Silverlight MD5 implementation

NEW POST: Different results compared to MD5CryptoServiceProvider

ReidB — Thu, 04 Jun 2009 20:32:41 GMT

Thank you for your bug report. This issue has been fixed in the v2 release, now available. Additionally, v2 adds better stream support for use on large data sets with minimal overhead. Please upgrade to v2 as soon as possible.

CLOSED ISSUE: Bytes more than 56 will get the wrong result

ReidB — Thu, 04 Jun 2009 20:28:36 GMT

code: md5.cs

in: //CHUNK IMPLEMENTATION

when do the copy from input to working, always start from the beginning is wrong!
The correct way is:

if (i == chunks)
{...
Buffer.BlockCopy(input, i * 64, working, 0, remainder);
...
}
else if ((0 > remainder) && ((i + 1) == chunks))
{
Buffer.BlockCopy(input, i * 64, working, 0, 64+remainder);
...
}
else
{
Buffer.BlockCopy(input, i * 64, working, 0, 64);
...
}

for those buffer less than 56 Bytes, the wrong code can get the correct result because:
56 + 64 + 8 = 128 = 2 * 64

Comments: Bug has been fixed in Silverlight MD5 v2. Please see the downloads tab for the fixed code.

UPDATED RELEASE: Silverlight MD5 Implementation v1 (Nov 18, 2008)

Thu, 04 Jun 2009 20:27:41 GMT

UPDATED RELEASE: Silverlight MD5 Implementation v2 (Jun 04, 2009)

Thu, 04 Jun 2009 20:27:22 GMT

UPDATED RELEASE: Silverlight MD5 Implementation v2

Thu, 04 Jun 2009 20:26:43 GMT

UPDATED RELEASE: Silverlight MD5 Implementation v2

Thu, 04 Jun 2009 20:26:28 GMT

NEW POST: Different results compared to MD5CryptoServiceProvider

EricSmith — Tue, 12 May 2009 18:10:40 GMT

On the home page, carstenh asserted that this implementation has a bug. I've duplicated those results.

It seems like there might be a problem when the length of hashed data reaches 120 bytes. In my tests, hashes up to that length are the same as those produced by MD5CryptoServiceProvider, but longer ones started to be different.

COMMENTED ISSUE: Bytes more than 56 will get the wrong result

KeithKoons — Fri, 06 Mar 2009 19:47:00 GMT

code: md5.cs

in: //CHUNK IMPLEMENTATION

when do the copy from input to working, always start from the beginning is wrong!
The correct way is:

if (i == chunks)
{...
Buffer.BlockCopy(input, i * 64, working, 0, remainder);
...
}
else if ((0 > remainder) && ((i + 1) == chunks))
{
Buffer.BlockCopy(input, i * 64, working, 0, 64+remainder);
...
}
else
{
Buffer.BlockCopy(input, i * 64, working, 0, 64);
...
}

for those buffer less than 56 Bytes, the wrong code can get the correct result because:
56 + 64 + 8 = 128 = 2 * 64

Comments: Thank you for this fix. I just spent a couple hours tracking this down myself, guess I should have checked here first :)

UPDATED RELEASE: Silverlight MD5 Implementation V1 (Nov 18, 2008)

Tue, 03 Feb 2009 19:23:13 GMT

UPDATED RELEASE: Silverlight MD5 Implementation V1 (Nov 18, 2008)

Tue, 03 Feb 2009 19:22:46 GMT

EDITED ISSUE: Bytes more than 56 will get the wrong result

ReidB — Tue, 03 Feb 2009 19:20:33 GMT

CREATED RELEASE: Silverlight MD5 Implementation V1 .1

Tue, 03 Feb 2009 19:19:05 GMT

CREATED ISSUE: Bytes more than 56 will get the wrong result

OneyWang — Tue, 20 Jan 2009 09:17:06 GMT

NEW POST: Case of Hash

ReidB — Thu, 11 Dec 2008 22:26:19 GMT

Unfortunately there is no standard method for returning a hash in ASCII encoded hex. Because this issue isn’t addressed in RFC 1321, I had to make an arbitrary decision as to the best way to represent the stream. Since the default BitConverter will convert a bitstream to ASCII encoding using uppercase characters, so I believed that this behavior would be more expected by users of the .NET runtime.

Whenever you are comparing two ASCII encoded hashes of different systems, it’s important to normalize the strings before doing comparison because of the lack of standard string representations. For your problem, the following code change will allow you to generate hashes in lower case:

Substitute all instances of the following:

string retval = BitConverter.ToString(GetHash(input));
retval = retval.Replace("-", "");

With:

string retval = BitConverter.ToString(GetHash(input));
retval = retval.Replace("-", "");
retval = retval.ToLower(CultureInfo.InvariantCulture);

Hope this helps!

Reid

NEW POST: Case of Hash

InquisitorJax — Wed, 10 Dec 2008 12:06:14 GMT

Shouldn't the MD5 hash be in lower case?

Currently GetHashString method is returning an uppercase hash, which would obviously fail a string comparison.

UPDATED WIKI: MD51Managed()

ReidB — Mon, 08 Dec 2008 23:16:14 GMT

MD5Managed Constructor

Initializes a new instance of the MD5Managed Class.

public MD5Managed()

UPDATED WIKI: Initialize()

ReidB — Mon, 08 Dec 2008 23:14:59 GMT

MD5Managed.Initialize Method ()

Initializes an instance of MD5Managed

public override void Initialize()

UPDATED WIKI: MD5Managed

ReidB — Mon, 08 Dec 2008 23:05:12 GMT

MD5Managed Class

Computes the MD5 hash value for the input data. This class is intended to be used in an object oriented manner. This is similar to the MD5CryptoServiceProvider class in the full .NET runtime, and is suitable for use in the CryptoStream infrastructure.

Any public static (Shared in Visual Basic) members of this type are thread safe. Any instance members are not guaranteed to be thread safe.

MD5Managed Constructor

Name	Description
MD51Managed()	Initializes a new instance of MD5Managed

MD5Managed Methods

Name	Description
Initialize()	Initializes an instance of MD5Managed
ComputeHash(byte[])	Computes the hash value for the specified byte array. (Inherited from HashAlgorithm.)
ComputeHash(stream)	Computes the hash value for the specified Stream object. (Inherited from HashAlgorithm.)
ComputeHash(byte[], int, int)	Computes the hash value for the specified region of the specified byte array. (Inherited from HashAlgorithm.)
HashCore(byte[], int, int)	Routes data written to the object into the MD5Managed hash algorithm for computing the hash. (protected)
HashFinal()	Returns the computed MD5 hash value after all data has been written to the object. (protected)

UPDATED WIKI: MD5Core

ReidB — Mon, 08 Dec 2008 03:38:47 GMT

MD5Core Class

Computes the MD5 hash value for the input data. This function is intended to be used in a static manner, MD5Managed can be used for an instance object.

This class is thread-safe.

MD5Core Methods

Name	Description
GetHash(byte[])	Computes the hash value for the specified byte array, and returns a byte array.
GetHash(string)	Computes the hash value for the specified string, and returns a byte array.
GetHash(string, Encoding)	Computes the hash value for the specified string using the provided encoding, and returns a byte array.
GetHashString(byte[])	Computes the hash value for the specified byte array, and returns a string.
GetHashString(string)	Computes the hash value for the specified string, and returns a string.
GetHashString(string, Encoding)	Computes the hash value for the specified string using the provided encoding, and returns a string.
r1(uint, uint, uint, uint, uint, int, uint)	Performs the first round and auxiliary [F()] functions. (private)
r2(uint, uint, uint, uint, uint, int, uint)	Performs the second round and auxiliary [G()] functions. (private)
r3(uint, uint, uint, uint, uint, int, uint)	Performs the third round and auxiliary [H()] functions. (private)
r4(uint, uint, uint, uint, uint, int, uint)	Performs the fourth round and auxiliary [I()] functions. (private)
LSR(uint, int)	Implements left-shift-and-rotate. (private)
Converter(byte[])	Converts an input byte array into an array of uints. (private)

UPDATED WIKI: GetHashString(string, Encoding)

ReidB — Mon, 08 Dec 2008 03:38:19 GMT

MD5Core.GetHashString Method (String, Encoding)

Computes the hash value for the specified string using the provided encoding, and returns the human readable representation of the output.

public static byte[] GetHashString(
     string input,
     Encoding encoding
)

Parameters

input
Type: System.String
The input to compute the hash code for.
NOTE: The input string is treated as a sequence of bytes. This means differences such as language, capitalization, and even punctuation will result in completely different outputs.

encoding
Type: System.Text.Encoder
The encoder to use to decode the input string into a series of bytes.

Return Value

Type: System.String
The computed hash code, represented as a string of upper case hexadecimal pairs, for instance "9A06BECFED4DFCC83F66406A1DB1BC6C" in the case of the input "Silverlight".

Exceptions

Exception	Condition
ArgumentNullException	input or encoding is a null reference (Nothing in Visual Basic).