Simplify RenderFrameImpl::decidePolicyForNavigation and avoid DCHECK

content/renderer/render_frame_impl.cc

Index: content/renderer/render_frame_impl.cc
diff --git a/content/renderer/render_frame_impl.cc b/content/renderer/render_frame_impl.cc
index 0fe03c647e108e470c63b818949ba1f795a02eed..9ca4ebab8c220849b74f3755170559ea6334a3bd 100644
--- a/content/renderer/render_frame_impl.cc
+++ b/content/renderer/render_frame_impl.cc
@@ -4984,7 +4984,6 @@ WebNavigationPolicy RenderFrameImpl::decidePolicyForNavigation(
       GURL source_url(old_url);
       if (is_initial_navigation && source_url.is_empty() && frame_->opener())
         source_url = frame_->opener()->top()->document().url();

[Charlie Reis, 2016/07/14 21:16:47]

Do we gain much by having the opener check at all?  I ask because it's broken
for OOPIFs, in which case top()->document() may be null.  (We're going to have
to come back to fix this in https://crbug.com/466297.)

If it's safe to fork for file-to-file navigations when the opener isn't found,
it's probably safe to do it even when there is an opener.  And that's good,
because I think some tab (or frame) other than the opener might be able to cause
this navigation as well.

Maybe we can change the comment to something like:
// Fork non-file to file opens.  Note that this may fork unnecessarily if
another tab (hosting a file or not) targeted this one before its initial
navigation, but that shouldn't cause a problem.

[Tom Sepez, 2016/07/14 21:27:59]

Done.

-      DCHECK(!source_url.is_empty());
       should_fork = !source_url.SchemeIs(url::kFileScheme);
     }