On 2014/08/09 at 03:53:29, caitpotter88 wrote:
On 2014/08/09 03:14:57, caitp wrote:
> On 2014/08/09 00:01:54, caitp wrote:
> > There are issues with unscopables tests, I will have to see if there's
more to
> > be done on my end, or if it's blocked on some changes to unscopables.
> >
> > Will look closer at that in a few
>
> Okay, what happens is `Object.defineProperty( [object Array], [object
Symbol],
> ... )` throws due to trying to convert the Symbol to an index. If this
operation
> returned NaN previously, then it's hard to imagine how this could have
possibly
> worked correctly before, but I'm still looking!

 From the spec, it looks like we're right to throw here, actually. This is
possibly a bug in the spec, but the result of performing ToUint32() on a Symbol should always throw --- for now I'll add code to ignore the Array case, I guess.

The spec just says "Else if P is an array index, then" which seems correct to
me.

The problem is that the way we detect the above is that we do a ToUint32. If we
wrap this in a `typeof p !== 'symbol` I think we are good to go.

https://code.google.com/p/chromium/codesearch#chromium/src/v8/src/v8natives.js&l=927

https://codereview.chromium.org/454233002/

--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
--- You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to