Prevent commented out tests

[sindresorhus]

We should recommend using test.skip() instead.

This:

// test('foo', => {
//  t.pass();
// });

Should instead be:

test.skip('foo', => {
    t.pass();
});

[jamestalmage]

👍

Interesting proposal.

How should we accomplish it? Try parsing the contents of the comments, or just use regexp?

Trying to parse will have more false negatives, regexp more false positives.

[jfmengels]

I found https://github.com/bahmutov/eslint-rules#no-commented-out-code that looks like this. I have no idea if it works well, but it might be a good lead.

[jamestalmage]

It looks like it parses line by line, and warns if any single line is valid JS.

This is a little more difficult to do with parsing because tests will (very likely) cover multiple lines:

test(t => {
  t.pass();
});

The first line above isn't valid JS on it's own, you need the whole thing. But if they do this when they comment it out:

// I am skipping this because: reasons.
// 
// test(t => {
//   t.pass();
// });

Then parsing the entire contents of the comments creates a syntax error.

I guess we could create a regexp to try and find the potential first line of a test, and then add lines incrementally until we found something that parsed.

Or maybe we could try the parse_dammit method from acorn.

[sindresorhus]

I was thinking regex.

[jamestalmage]

I was thinking regex.

Uh oh: #85

My concern with regexp is the likelihood of false positives.

[jamestalmage]

Acorns loose parser is surprisingly good (at making bad code into a usable AST), it might allow just as many false positives through as a regex would.

[sindresorhus]

Yeah, parsing looks like a more solid solution.

[jamestalmage]

I think this would be difficult to handle without parsing.

[jfmengels]

I found @platinumazure's eslint-plugin-qunit, which has this rule alreadby built-in (https://github.com/platinumazure/eslint-plugin-qunit/blob/master/lib/rules/no-commented-tests.js), and uses a regex to find it. In QUnit's case, there are bounds to be less false positives.

[revelt]

I came here looking to check does it exist and if necessary, to propose it :)

Imho, this rule would be one of the most useful-ones from what I've seen so far in the proposals. It's very difficult to pinpoint commented-out rule when the tests file is long.