This is a C++ implementation of the SHA-3 family of hash functions, based on 
the documentation found at 
http://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.202.pdf
Download binaries at: http://www.danielhorne.co.uk/Projects/SHA-3/SHA3.html

Usage: sha3sum [command]* file*

 where command is an optional parameter that can set either the algorithm, as
 there is a slight difference between the bare keccak function and the SHA-3
 variant.
 
 Algorithm 
 
 -a=s   :  Set algorithm to SHA-3 (default).
 -a=k   :  Set algotithm to Keccak.
 -a=h   :  Set algorithm to SHAKE
 
 Size
 
 -w=224 :  Set width to 224 bits.
 -w=256 :  Set width to 256 bits.
 -w=384 :  Set width to 384 bits.
 -w=512 :  Set width to 512 bits (default).
 
 Digest size (SHAKE)
 
 -d=number : Set the SHAKE digest size. Should be less than or equal to the hash size.
       Should be multiple of 8
       Only relevant for SHAKE - For SHA-3 and keccak, digest size is equal to sponge size.

 Output format
 
 -o=b : Base 64 encoded.
 -o=h : hexadecimal (default)
 
Any number of files can be specified. Files will be processed with the most
recently specified options - for example:

  sha3sum test.txt -a=k -w=384 test.txt -a=s -w=256 text.txt

will hash "test.txt" three times - First with 512-bit SHA-3, then with 384-bit
keccak, then finally with 256-bit SHA-3.



The intent with this implementation is to make a performant implementation in 
pure C++, following modern idioms: Favouring inline functions over macros, 
leaning on the compiler's optimiser instead of resorting to inline assembly,
for example. 

ToDo:

- Add command option to set the chunk size for reading.

- Attempt formal proof of important aspects of this implementation. 

Bugs:

- Any bugs, please either add them to the github repo 
  (https://github.com/DuSTman31/SHA-3/issues), or e-mail me at
  d-.-h-o-r-n-e-@-d-a-n-i-e-l-h-o-r-n-e-.-c-o-.-u-k (remove "-"s).
  
- Will document the design and verification on www.danielhorne.co.uk when I 
  get round to it.