[Issue 7067] std.random.RandomSample and RandomCover are poorly designed

[via Digitalmars-d-bugs]

https://issues.dlang.org/show_bug.cgi?id=7067

--- Comment #21 from Martin Nowak <code at dawg.eu> ---
(In reply to Joseph Rushton Wakeling from comment #20)
> I guess what I don't like about this solution is that it requires the user
> to take responsibility for generating the RNG (or wrapper-of-RNG) as a
> reference type, when that ought to be the default behaviour.  But OTOH we
> could take advantage of the fact that RNG structs are templated, and rework
> their convenience aliases, e.g.:
> 
>     alias Mt19937 = RefCounted!(MersenneTwisterEngine!(uint, 32LU, 624LU,
> 397LU, 31LU, 2567483615u, 11LU, 7LU, 2636928640u, 15LU, 4022730752u, 18LU))
> 
> ... which would still leave the underlying struct available for use by
> someone who really wants to use it directly in that form.

I don't think that sharing rng state among multiple consumers is a good idea.
Of course it looks like you want random numbers and shouldn't care about the
order, but often fixed seeds are used to get a reproducible pseudo-random
range.
When you share such a RNG as in
    auto rng = Random(fixedSeed);
    assert(!equal(randomCover(a, rng), randomCover(a, rng)));
the sequences depends on the implementation of equal and randomCover.
Would be cleaner to use 2 RNGs with independent state here, so the fact that a 
shared state requires to be explicit is a good thing. Also we'd set a bad
precedent by making one of std's ranges a ref type. Ref ranges have a lot of
subtleties and aren't that well tested with std.algorithm.

--