Offlist discussion...

Due to some crossed wires, François and I ended up having a discussion off list…. below is a replay of the discussion:  


Marcos Caceres:  
> | FYI, I think we are really getting a good picture here of what needs to
> be done. Hopefully you are also finding the discussion fruitful.
>  

François REMY:  
>  
> Yes, for sure. Having people who did write polyfills for HTML specs is
> really a plus for me, I'm sure I can learn a ton from discussing with you.
>  
> For example, I think the fact you can't construct DOM Exception is a very
> interesting case. This is clearly a place where we should try to provide a
> polyfill to create DOM Exceptions (even if it's ugly) and advertise the fact
> we would like this to be natively implemented.
>  
> I see the goal of this group as being exactly that: writing polyfills making
> the life of polyfills writer easier, and trying to get our polyfills
> implemented natively in browsers. The more polyfills are in use, the more
> they'll be tempted to optimize our cases.



Marcos Caceres:  
> So, last time I tried to this, here is where I ended up (dead-end):

> http://lists.w3.org/Archives/Public/public-script-coord/2012JanMar/0481.html


François REMY wrote:
> > This is not really a dead end, you made a perfectly valid polyfill :
> > http://jsfiddle.net/yDNgL/31/ (even if it may need tweaks). This is exactly
> > the kind of things we want in the "polyfill framework" we're going to build
> > via GitHub.
>  

> >  
> > After that, the fact we still haven't a constructor is a "minor" issue in
> > the sense we can polyfill it. I really belive that this kind of "tricks" is
> > going to be fixed as soon as we can get our "polyfill framework" used. At
> > least, this is the whole idea of "forward polyfills": we start by using an
> > hacky JS, and we continue to use it as long as no native implementation is
> > available. If enough people use your hacks, standardiation will happen.
>  


Marcos Caceres:   
>  
>  
> So, here is a complete mess of code for building WebIDL objects and Web IDL exceptions (not working, yet):
> https://github.com/marcoscaceres/ECMAScriptInJS/blob/master/WebIDL.js
>  
> But I spent quite a lot of time on it… we could see we can use it as a base.
  

--  
Marcos Caceres

Received on Monday, 12 November 2012 20:01:37 UTC