remove the 'dereferenceable' attribute from Box

[RalfJung]

Fixes#66600

r? @eddyb@rkruppe

[jonas-schievink]

@bors try

[bors]

⌛ Trying commit be07911 with merge d7c557381942c3cac82b5e4585f195ec369c2f68...

[RalfJung]

@rust-timer queue

[rust-timer]

Awaiting bors try build completion

[bors]

☀️ Try build successful - checks-azure
Build commit: d7c557381942c3cac82b5e4585f195ec369c2f68 (d7c557381942c3cac82b5e4585f195ec369c2f68)

[rust-timer]

Queued d7c557381942c3cac82b5e4585f195ec369c2f68 with parent 083b5a0, future comparison URL.

[rust-timer]

Finished benchmarking try commit d7c557381942c3cac82b5e4585f195ec369c2f68, comparison URL.

[RalfJung]

@Mark-Simulacrum this said "finished benchmarking" but there are no results at the comparison URL yet?

[bjorn3]

@RalfJung It does now. It takes some time before the parent commit gets benchmarked after @rust-timer says it has finished benchmarking the child.

[RalfJung]

I see.

Timings look like noise. But this is not necessarily a good benchmark.

[eddyb]

r=me if there are no concerns left (e.g. perf)

[RalfJung]

@pnkfelix so how do we proceed? I guess we should nominate this for T-lang and T-compiler discussion?

[pnkfelix]

@bors r+ rollup=never

[bors]

📌 Commit d0b9833 has been approved by pnkfelix

[pnkfelix]

(I marked as rollup=never because its not totally certain that there won't be performance regressions associated with this, and I'd rather pay the cost of an extra bors cycle up front than have to track that down in a rollup or have it otherwise masked by unrelated improvements in a rollup.)

[bors]

⌛ Testing commit d0b9833 with merge c4375c9...

[bors]

☀️ Test successful - checks-azure
Approved by: pnkfelix
Pushing c4375c9 to master...