Yes, Fat Arrows are still broken

[garthk]

(Tracking issue to save time for the curious and those they'd otherwise hound.)

Developers champing at the bit for native fat arrows? Worried about them being hidden behind a flag? Hounded by that one developer who insists they tried it and it's fine? Worried about the ominous warnings on the ES6 on io.js page?

really buggy or even broken features like Arrow Functions

Your first stop should be V8 issue 2700. Check to see if it's blocked (left column, just above "add a comment"):

• Not blocked? Perhaps fat arrows are on their way! Add a comment or update this text.
• Still blocked? Send the list below to that developer. Done!

v8 4.2.77.18 in io.js 2.0.0 still has trouble with:

• io.js Arrow functions' lexical binding is not working correctly #735, in which you can strip the binding from an arrow function
• V8 #4030, in which calling super() from an arrow function crashes the engine
• V8 #4031, in which using super.x from an arrow function throws ReferenceError

If you need fat arrows, consider transpilation.

[cjihrig]

Does this need to remain open to be used as a reference?

[Fishrock123]

Fat Arrows are still broken

Yeeep.

Does this need to remain open to be used as a reference?

I don't think so. Closing, hopefully people will search it. :)

[friday]

Not blocked any more. It will will be shipped with Chrome 45 https://twitter.com/malyw/status/608943596918702080. However according to https://kangax.github.io/compat-table/es6/ support is still lacking for lexical "super" binding and lexical "new.target" binding.

[LunaBorowska]

@friday: That's because new.target doesn't work in Chrome at all without flags (so it cannot work with arrows, because it doesn't even work outside of those), and lexical "super" binding is here, but classes require strict mode in Chrome for some reason (this is problem with classes, not arrow functions).

Would say it's implemented.

[friday]

Thanks for clearing out the confusion @xfix :)

[albertorestifo]

Now arrow functions are shipped in V8. When can we expect them in iojs?

[Fishrock123]

Now arrow functions are shipped in V8.

No not quite. They aren't in a stable v8 yet afaik (they are in 4.5). The latest stable is 4.4 -- anything can happen before then.

[albertorestifo]

@Fishrock123 Oh right! I forgot they shipped it in the latest Chrome Beta, not the stable. My bad!

[steelbrain]

It would be nice to have a beta build of iojs that uses the beta v8 branch.

[sjmueller]

+1 on beta builds with latest v8. It would also be awesome to have documentation in iojs about which v8 options actually have an implementation exposed. For example, it's very misleading to have --harmony_modules when enabling yields no benefits whatsoever to the developer.

We are getting to the point in the server side where transpilation may not be necessary to have your es2015+ cake and eat it too. But without more clarity on where the line is drawn for a particular release, it's a game of enabling the flag, getting an error, then hunting down the version of v8 that iojs is using, then searching for the latest features in v8, then looking back in the history to see if/when it was implemented, etc etc.

[bnoordhuis]

The V8 API changes frequently and is often broken in development. Maintaining a working build of io.js would be a a big time sink for little gain. Few people run unstable releases in any serious capacity and bug reports would be mostly useless because of the ever moving target.

[sjmueller]

Are some harmony flags blocked in iojs? Using this command:

node --es_staging --harmony_arrow_functions server.js

results in this error: SyntaxError: Unexpected token =>

This is why it would be great to know the status of each harmony flag for the current release of iojs, so that it's easy to understand what can be experimented with and what can't. Running iojs --v8-options | grep "in progress" gives you a list, but its just misleading. Like I said in the previous post, the harmony_modules is on the list, but I don't think it can be used in any shape or form. Or can it? I don't know; maybe I'm doing something wrong here.