Babel: Possible to represent reserved keywords using unicode escape sequences.

[dashed]

Input Code

let n = 5;

while(n --> 0) {
  \u0062\u0072\u0065\u0061\u{006B};
}

console.log(n);

In action in babel REPL (Babel 6.18.0)

Expected Behavior

It shouldn't be possible to represent reserved keywords using unicode escape sequences (e.g. \uXXXX or \u{XXXX}).

See: http://www.ecma-international.org/ecma-262/7.0/#sec-reserved-words

Current Behavior

example:

\u0062\u0072\u0065\u0061\u{006B}; => break;

Possible Solution

_TBD_

Context

Your Environment

| software | version |
| --- | --- |
| Babel | 6.18.0 (via REPL) |
| node | n/a |
| npm | n/a |
| Operating System | n/a |

---

/cc @mathiasbynens (for his expertise 😄 )

[DrewML]

Nice catch! I think this is a bug in Babylon, based on the tree generated in ASTExplorer. The escape sequences are being translated to an Identifier of break (which itself is also invalid).

[mathiasbynens]

This is something that web browsers used to allow despite it not being mentioned in the spec. As of ES6/ES2015, this behavior is explicitly non-conforming. More background from https://mathiasbynens.be/notes/javascript-identifiers:

[…] For example, var v\u0061r = 42; alert(va\u0072); would alert 42. This is very confusing, so I wouldn’t recommend relying on this hack. Luckily, it looks like the ECMAScript 6 spec will explicitly make this behavior non-conforming. Firefox/Spidermonkey, Safari/JavaScriptCore, and IE/Chakra have already dropped this behavior.

[DrewML]

@dashed I made a copy of this issue in the Babylon repository, since it is going to need a fix in the parser. Thanks again for the report!.