Visual Basic

13-Jan-15 11:39

SQL

Hi there,

does anybody know where to find a DJB2 Hashing algorithm code or lib for VB.NET?

Richard MacCutchan13-Jan-15 21:53

13-Jan-15 21:53

See http://www.cse.yorku.ca/~oz/hash.html[^], that is not exactly difficult to convert to VB.

SepPax14-Jan-15 6:33

14-Jan-15 6:33

Well,

guess you are right about it should not be difficult to be done, but I'm smashing my head on the wall since some time already, keep getting overflow errors and cast errors... WTF | :WTF:

I'm probably not that good in that kind of VB programming, that's why I'm asking if anybody suceeded in have it done and working... and available to share it.

Thanks in advance.

Richard MacCutchan14-Jan-15 6:58

14-Jan-15 6:58

Well, I'm not a VB programmer but I can probably help if you show your code and what errors you receive.

SepPax14-Jan-15 9:54

14-Jan-15 9:54

I have tried so many variants that I lost the path Sigh | :sigh:

BTW, the following is one of them

Public Function CRC_calc_crc_bernstein(frame As String, len As Integer)

    Dim crc As UInt16 = 5831
    Dim tmp As UInt16 = 0
    Dim t As UInt16 = 0

    For i = 1 To len

        tmp = Mid(frame, i, 1)

        ' t = (crc << 5) + crc + tmp
        t = (crc * 33) + 1  'this is another variant
        t = t + crc
        t = t + tmp

    Next

    CRC_calc_crc_bernstein = crc

End Function

The main error I get is that usually after few cycles of For I get an Overflow exception
I tried the use of Ctype() and DirectCast() without success, so... I'm almost giving up.

Thanks in advance for your kind help.

Richard MacCutchan14-Jan-15 23:44

14-Jan-15 23:44

Your logic is not the same as described in the algorithm (not sure why). Try:

Public Function Hash_calc(frame As String)

    Dim hash As UInteger = 5831
    Dim nextChar As Char
    Dim temp As UInt32
    
    For i = 0 To frame.Length() - 1
    
        nextChar = frame.Substring(i, 1)
        temp = Convert.ToUInt32(nextChar)
        hash = hash * 33 + temp
    Next

    Hash_calc = hash
    
End Function

I have changed the function name to Hash_calc as that is more meaningful, this is a hashing algorithm, not a CRC. I also changed the integer declarations to UInteger as that makes more sense for use in a 32-bit or 64-bit application. Finally you need to switch off the overflow checking in your projects properties page in Visual Studio.

SepPax15-Jan-15 10:37

15-Jan-15 10:37

Many thanks,

guess the point was about disabling the overflow checking (which I was not aware about).

I need just one more improvement, about getting the Hash_calc result to become a value between 0000h and FFFFh (a WORD). Which would be the correct approach?

Thanks once more.

Richard MacCutchan15-Jan-15 21:38

15-Jan-15 21:38

I guess just go back to using UInt16.

SepPax18-Jan-15 1:55

18-Jan-15 1:55

seems the right choice is Ushort.
I get results within FFFFh now. Thumbs Up | :thumbsup:

Just need to find out why those results are not matching the originator's hash, will debug further.

Richard MacCutchan18-Jan-15 2:17

18-Jan-15 2:17

Make sure that the originator is using exactly the same algorithm, including the hash seed value.

SepPax18-Jan-15 3:32

18-Jan-15 3:32

yes, will ask to set some debugs in order to check it.
The seed is identical, I guess it's something related to the results or maybe Ushort is not the right choice due to positive/negative calculation results.

SepPax18-Jan-15 9:02