require-unicode-regexp

Without the u or v flag, a regular expression works on 16-bit units, not on characters. /^.$/ doesn't match a single emoji such as "\u{1F600}", because it takes two units. The flags also make the pattern strict, so a useless escape such as \a is a syntax error instead of being silently read as a.

The rule checks regular expression literals and calls to RegExp, with or without new. If the flags aren't a constant string, the call isn't reported.

The report comes with a suggestion that adds the flag, if the pattern is still valid with it.

Configure

Off by default. To turn it on:

[lint.rules]
"require-unicode-regexp" = "error" # or "warn"

Every option, set to its default. The ones that are commented out are unset by default.

[lint.rules]
"require-unicode-regexp" = ["error", {
  # requireFlag = <"u" or "v">,
}]

Examples

Reported:

const oneCharacter = /^.$/;
const word = new RegExp("\\w+", "g");

Not reported:

const oneCharacter = /^.$/u;
const word = new RegExp("\\w+", "gu");
const emoji = /^\p{RGI_Emoji}$/v;

function search(pattern, flags) {
  return new RegExp(pattern, flags);
}

Options

Option	Default	What it does
requireFlag	not set	Set it to "u" or "v" to accept only that flag. When it isn't set, either flag is accepted.

With requireFlag set to "v", the u flag is reported:

const oneCharacter = /^.$/u;