Set the original url correctly if the frame is loaded via loadData

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 209b20bd42ea3ed1cf75a673d3d1c1ef659e9f85..17be18ed370d4e89d0347bf7099834bae53a2ab7 100644
--- a/content/renderer/render_frame_impl.cc
+++ b/content/renderer/render_frame_impl.cc
@@ -2106,7 +2106,6 @@ void RenderFrameImpl::UpdateURL(blink::WebFrame* frame) {
   DCHECK(ds);
 
   const WebURLRequest& request = ds->request();
-  const WebURLRequest& original_request = ds->originalRequest();
   const WebURLResponse& response = ds->response();
 
   DocumentState* document_state = DocumentState::FromDataSource(ds);
@@ -2227,10 +2226,7 @@ void RenderFrameImpl::UpdateURL(blink::WebFrame* frame) {
     // Track the URL of the original request.  We use the first entry of the
     // redirect chain if it exists because the chain may have started in another
     // process.
-    if (params.redirects.size() > 0)
-      params.original_request_url = params.redirects.at(0);
-    else
-      params.original_request_url = original_request.url();
+    params.original_request_url = GetOriginalRequestUrl(ds);
 
     params.history_list_was_cleared =
         navigation_state->history_list_was_cleared();
@@ -2281,4 +2277,22 @@ void RenderFrameImpl::didStopLoading() {
   Send(new FrameHostMsg_DidStopLoading(routing_id_));
 }
 
+GURL RenderFrameImpl::GetOriginalRequestUrl(WebDataSource* ds) {

[darin (slow to review), 2014/03/11 20:04:32]

nit: We aren't terribly consistent, but a quick grep reveals that we have more
method names including "URL" than "Url". I'd probably go with "URL".

[hush (inactive), 2014/03/12 21:34:21]

Done

+  std::vector<GURL> redirects;

[darin (slow to review), 2014/03/11 20:04:32]

nit: I'd move this below the hasUnreachableURL branch. That way you don't bother
getting the redirect chain unless it is needed.

[hush (inactive), 2014/03/12 21:34:21]

Done

+  GetRedirectChain(ds, &redirects);
+
+  // The frame is loaded via WebFrame::loadData, where we pass the

[darin (slow to review), 2014/03/11 20:04:32]

nit: This comment needs some work. It appears to be referring to some variable
names "history_url_for_data_url" that don't have much meaning in the context of
this function.

I think it is enough to indicate that the unreachableURL parameter of
WebFrame::loadData() should be treated as the original URL if set.

[hush (inactive), 2014/03/12 21:34:21]

Okay. I just reworded the comments by only referring to baseURL and
unreachableURL. These two are the input params to WebFrame::loadData.

+  // history_url_for_data_url as unreachableURL in blink. The
+  // base_url_for_data_url will be in the redirect chain, but we
+  // never actually visited base_url_for_data_url. So we need to set
+  // history_url_for_data_url as the original_request_url.
+  if (ds->hasUnreachableURL())
+    return ds->unreachableURL();
+
+  if (redirects.size() > 0)

[darin (slow to review), 2014/03/11 20:04:32]

nit: I'd use the .empty() method here instead: if (!redirects.empty())

[hush (inactive), 2014/03/12 21:34:21]

Done

+    return redirects.at(0);
+
+  return ds->originalRequest().url();
+}
+
 }  // namespace content