Level Extreme platform
Subscription
Corporate profile
Products & Services
Support
Legal
Français
CRC32 returns negative on big files
Message
From
01/08/2011 03:39:33
 
 
To
01/08/2011 03:13:36
General information
Forum:
ASP.NET
Category:
Coding, syntax and commands
Environment versions
Environment:
VB 9.0
OS:
Windows 7
Network:
Windows 2003 Server
Database:
MS SQL Server
Application:
Web
Miscellaneous
Thread ID:
01519635
Message ID:
01519775
Views:
19
>>>>>>>>>(1) The short way is to cast the return result to UInteger, and change the return type of GetCrc32() to UInteger
>>>>>>>>
>>>>>>>>IIRC you can't do that cast in VB (although you can in C# using 'unchecked')
>>>>>>>
>>>>>>>
>>>>>>>Viv,
>>>>>>>
>>>>>>>I don't know, but I'd be very surprised if it weren't possible
>>>>>>
>>>>>>Just tried this which won't compile:
Dim test As UInteger = DirectCast(Int32.MaxValue, UInteger)
>>>>>>If it was a valid cast what resulting UInteger would you expect from a negative value signed int ?
>>>>>>In C#:
unchecked { UInt32 test = ((UInt32)Int32.MinValue); }
>yields 2147483648....
>>>>>
>>>>>
>>>>>Firstly, I'm surprised that it yields an even number
>>>>>
>>>>>Secondly, if that's the case, I guess you'll need t check whether is was negative, and if so, set the left most bit after the 'cast'
>>>>
>>>>But if the signed int is negative then there is no accurate equivalent for an unsigned int ?
>>>>I'm assuming the C# cast simple treats the 32 bits as being an unsigned value ?
>>>>FWIW:
unchecked { Int32 test = ((Int32)UInt32.MaxValue);
>gives -1
>>>
>>>
>>>Yes, of course, all bits set is signed -1 and unsigned the max value
>>>
>>>When the leftmost bit is set in a signed int type, then it is considered as negative
>>>
>>>>But if the signed int is negative then there is no accurate equivalent for an unsigned int ?
>>>
>>>Viv, this isn't about maths. The crc32 is just a structure that is 32 bits wide
>>
>>That's why I said : "I'm assuming the C# cast simple treats the 32 bits as being an unsigned value"
>>
>>>If you think about it as a signed 32 bit structure and the leftmost bit is set then the number is negative
>>>
>>>If you look at it as an unsigned 32 bit structure, the same pattern is a number >= 0
>>
>>I understand that - my point was that, in normal application usage, converting a negative signed int to an unsigned int would not make sense.
>
>
>Well, if the cast does not work, I think I have figured out how to - pardon the syntax - and untested
>
>If the number is not negative, I see no problem in converting

What if the unsigned positive value > Int32.MaxValue() ?

>
>If it is, then
> set the sign bit to 0
> convert to UInteger
> add 0x80000000UI
>endif
>
>
>
>Private Function GetCrc32(ByRef toStream As System.IO.Stream) As UInteger
>
>   '  the code
>   ' Return Not lnCRC32Result
>   dim result as UInteger
>   if( lnCRC32Result >= 0 )  then
>       result = CUInt( lnCRC32Result)
>  else
>       result = CUInt(lnCRC32Result and &7FFFFFFF) + &80000000UI
>  endif
>   return not result 
>
Previous
Next
Reply
Map
View

Click here to load this message in the networking platform