Re: "var" declarations shadowing properties from Window.prototype

On Oct 5, 2012, at 6:32 PM, Boris Zbarsky <bzbarsky@MIT.EDU> wrote:

> On 10/5/12 7:35 PM, Cameron McCormack wrote:
> 
>> * What was the conclusion on whether properties for IDL operations also
>> need to move down to be own properties?
> 
> I don't know that we reached one.  I think Brendan wanted to move them down.  I don't think I have a strong opinion.  Not sure what Travis thinks.  Opera already has them down there, I think.  The Chrome/Safari folks are being quiet as usual...
> 

For Safari (or more broadly for WebKit-based ports that use JavaScriptCore):

- We are ok with eventually changing non-global-object Web IDL properties to be getters/setters on the prototype instead of magical properties on the object itself, but we expect we'll have to do a fair bit of work to avoid this being a performance regression for our DOM bindings (theoretically it could even be a speedup if we do it right).

- I don't think we have a problem with making the global object an exception.

- Our current behavior is that methods of the Window object are on the prototype but other special properties are on the Window object itself. I don't personally know of a web compat issue with doing it one way or another. There doesn't seem to be a compelling reason to change it.

WebKit V8 bindings often but not always match WebKit JSC bindings, but I can't speak for the opinions of Chrome or other V8-based WebKit ports.

Regards,
Maciej

Received on Wednesday, 10 October 2012 06:48:42 UTC