Oilpan: update WebNotification and WebDOMMediaStreamTrack type conversions

Oilpan: update WebNotification and WebDOMMediaStreamTrack type conversions

These two public/web/ types hadn't been updated to use transition types
over their underlying Oilpan-converted objects.

R=tkent@chromium.org,haraken@chromium.org
BUG=357163
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175540

[sof, 2014-06-04 20:08:35]

Please take a look.

[sof, 2014-06-04 21:21:08]

https://codereview.chromium.org/310393004/diff/20001/public/web/WebNotificati...
File public/web/WebNotification.h (right):

https://codereview.chromium.org/310393004/diff/20001/public/web/WebNotificati...
public/web/WebNotification.h:96: WebNotification(const
WTF::RawPtr<WebCore::Notification>&);
I want to go further here and use a straight pointer, but WebPrivatePtr's set of
assign() overloads lead to ambiguity / failure to resolve.

Not normalizing the template argument types wrt 'const' is one issue for
WebPrivatePtr<T>, but the harder ambiguity problem is how to support the
transition type RawPtr<T> (and its implicit conversions) alongside 'bare'
pointers.

Hence opting for RawPtr<WebCore::Notification> here for the time being.

[tkent, 2014-06-04 23:30:24]

lgtm

[haraken, 2014-06-05 00:51:39]

LGTM

[sof, 2014-06-05 05:35:44]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2014-06-05 05:36:54]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sigbjornf@opera.com/310393004/20001

[commit-bot: I haz the power, 2014-06-05 06:53:27]

Change committed as 175540

[Mads Ager (chromium), 2014-06-06 06:45:17]

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
File Source/web/NotificationPresenterImpl.cpp (right):

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
Source/web/NotificationPresenterImpl.cpp:82: return
m_presenter->show(RawPtr<Notification>(notification));
I think these methods already take actual raw pointers, so can we just remove
the conversion here and pass the pointer as is?

[sof, 2014-06-06 06:50:29]

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
File Source/web/NotificationPresenterImpl.cpp (right):

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
Source/web/NotificationPresenterImpl.cpp:82: return
m_presenter->show(RawPtr<Notification>(notification));
On 2014/06/06 06:45:17, Mads Ager (chromium) wrote:
> I think these methods already take actual raw pointers, so can we just remove
> the conversion here and pass the pointer as is?

No, that runs into issues with converting a Notification* to a const
WebNotification&. We only have conversions over (const RawPtr<Notification>).

This is what I referred to in https://codereview.chromium.org/310393004/#msg2

[Mads Ager (chromium), 2014-06-06 07:13:20]

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
File Source/web/NotificationPresenterImpl.cpp (right):

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
Source/web/NotificationPresenterImpl.cpp:82: return
m_presenter->show(RawPtr<Notification>(notification));
On 2014/06/06 06:50:28, sof wrote:
> On 2014/06/06 06:45:17, Mads Ager (chromium) wrote:
> > I think these methods already take actual raw pointers, so can we just
remove
> > the conversion here and pass the pointer as is?
> 
> No, that runs into issues with converting a Notification* to a const
> WebNotification&. We only have conversions over (const RawPtr<Notification>).
> 
> This is what I referred to in https://codereview.chromium.org/310393004/#msg2

OK, thanks. :(

[sof, 2014-06-06 07:23:06]

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
File Source/web/NotificationPresenterImpl.cpp (right):

https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
Source/web/NotificationPresenterImpl.cpp:82: return
m_presenter->show(RawPtr<Notification>(notification));
On 2014/06/06 07:13:20, Mads Ager (chromium) wrote:
> On 2014/06/06 06:50:28, sof wrote:
> > On 2014/06/06 06:45:17, Mads Ager (chromium) wrote:
> > > I think these methods already take actual raw pointers, so can we just
> remove
> > > the conversion here and pass the pointer as is?
> > 
> > No, that runs into issues with converting a Notification* to a const
> > WebNotification&. We only have conversions over (const
RawPtr<Notification>).
> > 
> > This is what I referred to in
https://codereview.chromium.org/310393004/#msg2
> 
> OK, thanks. :(

There's stuff to be looked at wrt WebPrivatePtr<>, but I'd prefer to address
that separately.

(It is painfully slow (at this end) to make changes to WebPrivatePtr & test them
out.)

[Mads Ager (chromium), 2014-06-06 07:24:18]

On 2014/06/06 07:23:06, sof wrote:
>
https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
> File Source/web/NotificationPresenterImpl.cpp (right):
> 
>
https://codereview.chromium.org/310393004/diff/20001/Source/web/NotificationP...
> Source/web/NotificationPresenterImpl.cpp:82: return
> m_presenter->show(RawPtr<Notification>(notification));
> On 2014/06/06 07:13:20, Mads Ager (chromium) wrote:
> > On 2014/06/06 06:50:28, sof wrote:
> > > On 2014/06/06 06:45:17, Mads Ager (chromium) wrote:
> > > > I think these methods already take actual raw pointers, so can we just
> > remove
> > > > the conversion here and pass the pointer as is?
> > > 
> > > No, that runs into issues with converting a Notification* to a const
> > > WebNotification&. We only have conversions over (const
> RawPtr<Notification>).
> > > 
> > > This is what I referred to in
> https://codereview.chromium.org/310393004/#msg2
> > 
> > OK, thanks. :(
> 
> There's stuff to be looked at wrt WebPrivatePtr<>, but I'd prefer to address
> that separately.
> 
> (It is painfully slow (at this end) to make changes to WebPrivatePtr & test
them
> out.)

Thanks Sigbjørn, yes, that is painful. Addressing separately is just fine!