Update remoting/ to use WeakPtr<T>::get() instead of implicit "operator T*"

Update remoting/ to use WeakPtr<T>::get() instead of implicit "operator T*"

BUG=245942
R=wez@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=203796

[akalin, 2013-06-03 21:46:26]

+wez for review

may end up TBRing darin, as this is an automated pass.

Worth noting that there will be another automated pass to convert weak_ptr.get()
back to weak_ptr in a boolean context.

[Wez, 2013-06-03 22:02:54]

AFAICT all of the affected lines are already Boolean checks of the WeakPtr, with
one exception which just needs changing to a Boolean (see below).  Can we do
that rather than change them all to .get() and back again, please? :)

https://codereview.chromium.org/15927028/diff/1/remoting/client/frame_consume...
File remoting/client/frame_consumer_proxy.cc (right):

https://codereview.chromium.org/15927028/diff/1/remoting/client/frame_consume...
remoting/client/frame_consumer_proxy.cc:60: DCHECK(frame_consumer_.get() ==
NULL);
This should become a boolean.

[akalin, 2013-06-03 22:07:24]

On 2013/06/03 22:02:54, Wez wrote:
> AFAICT all of the affected lines are already Boolean checks of the WeakPtr,
with
> one exception which just needs changing to a Boolean (see below).  Can we do
> that rather than change them all to .get() and back again, please? :)

The plan is to convert weak_ptr to weak_ptr.get(), remove the implicit T*
operator, add the conversion-to-bool, then convert the boolean contexts back.

It's much easier to do it this way instead of leaving the boolean contexts
alone. Sorry. :/

>
https://codereview.chromium.org/15927028/diff/1/remoting/client/frame_consume...
> File remoting/client/frame_consumer_proxy.cc (right):
> 
>
https://codereview.chromium.org/15927028/diff/1/remoting/client/frame_consume...
> remoting/client/frame_consumer_proxy.cc:60: DCHECK(frame_consumer_.get() ==
> NULL);
> This should become a boolean.

It's worth noting that the second pass should take care of this case.

[Wez, 2013-06-03 22:17:17]

Assuming there's some reason why we can't leave Boolean uses of |weak_ptr| as
|weak_ptr| without breaking the |weak_ptr.get()| -> |weak_ptr| pass, LGTM!

[akalin, 2013-06-03 22:19:23]

On 2013/06/03 22:17:17, Wez wrote:
> Assuming there's some reason why we can't leave Boolean uses of |weak_ptr| as
> |weak_ptr| without breaking the |weak_ptr.get()| -> |weak_ptr| pass, LGTM!

Thanks. Yeah, the core reason is that a class can't have the operator bool() and
operator T*() at the same time, necessitating multiple passes. Sad.

[akalin, 2013-06-04 00:02:20]

Committed patchset #1 manually as r203796 (presubmit successful).