|Maintainer||Vincent Hanquez <firstname.lastname@example.org>|
Provide a safe abstraction for cryptographic pseudo random generator.
- data EntropyPool
- createEntropyPool :: IO EntropyPool
- grabEntropy :: Int -> EntropyPool -> SecureMem
- class CPRG gen where
- withRandomBytes :: CPRG g => g -> Int -> (ByteString -> a) -> (a, g)
- data SystemRNG
- createTestEntropyPool :: ByteString -> EntropyPool
Pool of Entropy. contains a self mutating pool of entropy, that is always guarantee to contains data.
Create a new entropy pool with a default size.
While you can create as many entropy pool as you want, the pool can be shared between multiples RNGs.
Cryptographic Pseudo Random Generator
Create a new CPRG using an object of the CryptoGenerator class and with an explicit reference to an EntropyPool.
Give the ability to set a threshold of byte generated that after
being exceeded will result in a reseed with some stateful entropy
after a call to
If this threshold is exceeded during the set operation, the rng should be reseeded here.
If this value is set to 0, no reseeding will be done and the output will be completely predicable. This is not a recommended level except for debugging and testing purpose.
Fork a CPRG into a new independent CPRG.
As entropy is mixed to generate safely a new generator, 2 calls with the same CPRG will not produce the same output.
Generate a number of bytes using the CPRG.
Given one CPRG, the generated bytes will always be the same.
However the returned CPRG might have been reseeded with entropy bits, so 2 calls with the same CPRG will not necessarily result in the same next CPRG.
Similar to cprgGenerate except that the random data is mixed with pure entropy, so the result is not reproducible after use, but it provides more guarantee, theorically speaking, in term of the randomness generated.
len random bytes and mapped the bytes to the function f.
This is equivalent to use Control.Arrow
System entropy generator.
This generator doesn't use the entropy reseed level, as the only bytes generated are comping from the entropy pool already.
This generator doesn't create reproducible output, and might be difficult to use for testing and debugging purpose, but otherwise for real world use case should be fine.