VideoCaptureHost/VideoCaptureControllerEventHandler cleanup

VideoCaptureHost/VideoCaptureControllerEventHandler cleanup

1. VideoCaptureControllerID is a struct inside VCCEventHandler, just
to wrap an int. It also defines operator< and operator== to enable
using it in std::map. This is not needed, so I replaced it with a
typedef int VideoCaptureControllerID;

And s/const VideoCaptureControllerID&/VideoCaptureControllerID/

This enabled VideoCaptureControllerEventHandler to become a
pure interface. Removed video_capture_controller_event_handler.cc.

2. VideoCaptureHost lives 100% in IO Thread, except ctor in UI thread.
There's this idiom on IPC message sending where there's two
methods OnBla() -> PostTask to DoBla() on IO -> Send msg.
But all methods are living on IO and Send is asynchronous anyway,
so those DoBla() dealing with Send() are folded into OnBla()
methods -- trhose remain they mean basically a sophisticated
DeleteLater().

3. A few methods use base::TimeTicks as param, when they can
use const base::TimeTicks&, replaced them (one less copy in the
critical capture path).

BUG=440843
Committed: https://crrev.com/abc12f001172a3df76199a07bc6972587a4f9775
Cr-Commit-Position: refs/heads/master@{#320933}

[mcasas, 2015-03-17 03:26:59]

Patchset #3 (id:40001) has been deleted

[mcasas, 2015-03-17 03:42:39]

mcasas@chromium.org changed reviewers:
+ perkj@chromium.org, tommi@chromium.org

[mcasas, 2015-03-17 03:42:39]

tommi@/perkj@ PTAL.

[perkj_chrome, 2015-03-17 06:46:59]

lgtm if the extra bindtocurrentloop is removed or clarified why its needed.

https://codereview.chromium.org/1017503002/diff/60001/content/browser/rendere...
File content/browser/renderer_host/media/video_capture_host.cc (right):

https://codereview.chromium.org/1017503002/diff/60001/content/browser/rendere...
content/browser/renderer_host/media/video_capture_host.cc:201:
media::BindToCurrentLoop(
Why do you do this extra bindtocurrentloop? OnControllerAdded already checks
DCHECK_CURRENTLY_ON(BrowserThread::IO);

[mcasas, 2015-03-17 15:57:07]

The CQ bit was checked by mcasas@chromium.org

[mcasas, 2015-03-17 15:57:08]

The patchset sent to the CQ was uploaded after l-g-t-m from perkj@chromium.org
Link to the patchset: https://codereview.chromium.org/1017503002/#ps80001
(title: "Remove unneeded BindToCurrentLoop()")

[commit-bot: I haz the power, 2015-03-17 15:57:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1017503002/80001

[mcasas, 2015-03-17 15:57:44]

https://codereview.chromium.org/1017503002/diff/60001/content/browser/rendere...
File content/browser/renderer_host/media/video_capture_host.cc (right):

https://codereview.chromium.org/1017503002/diff/60001/content/browser/rendere...
content/browser/renderer_host/media/video_capture_host.cc:201:
media::BindToCurrentLoop(
On 2015/03/17 06:46:59, perkj wrote:
> Why do you do this extra bindtocurrentloop? OnControllerAdded already checks
> DCHECK_CURRENTLY_ON(BrowserThread::IO);

Not needed indeed. I think I started by changing this
line before I decided it was worth tackling the whole
move-into-IO thing.

[commit-bot: I haz the power, 2015-03-17 17:30:03]

Committed patchset #4 (id:80001)

[commit-bot: I haz the power, 2015-03-17 17:30:39]

Patchset 4 (id:??) landed as
https://crrev.com/abc12f001172a3df76199a07bc6972587a4f9775
Cr-Commit-Position: refs/heads/master@{#320933}