13499 2011-08-01 11:35:57 +0000 Introduce "assume-null-if-not-object" flag 2011-10-01 08:50:04 +0000 1 1 1 Unclassified WebAppsWG WebIDL unspecified PC All RESOLVED FIXED P2 normal --- 13433 14351 14352 1 annevk cam bzbarsky jonas mike public-script-coord public-webapps-bugzilla oldest_to_newest 51897 0 annevk 2011-08-01 11:35:57 +0000 Event handler IDL attributes either take a Function object or null. However, the behavior in user agents is such that whenever you assign a non-Function null is assumed (e.g. "test", 2, []). Maybe this can be combined with [AllowAny] somehow? Not sure if that is worth it though. 51909 1 jonas 2011-08-01 17:26:22 +0000 Which user agents behave this way? I suspect Gecko does, but I'd be interested to know what others do. 51911 2 bzbarsky 2011-08-01 17:27:44 +0000 All of them, basically. See http://www.w3.org/Bugs/Public/show_bug.cgi?id=13433 (which should probably have been referenced in comment 0) for details. 51982 3 annevk 2011-08-02 11:26:44 +0000 addEventListener()'s EventListener also seems affected by this. Have not tested other callbacks, but did notice that the timeout API already accepts "any". 51988 4 bzbarsky 2011-08-02 13:40:18 +0000 > addEventListener()'s EventListener also seems affected by this. Only in some browsers. 55791 5 cam 2011-08-25 04:45:06 +0000 It wasn't clear to me reading through bug 13433 and this one again whether we want to treat as null (a) non-callable objects, (b) non-Function objects or (c) non-objects. I've assumed (a) for now and added [TreatNonCallableAsNull]. http://dev.w3.org/cvsweb/2006/webapi/WebIDL/Overview.html.diff?r1=1.355;r2=1.356;f=h Let me know if that works or needs changes.