Overhaul token cursors

[nnethercote]

Some nice cleanups here.

r? @davidtwco

[nnethercote]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 7874297 with merge 29ae3e8...

[bors]

☀️ Try build successful - checks-actions
Build commit: 29ae3e8 (29ae3e83e7d9a22117bca546c540f18c2cb36a8b)

[rust-timer]

Finished benchmarking commit (29ae3e8): comparison URL.

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

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

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

Instruction count

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.4%	[0.3%, 0.4%]	3
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.1%]	2
Improvements ✅ (secondary)	-0.3%	[-0.5%, -0.2%]	4
All ❌✅ (primary)	-0.2%	[-0.2%, -0.1%]	2

Max RSS (memory usage)

Results (primary 1.3%, secondary -0.2%)

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)	2.0%	[1.5%, 2.6%]	4
Regressions ❌ (secondary)	1.8%	[1.6%, 2.3%]	4
Improvements ✅ (primary)	-1.1%	[-1.1%, -1.1%]	1
Improvements ✅ (secondary)	-4.3%	[-5.1%, -3.4%]	2
All ❌✅ (primary)	1.3%	[-1.1%, 2.6%]	5

Cycles

Results (primary 6.0%, secondary -0.5%)

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)	6.0%	[3.3%, 9.8%]	3
Regressions ❌ (secondary)	9.4%	[9.4%, 9.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.0%	[-3.5%, -2.5%]	4
All ❌✅ (primary)	6.0%	[3.3%, 9.8%]	3

Binary size

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

Bootstrap: 767.862s -> 766.941s (-0.12%)
Artifact size: 331.04 MiB -> 331.06 MiB (0.01%)

[nnethercote]

I saw small but consistent icount improvements locally:

The most likely explanation when local wins aren't seen on CI is that PGO is involved. But not every platform has PGO applied, and the improvements here are also to readability, so I think this is still worth going forward.

[rustbot]

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

Some changes occurred in src/tools/rustfmt

cc @rust-lang/rustfmt

[nnethercote]

@rustbot ready

[nnethercote]

@petrochenkov: I know your time is limited at the moment. Unlike PRs such as #133436 this PR is not that complicated, so I will ask someone else to review.

r? @davidtwco

[rustbot]

Could not assign reviewer from: davidtwco.
User(s) davidtwco are either the PR author, already assigned, or on vacation, and there are no other candidates.
Use r? to specify someone else to assign.

[ytmimi]

I see that we're replacing calls to .look_ahead() with .peek(). Is .peek() equivalent to both .look_ahead(0) here and .look_ahead(1) above in MacroParser::parse?

[nnethercote]

Yes. Full details are in the individual commit messages.

[ytmimi]

Interesting that it doesn't matter if you use 0, 1, or 2 as the look ahead, and that things only start failing if you choose 3 or more. I know you mentioned it's due to the structure of the code, but I'm wondering if you could elaborate a little more to help me fully understand.

[nnethercote]

AIUI, the existing code says "If there are at least two tokens remaining, try to parse a branch". The new code says "if there is at least one token remaining, try to parse a branch". parse_branch has three self.toks.next()? calls in it, and all three must succeed for the function to succeed, i.e. there must actually be at least three tokens remaining.

[ytmimi]

Thanks for walking me through that. We're good on the rustfmt changes.

[nnethercote]

r? @spastorino

[nnethercote]

I rebased.

@bors r=spastorino rollup=maybe

[bors]

📌 Commit 464d0ee has been approved by spastorino

It is now in the queue for this repository.

[bors]

📌 Commit 8a85bdc has been approved by spastorino

It is now in the queue for this repository.