Port #[cfg_attr] to the new attribute parsing infrastructure

[JonathanBrouwer]

This work in progress, not ready for review.
PR mostly for ci/perf runs

[JonathanBrouwer]

r? @jdonszelmann
Not ready for review yet, just assigning you

[JonathanBrouwer]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: ff502cb (ff502cb3140362d4d93bde4d537bcdb21bf639be, parent: 4b94758d2ba7d0ef71ccf5fde29ce4bc5d6fe2a4)

[rust-timer]

Finished benchmarking commit (ff502cb): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@bors rollup=never
@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%, 0.2%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.2%	[-0.4%, -0.1%]	14
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -2.2%, secondary 0.9%)

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)	3.1%	[3.1%, 3.1%]	1
Improvements ✅ (primary)	-2.2%	[-2.2%, -2.2%]	1
Improvements ✅ (secondary)	-1.3%	[-1.3%, -1.3%]	1
All ❌✅ (primary)	-2.2%	[-2.2%, -2.2%]	1

Cycles

Results (secondary 3.5%)

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)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 475.312s -> 475.298s (-0.00%)
Artifact size: 388.16 MiB -> 388.09 MiB (-0.02%)

[rustbot]

Some changes occurred in compiler/rustc_attr_parsing

cc @jdonszelmann

[JonathanBrouwer]

This is ugly but not a regression, this was already broken before.
#147693
I'll fix this in a separate PR, since this one is already a little big

[JonathanBrouwer]

@jdonszelmann I think the perf looks reasonable, more improvements than regressions and it's all secondary benchmarks anyways. Feel free to disagree and I can look into the regressions more though.
(I made no changes to the code since the perf run, only docs and tests, so no need to rerun)

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[jdonszelmann]

wouldn't it be nicer to somehow use our new attribute parser logic here already? Especially with the prospect of throwing the old one out some day

[JonathanBrouwer]

The problem is that the argument of cfg_attr are not required to be a valid meta item. The sub-attribute could be a proc macro attribute, in which case we still need to have access to the TokenStream. I could not find a nice way to achieve this goal with the new attribute parser logic.

The old attribute parser is not actually used here afaik, we don't parse to a MetaItem here. So throwing the old one out should still be possible with this implementation