VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in
https://crrev.com/2365223002/

TEST=
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo

BUG=650113
Committed: https://crrev.com/25f8a1a991cb03ab2c386f25603b9bb468cde68c
Cr-Commit-Position: refs/heads/master@{#421824}

[mcasas, 2016-09-28 22:23:37]

Description was changed from

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

wip

BUG=
==========

to

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

wip

Follow up to some comments in https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo
==========

[mcasas, 2016-09-28 22:26:00]

Patchset #1 (id:1) has been deleted

[mcasas, 2016-09-28 22:39:45]

Description was changed from

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

wip

Follow up to some comments in https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo
==========

to

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in 
https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo
==========

[mcasas, 2016-09-28 23:03:26]

Description was changed from

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in 
https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo
==========

to

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in 
https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo

BUG=650113
==========

[mcasas, 2016-09-28 23:13:42]

mcasas@chromium.org changed reviewers:
+ chfremer@chromium.org, miu@chromium.org

[mcasas, 2016-09-28 23:13:42]

chfremer@ PTAL

miu@ FYI (because of offline conversation 
about DeInit()-derived expectations)

[chfremer, 2016-09-28 23:28:09]

lgtm % comment

https://codereview.chromium.org/2377163002/diff/20001/content/renderer/media/...
File content/renderer/media/video_capture_impl.cc (right):

https://codereview.chromium.org/2377163002/diff/20001/content/renderer/media/...
content/renderer/media/video_capture_impl.cc:122: // 
message_filter_->AddDelegate(this);
Can be removed?

[mcasas, 2016-09-28 23:49:52]

https://codereview.chromium.org/2377163002/diff/20001/content/renderer/media/...
File content/renderer/media/video_capture_impl.cc (right):

https://codereview.chromium.org/2377163002/diff/20001/content/renderer/media/...
content/renderer/media/video_capture_impl.cc:122: // 
message_filter_->AddDelegate(this);
On 2016/09/28 23:28:09, chfremer wrote:
> Can be removed?

oops, yeah, done.

[mcasas, 2016-09-29 00:57:37]

The CQ bit was checked by mcasas@chromium.org

[mcasas, 2016-09-29 00:57:38]

The patchset sent to the CQ was uploaded after l-g-t-m from
chfremer@chromium.org
Link to the patchset: https://codereview.chromium.org/2377163002/#ps40001
(title: "chfremer@ nit")

[commit-bot: I haz the power, 2016-09-29 00:58:11]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-29 01:44:52]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-29 01:44:53]

Failed to apply patch for content/renderer/media/video_capture_impl_manager.cc:
While running git apply --index -3 -p1;
  error: patch failed: content/renderer/media/video_capture_impl_manager.cc:48
  Falling back to three-way merge...
  Applied patch to 'content/renderer/media/video_capture_impl_manager.cc' with
conflicts.
  U content/renderer/media/video_capture_impl_manager.cc

Patch:       content/renderer/media/video_capture_impl_manager.cc
Index: content/renderer/media/video_capture_impl_manager.cc
diff --git a/content/renderer/media/video_capture_impl_manager.cc
b/content/renderer/media/video_capture_impl_manager.cc
index
1c052d70415517084a8a689cf3e868421d39d8fe..1293df831d4b6ff62e08fe8c524fa9f37a2c412c
100644
--- a/content/renderer/media/video_capture_impl_manager.cc
+++ b/content/renderer/media/video_capture_impl_manager.cc
@@ -48,9 +48,6 @@ VideoCaptureImplManager::~VideoCaptureImplManager() {
   // Forcibly release all video capture resources.
   for (const auto& device : devices_) {
     VideoCaptureImpl* const impl = device.second.second;
-    ChildProcess::current()->io_task_runner()->PostTask(
-        FROM_HERE,
-        base::Bind(&VideoCaptureImpl::DeInit, base::Unretained(impl)));
     ChildProcess::current()->io_task_runner()->DeleteSoon(FROM_HERE, impl);
   }
   devices_.clear();
@@ -63,11 +60,11 @@ base::Closure VideoCaptureImplManager::UseDevice(
   const VideoCaptureDeviceMap::iterator it = devices_.find(id);
   if (it == devices_.end()) {
     impl = CreateVideoCaptureImplForTesting(id, filter_.get());
-    if (!impl)
-      impl = new VideoCaptureImpl(id, filter_.get());
+    if (!impl) {
+      impl = new VideoCaptureImpl(id, filter_.get(),
+                                  ChildProcess::current()->io_task_runner());
+    }
     devices_[id] = std::make_pair(1, impl);
-    ChildProcess::current()->io_task_runner()->PostTask(
-        FROM_HERE, base::Bind(&VideoCaptureImpl::Init,
base::Unretained(impl)));
   } else {
     ++it->second.first;
   }
@@ -162,9 +159,6 @@ void VideoCaptureImplManager::UnrefDevice(
   --it->second.first;
   if (!it->second.first) {
     devices_.erase(id);
-    ChildProcess::current()->io_task_runner()->PostTask(
-        FROM_HERE,
-        base::Bind(&VideoCaptureImpl::DeInit, base::Unretained(impl)));
     ChildProcess::current()->io_task_runner()->DeleteSoon(FROM_HERE, impl);
   }
 }

[mcasas, 2016-09-29 05:29:54]

Patchset #3 (id:60001) has been deleted

[mcasas, 2016-09-29 05:30:23]

Patchset #3 (id:80001) has been deleted

[mcasas, 2016-09-29 05:36:40]

Patchset #3 (id:100001) has been deleted

[mcasas, 2016-09-29 05:37:03]

The CQ bit was checked by mcasas@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-29 05:37:14]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[mcasas, 2016-09-29 05:38:05]

The CQ bit was unchecked by mcasas@chromium.org

[miu, 2016-09-29 06:59:41]

lgtm % style issue:

https://codereview.chromium.org/2377163002/diff/120001/content/renderer/media...
File content/renderer/media/video_capture_impl.h (right):

https://codereview.chromium.org/2377163002/diff/120001/content/renderer/media...
content/renderer/media/video_capture_impl.h:54: const
scoped_refptr<base::SingleThreadTaskRunner>& io_task_runner);
Please use new smart pointer arg-passing style: The 3rd arg should be passed by
value, not by const-ref:

  scoped_refptr<base::SingleThreadTaskRunner> io_task_runner

Then, in the ctor initializer list:

  io_task_runner_(std::move(io_task_runner))

(to prevent the excessive ref count inc+dec operation)

[mcasas, 2016-09-29 14:40:29]

The CQ bit was checked by mcasas@chromium.org

[mcasas, 2016-09-29 14:40:30]

The patchset sent to the CQ was uploaded after l-g-t-m from
chfremer@chromium.org, miu@chromium.org
Link to the patchset: https://codereview.chromium.org/2377163002/#ps140001
(title: "miu@s comment on passing |io_task_runner| argument")

[commit-bot: I haz the power, 2016-09-29 14:40:40]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[mcasas, 2016-09-29 14:42:41]

https://codereview.chromium.org/2377163002/diff/120001/content/renderer/media...
File content/renderer/media/video_capture_impl.h (right):

https://codereview.chromium.org/2377163002/diff/120001/content/renderer/media...
content/renderer/media/video_capture_impl.h:54: const
scoped_refptr<base::SingleThreadTaskRunner>& io_task_runner);
On 2016/09/29 06:59:40, miu wrote:
> Please use new smart pointer arg-passing style: The 3rd arg should be passed
by
> value, not by const-ref:
> 
>   scoped_refptr<base::SingleThreadTaskRunner> io_task_runner
> 
> Then, in the ctor initializer list:
> 
>   io_task_runner_(std::move(io_task_runner))
> 
> (to prevent the excessive ref count inc+dec operation)

Done.

[commit-bot: I haz the power, 2016-09-29 15:24:02]

Description was changed from

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in 
https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo

BUG=650113
==========

to

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in 
https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo

BUG=650113
==========

[commit-bot: I haz the power, 2016-09-29 15:24:04]

Committed patchset #4 (id:140001)

[commit-bot: I haz the power, 2016-09-29 15:25:45]

Description was changed from

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in 
https://crrev.com/2365223002/

TEST= 
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo

BUG=650113
==========

to

==========
VideoCaptureImpl cleanup: merge ctor+Init() and DeInit()+dtor.

VideoCaptureImpl::ctor and Init() are always called in sequence,
and so are DeInit() and the destructor.  This CL just bundles
them. Some strange comments about threading are made
irrelevant. This CL follows up on some comments in
https://crrev.com/2365223002/

TEST=
./out/gn/content_unittests --gtest_filter=VideoCaptureImpl*
and video capture WAI using any demo

BUG=650113

Committed: https://crrev.com/25f8a1a991cb03ab2c386f25603b9bb468cde68c
Cr-Commit-Position: refs/heads/master@{#421824}
==========

[commit-bot: I haz the power, 2016-09-29 15:25:47]

Patchset 4 (id:??) landed as
https://crrev.com/25f8a1a991cb03ab2c386f25603b9bb468cde68c
Cr-Commit-Position: refs/heads/master@{#421824}