Remove noteOn and noteOff methods from OscillatorNode

Remove noteOn and noteOff methods from OscillatorNode

The WebAudio spec removed these methods a year ago in favor of start
and stop.  Chrome had removed these methods for
AudioBuffserSourceNodes, but not OscillatorNodes.

BUG=430995
TESTS=dom-exceptions updated

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=185070

[Raymond Toy, 2014-11-06 19:37:33]

rtoy@chromium.org changed reviewers:
+ kbr@chromium.org

[Raymond Toy, 2014-11-06 19:37:34]

PTAL

[Ken Russell (switch to Gerrit), 2014-11-07 00:50:39]

Do you have usage statistics for these methods?

Is it worth deprecating them first?

Have you run various popular Web Audio content (in particular, using the
OscillatorNode) to ensure it's not relying on these methods?

[Raymond Toy, 2014-11-07 03:19:15]

On 2014/11/07 00:50:39, Ken Russell wrote:
> Do you have usage statistics for these methods?
> 
> Is it worth deprecating them first?
> 
> Have you run various popular Web Audio content (in particular, using the
> OscillatorNode) to ensure it's not relying on these methods?

Unfortunately, there are no statistics.  

I can deprecate first if you want, but since AudioBufferSourceNode already
removed them, it seems odd that Oscillators will have it when
AudioBufferSource's don't.

I have not run this against popular content.

[Raymond Toy, 2014-11-07 16:55:11]

On 2014/11/07 03:19:15, Raymond Toy wrote:
> On 2014/11/07 00:50:39, Ken Russell wrote:
> > Do you have usage statistics for these methods?
> > 
> > Is it worth deprecating them first?
> > 
> > Have you run various popular Web Audio content (in particular, using the
> > OscillatorNode) to ensure it's not relying on these methods?
> 
> Unfortunately, there are no statistics.  
> 
> I can deprecate first if you want, but since AudioBufferSourceNode already
> removed them, it seems odd that Oscillators will have it when
> AudioBufferSource's don't.
> 
> I have not run this against popular content.

Oh, there are some statistics for noteOn, but I don't know if it's for all
noteOn or just for AudioBufferSource.noteOn.  But as of Nov 1, it's 0.0004
percent. (Peak usage appears to be about 0.0145 percent.)

[Ken Russell (switch to Gerrit), 2014-11-08 00:38:49]

On 2014/11/07 16:55:11, Raymond Toy wrote:
> > I have not run this against popular content.

Please do so before committing. Thanks.

> Oh, there are some statistics for noteOn, but I don't know if it's for all
> noteOn or just for AudioBufferSource.noteOn.  But as of Nov 1, it's 0.0004
> percent. (Peak usage appears to be about 0.0145 percent.)

OK. LGTM

[Raymond Toy, 2014-11-10 22:35:10]

On 2014/11/08 00:38:49, Ken Russell wrote:
> On 2014/11/07 16:55:11, Raymond Toy wrote:
> > > I have not run this against popular content.
> 
> Please do so before committing. Thanks.

Ran some tests, but it's hard to tell if they're using oscillators or not
without examining all of the code. But for the few demos that I know use
oscillators, everything is working or is not working for other reasons. (For
example, setting the oscillator type to a number instead of a string.)

> 
> > Oh, there are some statistics for noteOn, but I don't know if it's for all
> > noteOn or just for AudioBufferSource.noteOn.  But as of Nov 1, it's 0.0004
> > percent. (Peak usage appears to be about 0.0145 percent.)
> 
> OK. LGTM

[Raymond Toy, 2014-11-10 22:35:18]

The CQ bit was checked by rtoy@chromium.org

[commit-bot: I haz the power, 2014-11-10 22:35:55]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/707753006/1

[commit-bot: I haz the power, 2014-11-10 22:37:58]

Committed patchset #1 (id:1) as 185070