fix(linter): Report implicit config parse errors (#12258)

[simonbuchan]

Fixes #12258

Not super sure about the config path normalization, but it Works On My Machine™️! Only other example of <cwd> in linter snapshots is the tsconfig error, which explicitly formats the error at that location.

Might be nicer to regex the snapshot for any paths to handle Windows, the current cwd detection in there seems a bit hacky anyway.

[graphite-app]

How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

• 0-merge - adds this PR to the back of the merge queue
• hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

[simonbuchan]

Dang it. absolute() doesn't actually validate the file exists...

[simonbuchan]

I've tried a few options, but they're all pretty messy - the simplest I can think of is to introduce Oxlintrc::try_from_file() that returns a richer error that can be introspected or Oxlintrc::from_file_if_exists(path: &Path) -> Result<Option<Self>, OxcDiagnostics>?

[camc314]

oh thatnks for this, i've seen this error before, let me fix it

[simonbuchan]

Yeah, plenty of ways to fix this but I'm not sure what the conventions are well enough to start hacking on something this core.

[camc314]

basically we need to use

        let handler = if cfg!(any(test, feature = "force_test_reporter")) {
            GraphicalReportHandler::new_themed(miette::GraphicalTheme::none())
        } else {
            GraphicalReportHandler::new()
        };

to report the error else you get diff snapshots in CI vs locally

[simonbuchan]

If you're picking this up, seems my issue is a duplicate of #12073, which also points out the nested override config files have the same issue.

[simonbuchan]

I'll take another stab tomorrow otherwise!

[simonbuchan]

Note the actual test failures are because absolute() isn't checking the file exists, despite the existing logic implying that it does. The from_file() method buries the distinction in an OxcDiagnostic which prevents the caller from treating them separately. Switching to from_string() does work somewhat but breaks the error messages. Fixing that basically impacts the whole CLI though, which is a bit too scary for me right now!

[camc314]

touch wood, this update should work now 🤞

[camc314]

oxc/apps/oxlint/src/lint.rs

Lines 508 to 513 in 42d6bd7

	Oxlintrc::from_file(&possible_config_path).map_err(|e| {
	let handler = GraphicalReportHandler::new();
	let mut err = String::new();
	handler.render_report(&mut err, &e).unwrap();
	err
	})

we should also remove use of GraphicalReportHandler::new() here and instead propogate the OxcDiagnostic

[simonbuchan]

Simple enough in the end, thanks for cleaning my mess up!

[camc314]

thanks for contributing!