Predicate versions of allequal & allunique

[mcabbott]

Closes #47005, and does the same for allunique.

First, a33ac25 uses Iterators.peel to improve allequal(itr) as suggested by @jlapeyre in here: #47005 (comment) .

After that, allequal(f, xs) = allequal(f(x) for x in xs) should never call f twice on the same x, nor more times than necessary.

For allunique, at present this has a path which collects short iterators (<32) as this is faster than making a Set. Thus calling allunique(f(x) for x in xs) will sometimes call f more times than it has to. This PR doesn't change that, just adds a method allunique(f, xs).

Since generators tend to have eltype Any, it pays to use Set{@default_eltype(C)}() now. I think this ought to be safe.

Finally, methods allequal(f, xs::Tuple) and allunique(f, xs::Tuple) seem to be worthwhile. This gist times a few variants.

[rfourquet]

very nitpicky, but using different names for the iterator seems confusing (even though it's maybe helpful for the current definition of this second form).

[mcabbott]

That is a bit ugly, I think I wanted to say itr = (f(x) for x in xs) below... but can say it another way.

Think it's OK to roll these definitions into the same docstring? They could be separate but probably that's just more text to dig through.

[mcabbott]

See if you like 21961ec

[rfourquet]

Yes looks good. An alternative could be

""""
    allunique([f=identity], itr) -> Bool

Return `true` if all values from `(f(x) for x in itr)` are distinct when compared with [`isequal`](@ref).
"""

(but that makes the simpler case (allunique(itr)) more difficult to understand...)

[mcabbott]

Maybe it should warn you that how many times f is called is regarded as an implementation detail? It may not be called at all (if length 1 is known) and it may be called even after a non-unique elements has been seen (as _indexed_allunique(collect(C)) makes the whole array & then checks).

Not sure how surprising that is. In my head at least, f ought to be fairly cheap, because you are obviously throwing away what it returns, and probably pure too.

[mcabbott]

724f048 adds such a note.

[3f6a]

Would be nice to have this. Is there something missing on this PR, or is this ready to be merged?

[mcabbott]

Pre-1.11 bump?

[vtjnash]

LGTM. I polished up the commit message a bit on merge.