[Cryptography] Dieharder & /dev/urandom

[Michel Arboi]

Le mar. 14 mai 2019 à 18:51, Natanael <natanael.l at gmail.com> a écrit :

> So passing or failing some tests is not the interesting question - it is *how many* tests you pass or fail and of what kind.

I (nearly?) always get a couple of "WEAK" results with urandom, not with random.

I'm confused, as this is to be the same PRNG in fact -- random is just
guaranteed to be regularly reseeded if I understood correctly.