Unnecessary escape character?

[dcousens]

Unnecessary escape character: \]. (no-useless-escape)

The regex:

/Expected property "1" of type \[Buffer\], got String "foobar"/

Is this one for eslint? Or am I misunderstanding?

edit: This apparently works... is it even sane though? It feels so odd.

/Expected property "1" of type \[Buffer], got String "foobar"/

[dcousens]

This only started happening in a recent update.
(forcing me to do bitcoinjs/bitcoinjs-lib@579fcce)

[feross]

The no-useless-escape rule got stricter in the latest version of eslint. I think they're right that it's technically useless to escape \] but I agree that this is a tad bit aggressive.

This might be the source of the issue: eslint/eslint#7424

[feross]

One good argument for why ESLint should not warn about \] is that regex's "Unicode mode" (which @mathiasbynens recently described as "strict mode" for regex literals) doesn't warn about it.

/\a/u - throws because a is always an invalid escape sequence. Prevents user from making a mistake.
/\)/u - does not throw even though ) does not need to be escaped. This is because ) sometimes does need to be escaped, and it doesn't hurt to just allow it. Very unlikely this is a user mistake.

This let's users keep things nice and parallel in their regex, like: /\[Buffer\]/u. My opinion: if it's good enough for regex "Unicode mode", then it's good enough for ESLint. 👍

[feross]

I opened an issue: eslint/eslint#7656

[dcousens]

Thanks for championing this @feross

[dcousens]

Close in favour of eslint/eslint#7656?

[feross]

Sure. If they end up adding a way to specify exceptions to this rule, we can open a new issue for that.

[feross]

FYI, the original eslint issue was fixed: eslint/eslint#7656