It should set the value to the empty string if the passed value is null or a Date with NaN value. Otherwise, it should throw a TypeError, per WebIDL spec.

(I forgot to refresh the patch...)

Comment on attachment 697485 [details] [diff] [review] Patch >+ if (aDate.isNull()) { isNullOrUndefined(), please. >+ ok(caught, "Setting " + value + " to .valueAsDate should throw"); Either s/Setting/Assigning/ or "setting .valueAsDate to " + value + " should throw". r=me with those nits.

(In reply to Boris Zbarsky (:bz) from comment #2) > Comment on attachment 697485 [details] [diff] [review] > Patch > > >+ if (aDate.isNull()) { > > isNullOrUndefined(), please. The HTML specs say: "if the new value is null or a Date object representing the NaN time value, then set the value of the element to the empty string". As far as I understand it, that means "undefined" should be considered as a TypeError. Or did I missed something? I've added a test for this to make it clearer that undefined should throw.

> "if the new value is null That's the IDL value, not the JS value. Nullable IDL values turn both JS null and JS undefined into null IDL values. If you were actually using our WebIDL bindings, this would happen automatically before you ever got into this code. Specifically, see http://dev.w3.org/2006/webapi/WebIDL/#es-nullable-type step 3 under "An ECMAScript value V is converted to an IDL nullable type T? value (where T is the inner type) as follows".

Fixed. Thanks for the explanations :)

Comment on attachment 697485 [details] [diff] [review] Patch [Approval Request Comment] Bug caused by (feature/regressing bug #): bug 769370 User impact if declined: adding a new feature with a known bug Testing completed (on m-c, etc.): m-c Risk to taking this patch (and alternatives if risky): low, we could introduce another bug in the feature by fixing that one but we have a pretty good test coverage String or UUID changes made by this patch: none

Comment on attachment 697485 [details] [diff] [review] Patch Low risk uplift for a new feature in Fx20.Approving on aurora.

See bug 866440 for documentation.