Update several WebContentsObserver loading callbacks to use RFH.

Update several WebContentsObserver loading callbacks to use RFH.

This code runs in the browser process, so there's no need to pass
around routing IDs for frames, since it can just use RFH directly.
This also simplifies the callbacks by removing several parameters
that are now redundant with RFH. This has the side effect of
making it impossible for tests to supply 'impossible' data to
WebContentsObserver callbacks.

Tests for WebContentsObserver loading callbacks can use the new
RenderFrameHostTester helper if they need to create subframe
RenderFrameHosts.

[email protected]
[email protected]

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=281241

[dcheng, 2014-07-01 18:58:26]

This updates three of the frame loading callbacks (DidStart, DidFail, and
DidCommit). There are a few more I could update, but this patch is getting
rather large and unwieldy already.

Regarding the addition of IsMainFrame(): I think this will be easier to remove
if we end up not needing it, rather than trying to retroactively fix call sites
if IsMainFrame() is no longer tied to GetParent().

[jam, 2014-07-02 16:02:42]

On 2014/07/01 18:58:26, dcheng (OOO) wrote:
> This updates three of the frame loading callbacks (DidStart, DidFail, and
> DidCommit). There are a few more I could update, but this patch is getting
> rather large and unwieldy already.
> 
> Regarding the addition of IsMainFrame(): I think this will be easier to remove
> if we end up not needing it, rather than trying to retroactively fix call
sites
> if IsMainFrame() is no longer tied to GetParent().

I think it's very confusing to have GetParent() return non-NULL if IsMainFrame
is true. Let's fix that instead of adding confusion. Most devs don't know what
webview is.

[dcheng, 2014-07-02 19:58:58]

PTAL. I've removed IsMainFrame() but I'm still trying to understand the test
failure from the DCHECK_EQ that I added.

[dcheng, 2014-07-02 21:53:22]

Update: I've also figured out why the test was failing. It looks like in the
past, when we generated AUTO_SUBFRAME navigation events, it might get targeted
at the main frame (see NavigatorImpl::DidNavigate, search for AUTO_SUBFRAME).
However, this doesn't appear to ever happen anymore. The test harness didn't
match the new model (it always dispatched the navigate events against frame
routing ID 1), so I've updated the test harness. I've also manually verified
that we don't hit the DCHECK_EQ during normal auto subframe loads browsing
around in Chrome.

[jam, 2014-07-02 22:24:51]

lgtm

https://codereview.chromium.org/365563002/diff/100001/content/browser/site_pe...
File content/browser/site_per_process_browsertest.cc (right):

https://codereview.chromium.org/365563002/diff/100001/content/browser/site_pe...
content/browser/site_per_process_browsertest.cc:37: content::RenderFrameHost*
render_frame_host,
nit: no "content::"

[dcheng, 2014-07-03 05:48:20]

TBRing jochen@ for the changes outside of content.

[dcheng, 2014-07-03 05:48:30]

The CQ bit was checked by [email protected]

[commit-bot: I haz the power, 2014-07-03 05:49:26]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/[email protected]/365563002/120001

[dcheng, 2014-07-03 09:13:38]

The CQ bit was checked by [email protected]

[commit-bot: I haz the power, 2014-07-03 09:14:11]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/[email protected]/365563002/140001

[commit-bot: I haz the power, 2014-07-03 12:17:54]

Change committed as 281241