this also should be reported to bugs.ecmascript.org
This is a very good catch.
Allen
On Jun 24, 2012, at 6:45 AM, Yusuke Suzuki wrote:
> Hello everyone,
>
> In 15.4 ArrayCreate algorithm step 7, we use [[DefineOwnProperty]] of A.
>
> Call the [[DefineOwnProperty]] internal method of A with arguments
> "length", Property Desciptor {[[Value]]: length, [[Writable]]: true,
> [[Enumerable]]: false, [[Configurable]]: false}, and false.
>
> ArrayCreate set A's [[DefineOwnProperty]] to 15.4.5.1 method in step 4, so
> step 7 calls 15.4.5.1 Array's [[DefineOwnProperty]].
>
> Array's [[DefineOwnProperty]] assumes that 'length' property is always set.
> Seeing 15.4.5.1 step 1
>
> Let oldLenDesc be the result of calling the [[GetOwnProperty]] internal
> method of A passing "length" as the argument. The result will never be
> undefined or an accessor descriptor because Array objects are created with a
> length data property that cannot be deleted or reconfigured.
>
> But, in ArrayCreate phase (step 7), A's 'length' property is not set yet.
> Then assertion in 15.4.5.1 step 1 always fails.
>
> So we should use the default [[DefineOwnProperty]] internal method instead of
> Array's.
> I suggest modifying ArrayCreate step 7,
>
> Call the default [[DefineOwnProperty]] internal method (8.12.9) on A
> with arguments "length", Property Desciptor {[[Value]]: length, [[Writable]]:
> true, [[Enumerable]]: false, [[Configurable]]: false}, and false.
>
> Regards,
> Yusuke Suzuki
> _______________________________________________
> es-discuss mailing list
> [email protected]
> https://mail.mozilla.org/listinfo/es-discuss
_______________________________________________
es-discuss mailing list
[email protected]
https://mail.mozilla.org/listinfo/es-discuss
