No it's not, from §4.4:
> Set all the other bits to randomly (or pseudo-randomly) chosen values.
I can use a PRNG with a cycle length of 8 and it would be fully correct according to the RFC, but it would be trivial to brute-force all the values.
No it's not, from §4.4:
> Set all the other bits to randomly (or pseudo-randomly) chosen values.
I can use a PRNG with a cycle length of 8 and it would be fully correct according to the RFC, but it would be trivial to brute-force all the values.
In the same section of uuid RFC: "See Section 4.5 for a discussion on random numbers.", and then: "Advice on generating cryptographic-quality random numbers can be found in RFC1750 [5].", so it's not like authors were unaware of the implications of trivial PRNG. It's just up to the implementation whether CSPRNG is required (user-visible security tokens), or just a PRNG (internal identifiers).
Generating correlated v4 UUIDs isn't wrong, but it is stupid in the sense of being self-defeating: nobody who wants to generate v4 UUIDs wants them with insufficient randomness, and will have no reason to not consider a bad random-number source to be a bug, because it's interfering with getting them the thing they want to get by generating v4 UUIDs in the first place.
(The real point of the explicit UUID spec, meanwhile, is in saying what constitutes a valid UUID—and it's certainly valid to generate a UUIDv4 using insufficient entropy. There's no way for a peer receiving such UUIDs to guess that they're maybe-predictable-with-enough-effort and reject them on that basis, which is all "validity" can ever mean: what can or cannot be technically enforced by protocol peers.)