- From: Cameron McCormack <cam@mcc.id.au>
- Date: Wed, 28 Dec 2011 17:16:02 +1100
- To: Jonas Sicking <jonas@sicking.cc>
- CC: Ian Hickson <ian@hixie.ch>, Alex Russell <slightlyoff@google.com>, "public-script-coord@w3.org" <public-script-coord@w3.org>, Anne van Kesteren <annevk@opera.com>
Cameron McCormack:
> People will weight the pros and cons differently, I imagine. But for me
> based on the above I prefer (c).
Based on that, I have changed the spec in this direction so that we move 
towards accepting Function objects rather than Function objects or plain 
objects with properties of particular names.
I've incorporated a variation of the "callback" proposal (just the 
single function version) from 
http://www.w3.org/mid/4EE7F7F7.9080600@mcc.id.au without the function 
identifier, and this is what specification authors are now recommended 
to use in place of:
   [Callback,NoInterfaceObject]
   interface Blah {
     ...
   };
For new APIs (and older ones that don't support the property style), you 
would write for example:
   callback CheeseHasArrivedNotification = void (DOMString cheeseType);
   interface CheeseManager {
     void requestAsyncCheese(CheeseHasArrivedNotification n);
   };
and requestAsyncCheese would take only a Function object.
[Callback] is now gone altogether, and for the cases where we do need to 
support a callback object with a property, or any other cases that would 
currently be handled by [Callback], you define a callback interface. 
For EventListener, since we need it to continue to support objects with 
a handleEvent property, we would write:
   callback interface EventListener {
     void handleEvent(Event evt);
   };
with no need for [NoInterfaceObject] since callback interface don't 
cause interface objects to exist normally.
http://dev.w3.org/cvsweb/2006/webapi/WebIDL/Overview.xml.diff?r1=1.432;r2=1.433;f=h
Anne (being the original commentor), I'll assume you are OK with this 
change unless I hear otherwise.
Thanks,
Cameron
Received on Wednesday, 28 December 2011 06:16:59 UTC