Use clang-format in tidy to check the C++ code style under llvm-wrapper

[DianQK]

Fixes #123510.

Based on the discussion at https://rust-lang.zulipchat.com/#narrow/stream/233931-t-compiler.2Fmajor-changes/topic/Enable.20.60clang-format.60.20for.20.60rustc.60.20compiler-team.23756/near/443562800, we can use clang-format from pip to achieve the code formatting.

r? @Kobzol

[bors]

☔ The latest upstream changes (presumably #126540) made this pull request unmergeable. Please resolve the merge conflicts.

[DianQK]

@rustbot ready
@rustbot label -S-experimental

[rustbot]

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.

[DianQK]

The job mingw-check-tidy failed! Check out the build log: (web) (plain)
Click to see the possible cause of the failure (guessed by this bot)

This is the unformatted actual output.

[Kobzol]

Looks great, thank you!

@bors r+

[bors]

📌 Commit c163d5c has been approved by Kobzol

It is now in the queue for this repository.

[bors]

⌛ Testing commit c163d5c with merge 7033f9b...

[bors]

☀️ Test successful - checks-actions
Approved by: Kobzol
Pushing 7033f9b to master...

[rust-timer]

Finished benchmarking commit (7033f9b): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

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

Max RSS (memory usage)

Results (secondary -3.8%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.8%	[-3.8%, -3.8%]	2
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary -2.0%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.0%	[-2.0%, -2.0%]	1
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 696.035s -> 694.579s (-0.21%)
Artifact size: 326.75 MiB -> 326.68 MiB (-0.02%)

[nikic]

What is the command to call this locally?

I tried ./x.py test tidy -- tidyselftest --extra-checks=cpp:fmt (which is probably not the usual way to spell it?) but it did not report formatting errors that CI does.

[DianQK]

What is the command to call this locally?

I tried ./x.py test tidy -- tidyselftest --extra-checks=cpp:fmt (which is probably not the usual way to spell it?) but it did not report formatting errors that CI does.

It's ./x test --stage 0 src/tools/tidy tidyselftest --extra-checks=cpp:fmt.

[nikic]

I see... It seems like the minimal command to make this work is ./x.py test tidy --extra-checks=cpp:fmt.

What confused me is that ./x.py tidy --extra-checks=cpp:fmt does not work (and suggests using -- --extra-checks=cpp:fmt instead, which does not work either).

[nikic]

I'm also wondering why TIDY_PRINT_DIFF=1 is not enabled by default. The default output seems to be unhelpful to the point of uselessness.

[DianQK]

What confused me is that ./x.py tidy --extra-checks=cpp:fmt does not work (and suggests using -- --extra-checks=cpp:fmt instead, which does not work either).

I guess this is unexpected. It should print:

error: 'x.py' requires a subcommand but one was not provided
 [subcommands: build, b, check, c, clippy, fix, fmt, doc, d, test, t, miri, bench, clean, dist, install, run, r, setup, suggest, vendor, perf]

I'm also wondering why TIDY_PRINT_DIFF=1 is not enabled by default. The default output seems to be unhelpful to the point of uselessness.

I guess it's because we can use --bless to fix the formatting problem and lint doesn't have diff to show.

cc @Kobzol

[Kobzol]

I agree that displaying the diff by default is a good idea (not sure why we didn't do it before). When you run x test tidy and some Rust code is unformatted, it also prints the diff by default, even without any special flags.

@tgross35 Any objections against just printing the Python/C++ linting/formatting diff by default?

[tgross35]

That sounds good to me, I don't know why I didn't just do that in the first place.

I have also been meaning to change

rust/src/bootstrap/src/core/config/flags.rs

Lines 363 to 365 in f25e92b

	/// comma-separated list of other files types to check (accepts py, py:lint,
	/// py:fmt, shell)
	extra_checks: Option<String>,

to a Vec<ext_tool_checks::Arg> + enum Arg { Py, PyLint, Cpp, ... } so Clap handles printing the available options, since that doc comment is out of date. I'll probably get to that this week but if somebody beats me to it, be my guest 😆