OOPIFs: Deduplicating MHTML parts across frames.

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 beeed94535d7f1265d40a753b43c2066b76af33a..6ea1f8e15350d848cebc7a2d38802f982edb3d0d 100644
--- a/content/renderer/render_frame_impl.cc
+++ b/content/renderer/render_frame_impl.cc
@@ -124,6 +124,7 @@
 #include "content/renderer/web_frame_utils.h"
 #include "content/renderer/web_ui_extension.h"
 #include "content/renderer/websharedworker_proxy.h"
+#include "crypto/sha2.h"
 #include "gin/modules/module_registry.h"
 #include "media/audio/audio_output_device.h"
 #include "media/base/audio_renderer_mixer_input.h"
@@ -168,6 +169,7 @@
 #include "third_party/WebKit/public/web/WebView.h"
 #include "third_party/mojo/src/mojo/edk/js/core.h"
 #include "third_party/mojo/src/mojo/edk/js/support.h"
+#include "url/gurl.h"
 #include "url/url_util.h"
 
 #if defined(ENABLE_PLUGINS)
@@ -4708,16 +4710,12 @@ void RenderFrameImpl::OnGetSerializedHtmlWithLocalLinks(
 }
 
 void RenderFrameImpl::OnSerializeAsMHTML(
-    int job_id,
-    IPC::PlatformFileForTransit file_for_transit,
-    const std::string& std_mhtml_boundary,
-    const std::map<int, std::string>& frame_routing_id_to_content_id,
-    bool is_last_frame) {
+    const FrameMsg_SerializeAsMHTML_Params& params) {
   // Unpack IPC payload.
-  base::File file = IPC::PlatformFileForTransitToFile(file_for_transit);
-  WebString mhtml_boundary = WebString::fromUTF8(std_mhtml_boundary);
+  base::File file = IPC::PlatformFileForTransitToFile(params.destination_file);
+  WebString mhtml_boundary = WebString::fromUTF8(params.mhtml_boundary_marker);
   std::vector<std::pair<WebFrame*, WebString>> web_frame_to_content_id;
-  for (const auto& it : frame_routing_id_to_content_id) {
+  for (const auto& it : params.frame_routing_id_to_content_id) {
     const std::string& content_id = it.second;
     WebFrame* web_frame = GetWebFrameFromRoutingIdForFrameOrProxy(it.first);
     if (!web_frame)
@@ -4726,9 +4724,11 @@ void RenderFrameImpl::OnSerializeAsMHTML(
     web_frame_to_content_id.push_back(
         std::make_pair(web_frame, WebString::fromUTF8(content_id)));
   }
+  const std::string& salt = params.salt;
 
   WebData data;
   bool success = true;
+  std::set<GURL> uris_of_generated_mhtml_parts;
 
   // Generate MHTML header if needed.
   if (success && IsMainFrame()) {
@@ -4745,8 +4745,20 @@ void RenderFrameImpl::OnSerializeAsMHTML(
 
   // Generate MHTML parts.
   if (success) {
-    data = WebPageSerializer::generateMHTMLParts(
-        mhtml_boundary, GetWebFrame(), false, web_frame_to_content_id);
+    std::vector<WebURL> uris_to_skip;
+    WebVector<WebURL> all_urls =
+        WebPageSerializer::enumerateMHTMLResources(GetWebFrame());
+    for (const WebURL url : all_urls) {
+      std::string digest = crypto::SHA256HashString(salt + GURL(url).spec());
+      if (0 != params.digests_of_uris_to_skip.count(digest))
+        uris_to_skip.push_back(url);
+      else
+        uris_of_generated_mhtml_parts.insert(url);
+    }
+
+    data = WebPageSerializer::generateMHTMLParts(mhtml_boundary, GetWebFrame(),
+                                                 false, web_frame_to_content_id,
+                                                 uris_to_skip);
     // TODO(jcivelli): write the chunks in deferred tasks to give a chance to
     //                 the message loop to process other events.
     if (file.WriteAtCurrentPos(data.data(), data.size()) < 0) {
@@ -4755,7 +4767,7 @@ void RenderFrameImpl::OnSerializeAsMHTML(
   }
 
   // Generate MHTML footer if needed.
-  if (success && is_last_frame) {
+  if (success && params.is_last_frame) {
     data = WebPageSerializer::generateMHTMLFooter(mhtml_boundary);
     if (file.WriteAtCurrentPos(data.data(), data.size()) < 0) {
       success = false;
@@ -4764,8 +4776,9 @@ void RenderFrameImpl::OnSerializeAsMHTML(
 
   // Cleanup and notify the browser process about completion.
   file.Close();  // Need to flush file contents before sending IPC response.
-  Send(new FrameHostMsg_SerializeAsMHTMLResponse(routing_id_, job_id, success,
-                                                 mhtml_boundary.utf8()));
+  Send(new FrameHostMsg_SerializeAsMHTMLResponse(
+      routing_id_, params.job_id, success, mhtml_boundary.utf8(),
+      uris_of_generated_mhtml_parts));
 }
 
 void RenderFrameImpl::OpenURL(const GURL& url,