In CloseEvent.idl, [ConvertingNullStringTo] is a typo of [ConvertNullStringTo]. In bug 78108, [ConvertNullStringTo] was renamed to [TreatReturnedNullStringAs]. In conclusion, we should replace [ConvertingNullStringTo] with [TreatReturnedNullStringAs].
Created attachment 126464 [details] Patch
Comment on attachment 126464 [details] Patch Have you cross-checked this with the spec?
Created attachment 126467 [details] Patch
(In reply to comment #2) > (From update of attachment 126464 [details]) > Have you cross-checked this with the spec? The spec says "The reason attribute must return the value it was initialized to. When the object is created, this attribute must be initialized to empty string.". http://dev.w3.org/html5/websockets/#event-definitions So we should simply remove [ConvertingNullStringTo]. Uploaded a revised patch.
Comment on attachment 126467 [details] Patch Clearing flags on attachment: 126467 Committed r107382: <http://trac.webkit.org/changeset/107382>
All reviewed patches have been landed. Closing bug.
All sounds fine, but what about test coverage?
(In reply to comment #7) > All sounds fine, but what about test coverage? Darin: The following existing test is in close-event-constructor.html: shouldBeEqualToString("new CloseEvent('eventType', { reason: '' }).reason", ""); Fortunately, we didn't need to change the test result (since [ConvertingNullStringTo] was typo and had not been working at all).
(In reply to comment #8) > Fortunately, we didn't need to change the test result So if we had added the correctly spelled attribute the test would have failed?
(In reply to comment #9) > (In reply to comment #8) > > Fortunately, we didn't need to change the test result > > So if we had added the correctly spelled attribute the test would have failed? Yes.