1317379 - B.3.4 shouldn't be applied for generator and async functions

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[André Bargull [:anba]]

`if (true) function* g() {}` and `if (true) async function f() {}` should both throw a SyntaxError because B.3.4 only applies for FunctionDeclaration.


ES2017 spec:
https://tc39.github.io/ecma262/#sec-functiondeclarations-in-ifstatement-statement-clauses

Comment 1

[Jeff Walden [:Waldo]]

Attachment: Disallow generator functions and async functions as direct children of if/else

Comment 2

[Tooru Fujisawa [:arai]]

Comment on attachment 8810674 [details] [diff] [review]
Disallow generator functions and async functions as direct children of if/else

Review of attachment 8810674 [details] [diff] [review]:
-----------------------------------------------------------------

Thanks!

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/4c223c3ddf73
Disallow generator functions and async functions as direct children of if/else.  r=arai

Comment 4

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/4c223c3ddf73

Comment 5

[Hannes Verschore [:h4writer]]

Sorry about all these requests. Not sure if it will work to backport everything. Let us at least try to backport the easy / lowest risk ones. I also assume that all these issues probably go back before F52? Though I don't think we can / need to backport to beta. I assume only serious security issues are allowed currently.

Comment 6

[André Bargull [:anba]]

I think Waldo should decide if it's safe to backport this one.

Comment 7

[Jeff Walden [:Waldo]]

I can't think of a reason why we'd need to backport this.  It's enough of an edge case that I doubt anyone will come to depend on this bugginess, and that seems like the only reason to accelerate a fix.

Even if we did decide this was important to fix earlier, the changes span sufficient number of lines, and are picky enough about peeking and lookahead and such, that it presents *some* risk of containing a mistake.  And when there's not much to gain, some risk is too much risk.

I say we don't even consider taking this.