[dom_distiller] Improve distillation speed.

Wait for DOMContentLoaded instead of the full page to load. Also, stop
page loading at this time. With this change, distillation results are
calculated sooner and network traffic is reduced as unneeded
subresources aren't fetched. The downside is that this may break some
pages which are using Javascript to write the main document although
this uncommon for document-centric pages.

Also adds some logs to see when the distiller is running:
(e.g. out/Debug/chrome --enable-dom-distiller --vmodule="*distiller*=1")

BUG=288015
NOTRY=true
R=cjhopman@chromium.org

Review URL: https://codereview.chromium.org/209333002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@259030 0039d316-1c4b-4281-b951-d872f2087c98

3 files changed, 18 insertions, 15 deletions

diff --git a/components/dom_distiller/content/distiller_page_web_contents.cc b/components/dom_distiller/content/distiller_page_web_contents.cc
index 2333d03..9cf8388 100644
--- a/components/dom_distiller/content/distiller_page_web_contents.cc
+++ b/components/dom_distiller/content/distiller_page_web_contents.cc
@@ -35,9 +35,9 @@ DistillerPageWebContents::~DistillerPageWebContents() {
 
 void DistillerPageWebContents::InitImpl() {
   DCHECK(browser_context_);
-  web_contents_.reset(
-      content::WebContents::Create(
-          content::WebContents::CreateParams(browser_context_)));
+  content::WebContents::CreateParams create_params(browser_context_);
+  create_params.initially_hidden = true;
+  web_contents_.reset(content::WebContents::Create(create_params));
 }
 
 void DistillerPageWebContents::LoadURLImpl(const GURL& gurl) {
@@ -57,14 +57,12 @@ void DistillerPageWebContents::ExecuteJavaScriptImpl(
                                       web_contents_->GetLastCommittedURL()));
 }
 
-void DistillerPageWebContents::DidFinishLoad(int64 frame_id,
-                                             const GURL& validated_url,
-                                             bool is_main_frame,
-                                             RenderViewHost* render_view_host) {
-  // TODO(shashishekhar): Find a better way to detect when it is safe to run the
-  // distillation script. Waiting for the entire page to load is really slow.
-  if (is_main_frame) {
+void DistillerPageWebContents::DocumentLoadedInFrame(
+    int64 frame_id,
+    RenderViewHost* render_view_host) {
+  if (frame_id == web_contents_->GetMainFrame()->GetRoutingID()) {
     content::WebContentsObserver::Observe(NULL);
+    web_contents_->Stop();
     OnLoadURLDone();
   }
 }
diff --git a/components/dom_distiller/content/distiller_page_web_contents.h b/components/dom_distiller/content/distiller_page_web_contents.h
index 8d8bf3e..c6b94fc 100644
--- a/components/dom_distiller/content/distiller_page_web_contents.h
+++ b/components/dom_distiller/content/distiller_page_web_contents.h
@@ -46,10 +46,8 @@ class DistillerPageWebContents : public DistillerPage,
   virtual ~DistillerPageWebContents();
 
   // content::WebContentsObserver implementation.
-  virtual void DidFinishLoad(int64 frame_id,
-                             const GURL& validated_url,
-                             bool is_main_frame,
-                             RenderViewHost* render_view_host) OVERRIDE;
+  virtual void DocumentLoadedInFrame(int64 frame_id,
+                                     RenderViewHost* render_view_host) OVERRIDE;
 
   virtual void DidFailLoad(int64 frame_id,
                            const GURL& validated_url,
diff --git a/components/dom_distiller/core/page_distiller.cc b/components/dom_distiller/core/page_distiller.cc
index 9116b47..db941ab 100644
--- a/components/dom_distiller/core/page_distiller.cc
+++ b/components/dom_distiller/core/page_distiller.cc
@@ -40,11 +40,15 @@ void PageDistiller::DistillPage(const GURL& url,
   LoadURL(url);
 }
 
-void PageDistiller::LoadURL(const GURL& url) { distiller_page_->LoadURL(url); }
+void PageDistiller::LoadURL(const GURL& url) {
+  DVLOG(1) << "Loading for distillation: " << url.spec();
+  distiller_page_->LoadURL(url);
+}
 
 void PageDistiller::OnLoadURLDone() { GetDistilledContent(); }
 
 void PageDistiller::GetDistilledContent() {
+  DVLOG(1) << "Beginning distillation";
   std::string script = ResourceBundle::GetSharedInstance()
                            .GetRawDataResource(IDR_DISTILLER_JS)
                            .as_string();
@@ -53,6 +57,9 @@ void PageDistiller::GetDistilledContent() {
 
 void PageDistiller::OnExecuteJavaScriptDone(const GURL& page_url,
                                             const base::Value* value) {
+  DVLOG(1) << "Distillation complete; extracting resources for "
+      << page_url.spec();
+
   scoped_ptr<DistilledPageInfo> page_info(new DistilledPageInfo());
   std::string result;
   const base::ListValue* result_list = NULL;