Add internal collect_into_array[_unchecked] to remove duplicate code

[LukasKalbertodt]

Unlike the similar PRs #69985, #75644 and #79659, this PR only adds private functions and does not propose any new public API. The change is just for the purpose of avoiding duplicate code.

Many array methods already contained the same kind of code and there are still many array related methods to come (e.g. Iterator::{chunks, map_windows, next_n, ...}, [T; N]::{cloned, copied, ...}, ...) which all basically need this functionality. Writing custom unsafe code for each of those doesn't seem like a good idea. I added two functions in this PR (and not just the unsafe version) because I already know that I need the Option-returning version for Iterator::map_windows.

This is closely related to #81615. I think that all options listed in that issue can be implemented using the function added in this PR. The only instance where collect_array_into might not be general enough is when the caller want to handle incomplete arrays manually. Currently, if iter yields fewer than N items, None is returned and the already yielded items are dropped. But as this is just a private function, it can be made more general in future PRs.

And while this was not the goal, this seems to lead to better assembly for array::map: https://rust.godbolt.org/z/75qKTa (CC @JulianKnodt)

Let me know what you think :)

CC @matklad@bstrie

[rust-highfive]

r? @dtolnay

(rust-highfive has picked a reviewer for you, use r? to override)

[scottmcm]

Very nice! I was going to propose my approach from #75571 (comment) (https://godbolt.org/z/e1GdEM), but it looks like the one you have here is giving the same elegance (https://rust.godbolt.org/z/66nrWY):

definevoid@_ZN7example12map_ints_new17hd42cfbe0c1377431E([8 x float]* noaliasnocapturesretdereferenceable(32) %0, [8 x i32]* noaliasnocapturereadonlydereferenceable(32) %s) unnamed_addr #0personalityi32 (i32, i32, i64, %"unwind::libunwind::_Unwind_Exception"*, %"unwind::libunwind::_Unwind_Context"*)* @rust_eh_personality!dbg!166 {
start:
%1 = bitcast [8 x i32]* %sto <8 x i32>*, !dbg!167%2 = load <8 x i32>, <8 x i32>* %1, align4, !dbg!167%3 = uitofp <8 x i32> %2to <8 x float>, !dbg!168%4 = bitcast [8 x float]* %0to <8 x float>*, !dbg!180store <8 x float> %3, <8 x float>* %4, align4, !dbg!180, !alias.scope!181, !noalias!184retvoid, !dbg!186
}

Consider whether it'd be worth having a codegen test for one of these examples, since it seems very easy for small permutations to the code to give much worse output...

[bstrie]

I'll second scottmcm's idea of adding a codegen test, if it's not too onerous.

[fbstj]

Suggested change

	/// If `iter.next()` panicks, all items already yielded by the iterator are
	/// If `iter.next()` panics, all items already yielded by the iterator are

[dtolnay]

Nice!

[dtolnay]

@bors r+

[bors]

📌 Commit c675af8 has been approved by dtolnay

[LukasKalbertodt]

Forgot to check back in here. Now the typo was merged, but I guess that's not a huge disaster for a private function.

I also tried to add codegen checks, as suggested, but I don't have any experience with those and ultimately couldn't get it to work in the time I had. Of course, if someone wants to add those still, that would be helpful. See #75243 for more information about the codegen of array::map.