Band-aid fix for bug 487406

content/browser/frame_host/render_frame_proxy_host.cc

Index: content/browser/frame_host/render_frame_proxy_host.cc
diff --git a/content/browser/frame_host/render_frame_proxy_host.cc b/content/browser/frame_host/render_frame_proxy_host.cc
index fe62776995198a81592cc5447b7a85eb01f11c0c..066f17df56969601acddb34810d44b18faec1236 100644
--- a/content/browser/frame_host/render_frame_proxy_host.cc
+++ b/content/browser/frame_host/render_frame_proxy_host.cc
@@ -99,6 +99,7 @@ RenderViewHostImpl* RenderFrameProxyHost::GetRenderViewHost() {
 
 void RenderFrameProxyHost::TakeFrameHostOwnership(
     scoped_ptr<RenderFrameHostImpl> render_frame_host) {
+  CHECK(render_frame_host_ == nullptr);

[nasko, 2015/05/14 14:09:01]

CHECK_EQ?

[ncarter (slow), 2015/05/14 17:16:13]

Tried five things:

CHECK(render_frame_host_ == nullptr);  // Compiles
CHECK_EQ(render_frame_host_, nullptr);  // Doesn't compile
CHECK_EQ(render_frame_host_.get(), nullptr);  // Doesn't compile either
CHECK_EQ(render_frame_host_, static_cast<RenderFrameHost*>(nullptr));  // Still
doesn't compile.
CHECK_EQ(render_frame_host_.get(), static_cast<RenderFrameHost*>(nullptr));  //
Compiles! But at what cost?

There are some (but not many) examples of option 5 in the codebase:
https://code.google.com/p/chromium/codesearch#search/&q=CHECK_EQ.*nullptr&sq=...

I think the version with operator== is easiest on the eyes. CHECK_EQ(ptr1, ptr2)
is a good idea if you have two pointers variables that you expect to be equal,
because then stdout will show you which of the mismatched address values is null
or not. But if you're testing the value of a single pointer value against
nullity, I don't think it helps much to see the particular non-null address on
stdout.

   render_frame_host_ = render_frame_host.Pass();
   render_frame_host_->set_render_frame_proxy_host(this);
 }