winaura: Do not create WindowObservers for BrowserPlugin.

winaura: Do not create WindowObservers for BrowserPlugin.

WindowObserver is not necessary for BrowserPlugin.
Moreover, if BrowserPlugin navigates to an Interstitial page,
tearing down its embedder causes a crash b/c it tries
to cast the Interstitial's WC's view to WCViewAura, which
actually is a WCViewGuest. The ToRenderWidgetHostViewAura does
not work in this scenario because the RWH's process points to
Interstitial's RenderProcess, whose IsGuest() is false. The
simpler fix hence is to get rid of WindowObserver for
BrowserPlugin.

This will also fix a flaky test on windows:
WebViewTest.InterstitialTeardown.


BUG=316298, 297014
Test=On Windows Aura, load an app with a <webview> pointing
to a page that shows an interstitial, e.g. try navigating
to https://test.com. After the interstitial is visible, close
the app. The fix would make the browser not crash.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=235600

[jam, 2013-11-14 17:42:15]

lgtm

https://codereview.chromium.org/72283002/diff/1/content/browser/web_contents/...
File content/browser/web_contents/web_contents_view_aura.cc (right):

https://codereview.chromium.org/72283002/diff/1/content/browser/web_contents/...
content/browser/web_contents/web_contents_view_aura.cc:101:
DCHECK(!process->IsGuest());
is this dcheck conclusive? i.e. can we still end up with a RHVGuest passed into
here that doesn't have RenderProcessHost::IsGuest?

https://codereview.chromium.org/72283002/diff/1/content/browser/web_contents/...
content/browser/web_contents/web_contents_view_aura.cc:1131: 
nit: undo

[lazyboy, 2013-11-14 19:04:50]

https://codereview.chromium.org/72283002/diff/1/content/browser/web_contents/...
File content/browser/web_contents/web_contents_view_aura.cc (right):

https://codereview.chromium.org/72283002/diff/1/content/browser/web_contents/...
content/browser/web_contents/web_contents_view_aura.cc:101:
DCHECK(!process->IsGuest());
On 2013/11/14 17:42:15, jam wrote:
> is this dcheck conclusive? i.e. can we still end up with a RHVGuest passed
into
> here that doesn't have RenderProcessHost::IsGuest?

IsGuest() = true here would mean either a. we incorrectly added WindowObserver
to a guest or b. for a WebContents, we navigated from a non-guest RVH to a guest
RVH.
Both of them is not expected in theory, so this DCHECK is covering these two
cases.

https://codereview.chromium.org/72283002/diff/1/content/browser/web_contents/...
content/browser/web_contents/web_contents_view_aura.cc:1131: 
On 2013/11/14 17:42:15, jam wrote:
> nit: undo

Done.

[lazyboy, 2013-11-14 21:34:54]

Can you take another quick look? Thanks.

https://chromiumcodereview.appspot.com/72283002/diff/1/content/browser/web_co...
File content/browser/web_contents/web_contents_view_aura.cc (right):

https://chromiumcodereview.appspot.com/72283002/diff/1/content/browser/web_co...
content/browser/web_contents/web_contents_view_aura.cc:101:
DCHECK(!process->IsGuest());
On 2013/11/14 19:04:50, lazyboy wrote:
> On 2013/11/14 17:42:15, jam wrote:
> > is this dcheck conclusive? i.e. can we still end up with a RHVGuest passed
> into
> > here that doesn't have RenderProcessHost::IsGuest?
> 
> IsGuest() = true here would mean either a. we incorrectly added WindowObserver
> to a guest or b. for a WebContents, we navigated from a non-guest RVH to a
guest
> RVH.
> Both of them is not expected in theory, so this DCHECK is covering these two
> cases.

TouchEditable also uses this method, which means the DCHECK cannot be here,
reverted this change here.

[jam, 2013-11-15 17:37:43]

lgtm

[commit-bot: I haz the power, 2013-11-15 17:55:50]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/lazyboy@chromium.org/72283002/200001

[commit-bot: I haz the power, 2013-11-15 22:28:21]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/lazyboy@chromium.org/72283002/200001

[commit-bot: I haz the power, 2013-11-16 01:16:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/lazyboy@chromium.org/72283002/200001

[commit-bot: I haz the power, 2013-11-18 04:11:11]

Change committed as 235600