Revert #107376 to fix potential bincode breakage and rustc-perf benchmark.

[lqd]

#107376 caused rustc-perf's webrender benchmark to break, by regressing on the bincode-1.3.3 crate.

This PR is a draft revert in case we can't land a fix soon enough, and we'd like to land the revert instead

(Though I myself think it'd be safer to do the revert, and run crater when relanding #107376.)

cc @aliemjay

[compiler-errors]

Though I myself think it'd be safer to do the revert, and run crater when relanding #107376.

I also think that this is probably the safest approach as well.

[compiler-errors]

Anyways, if you'd like to unblock perf sooner than later, r=me.

[lqd]

if you'd like to unblock perf sooner than later

That wasn't the primary motivation -- bincode seems popular and the next nightly might cause a bunch of breakage otherwise -- but now that you mention it, it's another good reason to do so :)

We're likely going to miss the nightly cutoff, unfortunately.

We could add the MCVE as a test here, but if we plan to run crater anyways, it can be added when we revert this revert.

Thanks @compiler-errors.

r? @compiler-errors @bors r=compiler-errors

[bors]

📌 Commit 5ad1083 has been approved by compiler-errors

It is now in the queue for this repository.

[compiler-errors]

@bors p=1

[bors]

⌛ Testing commit 5ad1083 with merge 18e305d...

[aliemjay]

Thank you for the revert and the MCVE!

I haven't thought enough about how this blocks perf. I should've prioritized the issue more.

[bors]

☀️ Test successful - checks-actions
Approved by: compiler-errors
Pushing 18e305d to master...

[lqd]

I haven't thought enough about how this blocks perf. I should've prioritized the issue more.

To be clear @aliemjay, perf.rlo wasn't blocked per se, the collector ignores benchmarks that fail but still measures the others. We just wouldn't see regressions or improvements to that specific crate. It also means that a real world crate was broken by a PR, in a way that CI didn't catch (but crater usually would) and that can be more problematic than the lack of benchmarking data for a short period of time.

[rust-timer]

Finished benchmarking commit (18e305d): 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)	1.1%	[0.6%, 1.4%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results

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)	-	-	0
Regressions ❌ (secondary)	3.2%	[2.8%, 3.5%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

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

[lqd]

Even if no triage needs to be done per-se: this is a revert, and the tt-muncher benchmark is currently noisy.