Call dispatchDidReceiveServerRedirectForProvisionalLoad before checking navigation policy for a red…

Call dispatchDidReceiveServerRedirectForProvisionalLoad before checking navigation policy for a redirect.

Repro steps:
1. Load an application that uses the Chromium WebKit API.
2. Redirect a request via net::URLRequestJob::IsRedirectResponse by populating
|location| and returning true.

The |isRedirect| value passed to WebFrameClient::decidePolicyForNavigation in
FrameLoaderClientImpl::dispatchDecidePolicyForNavigationAction should be true. However, WebKit revision 137607 (http://trac.webkit.org/changeset/137607) changed the implementation of MainResourceLoader::load() such that WebDataSourceImpl::appendRedirect() is no longer called for the redirected URL following this code path. As a result the |isRedirect| value is now false.

BUG=229154
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=150690

[Nate Chapin, 2013-05-17 17:41:53]

I'm not sure how to properly test this, besides seeing the callbacks in the
desired order in some layout tests.

[abarth-chromium, 2013-05-17 18:58:43]

What's the reason for making this change?  Does it fix a bug in CEF?  (I'm
guessing because you've CCed Marshall.)

If you'd be willing to improve the description, then LGTM.

[Nate Chapin, 2013-05-17 19:02:58]

On 2013/05/17 18:58:43, abarth wrote:
> What's the reason for making this change?  Does it fix a bug in CEF?  (I'm
> guessing because you've CCed Marshall.)
> 
> If you'd be willing to improve the description, then LGTM.

The bug mentioned in the description described that we're not properly reporting
the redirect chain for server redirects. I don't fully understand what that
breaks in practice.

[abarth-chromium, 2013-05-17 19:07:52]

Would you be willing to copy/paste that information into the description?  It's
too much work to click through bug links to sort out why we're making changes. 
/me is lazy.  :)

[Nate Chapin, 2013-05-17 19:15:56]

On 2013/05/17 19:07:52, abarth wrote:
> Would you be willing to copy/paste that information into the description? 
It's
> too much work to click through bug links to sort out why we're making changes.

> /me is lazy.  :)


Chromium revision 176706
WebKit revision 139548

1. Load an application that uses the Chromium WebKit API.
2. Redirect a request via net::URLRequestJob::IsRedirectResponse by populating
|location| and returning true.

The |isRedirect| value passed to WebFrameClient::decidePolicyForNavigation in
FrameLoaderClientImpl::dispatchDecidePolicyForNavigationAction should be true.
However, WebKit revision 137607 (http://trac.webkit.org/changeset/137607)
changed the implementation of MainResourceLoader::load() such that
WebDataSourceImpl::appendRedirect() is no longer called for the redirected URL
following this code path. As a result the |isRedirect| value is now false.

Former WebKit bug: https://bugs.webkit.org/show_bug.cgi?id=107175

[Marshall, 2013-05-20 19:49:16]

LGTM. I've confirmed that it fixes the CEF problem reported in the bug.

[commit-bot: I haz the power, 2013-05-20 19:51:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/japhet@chromium.org/15303010/1

[commit-bot: I haz the power, 2013-05-20 20:59:04]

Change committed as 150690