17722 2012-07-09 11:20:40 +0000 All interface objects should have a 'length' property 2012-07-10 08:29:24 +0000 1 1 1 Unclassified WebAppsWG WebIDL unspecified All All CLOSED FIXED P2 normal --- 1 Ms2ger cam annevk bzbarsky mike public-script-coord public-webapps-bugzilla oldest_to_newest 69733 0 Ms2ger 2012-07-09 11:20:40 +0000 In V8, Opera and Gecko's new bindings, they have a 'length' property set to 0. In IE and old Gecko, they don't have one. 69734 1 annevk 2012-07-09 11:34:06 +0000 Safari 5.1.7 also does not have one. 69761 2 bzbarsky 2012-07-10 06:43:58 +0000 Per spec, interface objects are always functions, functions in ES5 have a length. WebIDL's only business here is defining what that length should be. It does so for cases when there is a [Constructor]; presumably this bug is about the cases when there isn't one? 69764 3 cam 2012-07-10 07:25:09 +0000 Right, and it should be 0 in that case. http://dev.w3.org/cvsweb/2006/webapi/WebIDL/Overview.xml.diff?r1=1.559;r2=1.560;f=h 69768 4 Ms2ger 2012-07-10 08:29:24 +0000 (In reply to comment #2) > Per spec, interface objects are always functions, functions in ES5 have a > length. WebIDL's only business here is defining what that length should be. > It does so for cases when there is a [Constructor]; presumably this bug is > about the cases when there isn't one? And the fact that I was confused by the phrasing. (In reply to comment #3) > Right, and it should be 0 in that case. > > http://dev.w3.org/cvsweb/2006/webapi/WebIDL/Overview.xml.diff?r1=1.559;r2=1.560;f=h Thanks.