Use SpecialWrapFor over HTMLMediaElement

Use SpecialWrapFor over HTMLMediaElement.

To appropriately delegate the creation of a wrapper to
subclasses of HTMLMediaElement, annotate HTMLMediaElement's
interface with SpecialWrapFor.

R=haraken@chromium.org
BUG=379499
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180555

[sof, 2014-08-19 10:36:35]

Please take a look.

It seems unavoidable to add some custom code here.

[haraken, 2014-08-19 11:58:23]

Thanks for the fix!

Can we avoid adding the custom binding by using a [SpecialWrapFor] IDL
attribute?

+hoshi-san, shiino-san: FYI. This is the same issue we faced when moving
MediaControlsElement to Blink-in-JS :)

https://codereview.chromium.org/483863003/diff/1/LayoutTests/webaudio/mediael...
File LayoutTests/webaudio/mediaelementaudiosourcenode-wrapper.html (right):

https://codereview.chromium.org/483863003/diff/1/LayoutTests/webaudio/mediael...
LayoutTests/webaudio/mediaelementaudiosourcenode-wrapper.html:19: if (window.gc)

Nit: You don't need this check. js-test.js has window.gc.

[sof, 2014-08-19 12:22:58]

On 2014/08/19 11:58:23, haraken wrote:
> Thanks for the fix!
> 
> Can we avoid adding the custom binding by using a [SpecialWrapFor] IDL
> attribute?
> 

Very close&tempting, but SpecialWrapFor=Foo assumes a method isFoo(). The
element predicates are toplevel functions, not methods.

It seems overkill to add virtual predicates over HTMLMediaElement just for this.
Make the compiler know about instance testing for SpecialWrapFor types with an
"HTML" prefix?

[haraken, 2014-08-19 12:26:38]

On 2014/08/19 12:22:58, sof wrote:
> On 2014/08/19 11:58:23, haraken wrote:
> > Thanks for the fix!
> > 
> > Can we avoid adding the custom binding by using a [SpecialWrapFor] IDL
> > attribute?
> > 
> 
> Very close&tempting, but SpecialWrapFor=Foo assumes a method isFoo(). The
> element predicates are toplevel functions, not methods.

Thanks, understood...

> It seems overkill to add virtual predicates over HTMLMediaElement just for
this.
> Make the compiler know about instance testing for SpecialWrapFor types with an
> "HTML" prefix?

I'm fine with either approach. Both approach seems better than adding custom
bindings :)

[sof, 2014-08-19 13:09:33]

On 2014/08/19 12:26:38, haraken wrote:
> On 2014/08/19 12:22:58, sof wrote:
> > On 2014/08/19 11:58:23, haraken wrote:
> > > Thanks for the fix!
> > > 
> > > Can we avoid adding the custom binding by using a [SpecialWrapFor] IDL
> > > attribute?
> > > 
> > 
> > Very close&tempting, but SpecialWrapFor=Foo assumes a method isFoo(). The
> > element predicates are toplevel functions, not methods.
> 
> Thanks, understood...
> 
> > It seems overkill to add virtual predicates over HTMLMediaElement just for
> this.
> > Make the compiler know about instance testing for SpecialWrapFor types with
an
> > "HTML" prefix?
> 
> I'm fine with either approach. Both approach seems better than adding custom
> bindings :)

Yes, I wasn't aware of the intended usage of this newer attribute -- the
generated binding code now comes out the exact same as the custom code :)

Did that by adding HTMLMediaElement virtual predicates in the end (too many
special cases to consider to build this into code generation instead.) I hope
that's an acceptable addition.

[haraken, 2014-08-19 13:15:21]

On 2014/08/19 13:09:33, sof wrote:
> On 2014/08/19 12:26:38, haraken wrote:
> > On 2014/08/19 12:22:58, sof wrote:
> > > On 2014/08/19 11:58:23, haraken wrote:
> > > > Thanks for the fix!
> > > > 
> > > > Can we avoid adding the custom binding by using a [SpecialWrapFor] IDL
> > > > attribute?
> > > > 
> > > 
> > > Very close&tempting, but SpecialWrapFor=Foo assumes a method isFoo(). The
> > > element predicates are toplevel functions, not methods.
> > 
> > Thanks, understood...
> > 
> > > It seems overkill to add virtual predicates over HTMLMediaElement just for
> > this.
> > > Make the compiler know about instance testing for SpecialWrapFor types
with
> an
> > > "HTML" prefix?
> > 
> > I'm fine with either approach. Both approach seems better than adding custom
> > bindings :)
> 
> Yes, I wasn't aware of the intended usage of this newer attribute -- the
> generated binding code now comes out the exact same as the custom code :)
> 
> Did that by adding HTMLMediaElement virtual predicates in the end (too many
> special cases to consider to build this into code generation instead.) I hope
> that's an acceptable addition.

I agree that this is acceptable. LGTM.

I think we need to reconsider how the custom wraps should be organized when
shiino-san moves wrap() to 'virtual ScriptWrappable::wrap()'.

[sof, 2014-08-19 13:29:58]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2014-08-19 13:30:42]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sigbjornf@opera.com/483863003/20001

[commit-bot: I haz the power, 2014-08-19 14:12:26]

Committed patchset #2 (20001) as 180555