Hash algorithm implementations for .NET, including CRC (CRC32 and CRC16).
All hash algorithms support streaming (hashing a block at a time) and have full support for Span<T> to prevent unnecessary memory allocation. They also have excellent downlevel platform support (netstandard1.3).
CRC32 is a generic 32-bit CRC implementation, patterned after Boost's CRC library.
You can pass a CRC32.Definition instance into the constructor to control all the CRC algorithm parameters.
var defaultHasher = new CRC32(); // uses CRC32.Definition.Default
var bzipHasher = new CRC32(CRC32.Definition.BZip2);
var customHasher = new CRC32(new CRC32.Definition
{
TruncatedPolynomial = 0x1EDC6F41,
Initializer = 0xFFFFFFFF,
FinalXorValue = 0xFFFFFFFF,
ReverseDataBytes = true,
ReverseResultBeforeFinalXor = true,
});Once created, the CRC32 instance can be used like any other HashAlgorithm instance.
Built-in implementations include:
Default- an older CRC-32 defined by the old IEEE standard. Used by Ethernet, zip, PNG, and many other systems. Known as"CRC-32","CRC-32/ADCCP", and"PKZIP". This is the default CRC-32 definition because it is the most common; however, there are better polynomials if you are defining a new protocol/format.BZip2- a CRC-32 used by BZIP2. Known as"CRC-32/BZIP2"and"B-CRC-32".Castagnoli- a modern CRC-32 defined in RFC 3720 (iSCSI). Known as"CRC-32C","CRC-32/ISCSI", and"CRC-32/CASTAGNOLI".Mpeg2- a CRC-32 used by the MPEG-2 standard. Known as"CRC-32/MPEG-2".Posix- a CRC-32 used by the POSIXchksumcommand; note that thechksumcommand-line program appends the file length to the contents unless it is empty. Known as"CRC-32/POSIX"and"CKSUM".Aixm- a CRC-32 used in the Aeronautical Information eXchange Model. Known as"CRC-32Q".Xfer- a very old CRC-32, appearing in Numerical Recipes in C. Known as"XFER".
CRC16 is a generic 16-bit CRC implementation, patterned after Boost's CRC library.
You can pass a CRC16.Definition instance into the constructor to control all the CRC algorithm parameters.
var defaultHasher = new CRC16(); // uses CRC16.Definition.Default
var ccittHasher = new CRC16(CRC16.Definition.Ccitt);
var customHasher = new CRC16(new CRC16.Definition
{
TruncatedPolynomial = 0x1021,
Initializer = 0xFFFF,
FinalXorValue = 0xFFFF,
ReverseDataBytes = true,
ReverseResultBeforeFinalXor = true,
});Once created, the CRC16 instance can be used like any other HashAlgorithm instance.
Built-in implementations include:
Default- a common CRC-16, used by ARC and LHA.CcittFalse- a CRC-16 used by floppy disk formats, commonly misidentified as CCITT.Ccitt- a CRC-16 used by Kermit. Known as"CCITT". Appears in Numerical Recipes in C.XModem- a CRC-16 used by XMODEM and ZMODEM. Appears in Numerical Recipes in C.X25- a CRC-16 used by X.25, V.42, T.30, and RFC 1171 (PPP). Appears in Numerical Recipes in C.
The Nito.HashAlgorithm.Core NuGet library has a HashAlgorithmBase type that provides an easier way to define hash algorithms across the widest possible number of platforms (netstandard1.3 - netstandard2.1). It pulls in System.Memory on platforms that require it to provide support for Span<T>.
To use HashAlgorithmBase, your hash algorithm implementation should pass the size of the hash (in bits) to the HashAlgorithmBase constructor. HashAlgorithmBase will set HashSizeValue or override HashSize appropriately, depending on platform.
Then implement these two methods:
protected void DoHashCore(ReadOnlySpan<byte> source);
protected void DoHashFinal(Span<byte> destination);The DoHashCore method should hash the bytes in source. The DoHashFinal method should save the hash result into destination (hint: BinaryPrimitives is useful when implementing this method).
That's it; the HashAlgorithmBase takes care of implementing the HashAlgorithm methods, including overrides for efficient Span<T>-based hashing.