[DO NOT COMMIT] Random hacks to unblock work on replication.

[DO NOT COMMIT] Random hacks to unblock work on replication.

BUG=

[nasko, 2014-06-20 20:02:25]

Hey Daniel,
These are the hacks I've put in place to make progress on frame tree
replication. Ignore the commented out CRASH/NOT_REACHED, as I wanted to get as
deep as I could in debug builds without hitting those.
Thanks!
Nasko

[dcheng, 2014-06-20 20:49:12]

https://codereview.chromium.org/348833005/diff/1/Source/web/WebLocalFrameImpl...
File Source/web/WebLocalFrameImpl.cpp (right):

https://codereview.chromium.org/348833005/diff/1/Source/web/WebLocalFrameImpl...
Source/web/WebLocalFrameImpl.cpp:1662: bool isMainFrame = (parent() == 0);
I think Blink style would prefer bool isMainFrame = parent().

Yeah...

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp
File Source/web/WebViewImpl.cpp (right):

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp#n...
Source/web/WebViewImpl.cpp:1588: if (!mainFrameImpl())
Being less familiar with the rendering paths, is this something we're going to
need to handle for mirrored roots eventually? I feel like there are a lot of
locations that depend on a top-level FrameView, and I'm not sure we have a good
story for them yet.

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp#n...
Source/web/WebViewImpl.cpp:2533: if (frame->isRemoteFrame())
You can use WebFrame::fromFrame now, so you don't have to do this awkward check
(since the awkwardness is now centralized =)

[nasko, 2014-06-20 21:09:53]

Hey Ken,
There is a question that Daniel asked that I think you can handle best. Can you
look at it?
Thanks!
Nasko

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp
File Source/web/WebViewImpl.cpp (right):

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp#n...
Source/web/WebViewImpl.cpp:1588: if (!mainFrameImpl())
On 2014/06/20 20:49:12, dcheng wrote:
> Being less familiar with the rendering paths, is this something we're going to
> need to handle for mirrored roots eventually? I feel like there are a lot of
> locations that depend on a top-level FrameView, and I'm not sure we have a
good
> story for them yet.

I'll leave that for Ken to give input on.

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp#n...
Source/web/WebViewImpl.cpp:2533: if (frame->isRemoteFrame())
On 2014/06/20 20:49:12, dcheng wrote:
> You can use WebFrame::fromFrame now, so you don't have to do this awkward
check
> (since the awkwardness is now centralized =)

Yay! ?!

[abarth-chromium, 2014-06-20 23:31:55]

> Random hacks

...

[kenrb, 2014-06-23 15:21:19]

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp
File Source/web/WebViewImpl.cpp (right):

https://codereview.chromium.org/348833005/diff/1/Source/web/WebViewImpl.cpp#n...
Source/web/WebViewImpl.cpp:1588: if (!mainFrameImpl())
On 2014/06/20 21:09:53, nasko wrote:
> On 2014/06/20 20:49:12, dcheng wrote:
> > Being less familiar with the rendering paths, is this something we're going
to
> > need to handle for mirrored roots eventually? I feel like there are a lot of
> > locations that depend on a top-level FrameView, and I'm not sure we have a
> good
> > story for them yet.
> 
> I'll leave that for Ken to give input on.

Most of the places that access the top-level FrameView can be changed to instead
access the FrameView of the local Frame root (i.e. root of a connected subtree
of LocalFrames). For rendering that is typically going to be the case, anyway.
Certainly that is the case here, where the FrameView is being used to start a
layout.

I expect there to be exceptions but we can deal with those as we find them.

[kenrb, 2014-07-16 20:58:57]

I have reduced this to a smaller patch that is up for review:
https://codereview.chromium.org/397023003/

Closing this accordingly.