   End Function

   Function PbMain
      If SHA ("abc") = "A9993E364706816ABA3E25717850C26C9CD0D89D" Then MsgBox "ok1"
      If SHA ("abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq") = _
        "84983E441C3BD26EBAAE4AA1F95129E5E54670F1" Then MsgBox "ok2"
   End Function

------------------
E-MAIL: [email protected]

[Adrian C Aitken]

thanks greg, I've posted assembler versions of sha-1 and Rijndael (both freely distributable), just need some one who knows how to handle strings,bytes etc in 32-bit world. I only did .com programs (before .coms were massively profitable I might add !!!!)


Adrian

------------------

[Greg Turgeon]

Adrian--
   
Originally, I coded SHA-1 for the PB/DOS compiler.  I posted this 
code in a few places, including on the old PowerBASIC BBS, but I have 
no idea how far it's now travelled.  Since then I've sent it to 
people who have requested it. 
   
I've also coded SHA-1 for the PB 32-bit compilers.  I did so the 
first time for a client, which of course left me with code that I 
couldn't freely distribute, considering that someone had paid me to 
write it.  Since then I've revised the code extensively so that I can 
supply it, as a DLL, to clients for whom I work on larger projects.  
But once again, I find myself with code that I'd be uncomfortable 
distributing freely.
   
I'd very much like to redo the algorithm yet again, probably as an 
implementation of the recently released 256-bit SHA standard.  I'd 
write this code with the intention of posting it here.  All I need is 
the 60-hour day, which, surprisingly enough, neither major US 
presidential candidate has yet promised to initiate. 
   
Some people might recall that I used to post lots of stuff, back when 
I had more free time to do so.  I always tried to put together the 
type of posting which, I hope, offered code that suggested 
possibilities that people might not otherwise have considered.  I now 
have SHA/256 on my list of projects, and I'll be happy to post the 
results once they exist.

------------------
-- Greg
[email protected]

[Adrian C Aitken]

Hi Greg (or anyone else for that matter!)
You posted a message in 1999 about source code for SHA-1 - I don't have a Compuserve account so could you (or any else) please either email it to me or post it on the source code forum?

Regards

Adrian Aitken
[email protected]

------------------