Update thinvec to 0.2.19

[panstromek]

It contains may_dangle Drop impl under a feature flag. This allows us to use it in a lot more places, notable in types that hold &'tcx references. This would previously fail in borrowck in arena code.

We have to enable its "unstable" feature for this. That also enables TrustedLen impl, which is behind unstable flag, too. I left the flag disabled, because it regressed next-solver benchmarks, but that will be fixed in PR that needs this flag (see the comment: #159974 (comment))

[rustbot]

The parser was modified, potentially altering the grammar of (stable) Rust
which would be a breaking change.

cc @fmease

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

Some changes occurred in compiler/rustc_attr_parsing

cc @jdonszelmann, @JonathanBrouwer

[rustbot]

r? @davidtwco

rustbot has assigned @davidtwco.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 74 candidates
• Random selection from 16 candidates

[mejrs]

You can enable the unstable feature for rustc_type_ir as well, you just have to gate it behind its nightly feature (just like smallvec).

[panstromek]

You can enable the unstable feature for rustc_type_ir as well, you just have to gate it behind its nightly feature

Yes, but if I understand corectly, Cargo will also enable it through feature unification automatically when you build the compiler, right?

---

I tried to use the new version in #159975 and it had surprisingly negative perf results, so let's try measure this in isolation, just in case.

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 83b3c6e (83b3c6ea3bc5da808387eff3da89c310d09ad5e3)
Base parent: dfbea5b (dfbea5bc945a9d7ff017d3d4bfbd1af9295fe5b8)

[rust-timer]

Finished benchmarking commit (83b3c6e): comparison URL.

Overall result: ❌ regressions - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@bors rollup=never rustc-perf
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.6%	[0.2%, 2.5%]	16
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (secondary -2.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.0%	[-2.5%, -1.5%]	2
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 2.4%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.4%	[2.1%, 6.3%]	5
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.6%	[-2.6%, -2.6%]	1
All ❌✅ (primary)	-	-	0

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 486.74s -> 487.133s (0.08%)
Artifact size: 387.74 MiB -> 388.40 MiB (0.17%)

[panstromek]

Interesting, I wonder why is new-solver so sensitive to ThinVec. Anyway, let's first try to benchmark the update without unstable features - may_dangle changes how Drop is handled, and TrustedLen impacts which specializations are used.

@bors try @rust-timer queue

This should correspond to a smaller diff in thin-vec code:

• mozilla/thin-vec@de42024 which is a dead code for us, I believe
• mozilla/thin-vec@e903080 which adds a capacity overflow check on alloc path, doesn't seem super significant to me
• mozilla/thin-vec@4479f83 which moves a function, so it should be a no op

@rustbot author
(until I figure it out)

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[rust-bors]

☀️ Try build successful (CI)
Build commit: 70848ee (70848eecea0054e9979a011fba24323e191e2e21)
Base parent: dfbea5b (dfbea5bc945a9d7ff017d3d4bfbd1af9295fe5b8)

[rust-timer]

Finished benchmarking commit (70848ee): comparison URL.

Overall result: ❌ regressions - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -2.4%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.4%	[-2.4%, -2.4%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.4%	[-2.4%, -2.4%]	1

Cycles

Results (secondary -2.9%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.5%	[3.5%, 3.5%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.5%	[-6.2%, -2.4%]	4
All ❌✅ (primary)	-	-	0

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 486.74s -> 490.208s (0.71%)
Artifact size: 387.74 MiB -> 389.79 MiB (0.53%)

[panstromek]

Ok, this is great (include-blob is noise probably). It looks like the regression is in the unstable feature (probably TrustedLen impl on Drain?), but it turns out #160005 also needs this update and it will change the new-solver code to avoid or outweigh that regression.

I think we could do the update without "unstable" here and let #160005 enable the feature flag while fixing the regression. But we can also just close and wait for #160005 to do this all at once, but I'm not sure how long that's going to take.

@rustbot ready

[mejrs]

r? me @bors r+ rollup

It looks like the regression is in the unstable feature (probably TrustedLen impl on Drain?)

That's weird 🤔

View changes since this review

[rust-bors]

📌 Commit 02a54aa has been approved by mejrs

It is now in the queue for this repository.

🌲 The tree is currently closed for pull requests below priority 100. This pull request will be tested once the tree is reopened.

Reason for tree closure: spurious failures