Implement array::repeat

[joboet]

See rust-lang/libs-team#310.

I've decided to make the function use the input value as last element instead of cloning it to every position and dropping it, and to make this part of the API so that callers are not surprised by this behaviour.

TODO: open a tracking issue. I'll wait for the ACP to be accepted, first.

@rustbot label +T-libs-api +T-libs
r? libs

[scottmcm]

I've decided to make the function use the input value as last element instead of cloning it to every position and dropping it

As I wrote in the ACP, the way to do this is to use https://stdrs.dev/nightly/x86_64-unknown-linux-gnu/std/iter/fn.repeat_n.html:

from_trusted_iterator(iter::repeat_n(x, N))

No need to MaybeUninit manually here.

[joboet]

I've decided to make the function use the input value as last element instead of cloning it to every position and dropping it

As I wrote in the ACP, the way to do this is to use https://stdrs.dev/nightly/x86_64-unknown-linux-gnu/std/iter/fn.repeat_n.html:

from_trusted_iterator(iter::repeat_n(x, N))

No need to MaybeUninit manually here.

I fear that would lead to suboptimal codegen, since the optimizer has to figure out that counting backwards results in the same amount of iterations as iterating forwards, which in my experience it is not really able to do well.

[scottmcm]

You should use the direct implementation unless you have a codegen test demonstrating that the custom implementation is actually better. Counting down can actually be better than counting up, in assembly (example).

Regardless of the implementation, it might be good to have some simple tests like checking that repeat(123_u8) will optimize to a single memset. It looks like that'll happen for the iter::repeat_n approach, at least: https://rust.godbolt.org/z/EhG1azc67.

[scottmcm]

Finally approved in rust-lang/libs-team#310 (comment) !

I'm going to re-run CI as a double-check, since it's been quite a while.

[joboet]

Rebased, updated and tracking issue added.
@rustbot ready

[scottmcm]

Thanks! Excited to finally get this 🙂

@bors r+ rollup=iffy (new codegen test, which isn't tested in CI)

[bors]

📌 Commit 0aa3310 has been approved by scottmcm

It is now in the queue for this repository.

[bors]

⌛ Testing commit 0aa3310 with merge d8a38b0...

[bors]

☀️ Test successful - checks-actions
Approved by: scottmcm
Pushing d8a38b0 to master...

[rust-timer]

Finished benchmarking commit (d8a38b0): comparison URL.

Overall result: ❌ regressions - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.3%	[0.3%, 0.3%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 3.1%, secondary -3.3%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	3.1%	[2.5%, 3.8%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.3%	[-4.0%, -2.5%]	2
All ❌✅ (primary)	3.1%	[2.5%, 3.8%]	2

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 691.194s -> 692.485s (0.19%)
Artifact size: 323.71 MiB -> 323.83 MiB (0.04%)