Get signals working in the EXTRA_DATA section of MHTML

content/browser/download/mhtml_generation_manager.cc

Index: content/browser/download/mhtml_generation_manager.cc
diff --git a/content/browser/download/mhtml_generation_manager.cc b/content/browser/download/mhtml_generation_manager.cc
index b989af11932a1e2c49ec1451e7d55560874b8f28..343c3e5f7b26d676e15cf0f026a36ca3ec418ef5 100644
--- a/content/browser/download/mhtml_generation_manager.cc
+++ b/content/browser/download/mhtml_generation_manager.cc
@@ -32,6 +32,13 @@
 #include "content/public/common/mhtml_generation_params.h"
 #include "net/base/mime_util.h"
 
+namespace {
+const char kContentLocation[] = "Content-Location: ";
+const char kContentType[] = "Content-Type: ";
+const int kMHTMLExtraDataKey = 0;
+

[Dmitry Titov, 2017/03/18 00:58:34]

extra empty line

[Pete Williamson, 2017/03/20 18:26:53]

Done.

+}  // namespace
+
 namespace content {
 
 // The class and all of its members live on the UI thread.  Only static methods
@@ -101,9 +108,18 @@ class MHTMLGenerationManager::Job : public RenderProcessHostObserver {
   static std::tuple<MhtmlSaveStatus, int64_t> CloseFileOnFileThread(
       MhtmlSaveStatus save_status,
       const std::string& boundary,
-      base::File file);
+      base::File file,
+      const std::vector<MHTMLGenerationManager::MHTMLExtraSection>*
+          extra_sections);
   void AddFrame(RenderFrameHost* render_frame_host);
 
+  // If we have any extra data sections to write out, write them into the file
+  // while on the file thread.  Returns true for success.
+  static bool WriteExtraDataSectionsOnFileThread(
+      const std::string& boundary,
+      base::File& file,
+      const std::vector<MHTMLExtraSection>* extra_sections);
+
   // Creates a new map with values (content ids) the same as in
   // |frame_tree_node_to_content_id_| map, but with the keys translated from
   // frame_tree_node_id into a |site_instance|-specific routing_id.
@@ -156,6 +172,9 @@ class MHTMLGenerationManager::Job : public RenderProcessHostObserver {
   // destroyed by MHTMLGenerationManager::OnFileClosed).
   bool is_finished_;
 
+  // Any extra data sections that should be emitted into the output MHTML.
+  std::vector<MHTMLExtraSection>* extra_sections_;
+
   // RAII helper for registering this Job as a RenderProcessHost observer.
   ScopedObserver<RenderProcessHost, MHTMLGenerationManager::Job>
       observed_renderer_process_host_;
@@ -185,6 +204,9 @@ MHTMLGenerationManager::Job::Job(int job_id,
   DCHECK(!pending_frame_tree_node_ids_.empty());
   DCHECK(FrameTreeNode::GloballyFindByID(pending_frame_tree_node_ids_.front())
              ->parent() == nullptr);
+
+  // Save off any extra data.
+  extra_sections_ = MHTMLExtraData::FromWebContents(web_contents);
 }
 
 MHTMLGenerationManager::Job::~Job() {
@@ -355,7 +377,7 @@ void MHTMLGenerationManager::Job::CloseFile(
           &MHTMLGenerationManager::Job::CloseFileOnFileThread, save_status,
           (save_status == MhtmlSaveStatus::SUCCESS ? mhtml_boundary_marker_
                                                    : std::string()),
-          base::Passed(&browser_file_)),
+          base::Passed(&browser_file_), extra_sections_),
       callback);
 }
 
@@ -401,25 +423,34 @@ MhtmlSaveStatus MHTMLGenerationManager::Job::OnSerializeAsMHTMLResponse(
 
 // static
 std::tuple<MhtmlSaveStatus, int64_t>
-MHTMLGenerationManager::Job::CloseFileOnFileThread(MhtmlSaveStatus save_status,
-                                                   const std::string& boundary,
-                                                   base::File file) {
+MHTMLGenerationManager::Job::CloseFileOnFileThread(
+    MhtmlSaveStatus save_status,
+    const std::string& boundary,
+    base::File file,
+    const std::vector<MHTMLExtraSection>* extra_sections) {
   DCHECK_CURRENTLY_ON(BrowserThread::FILE);
+  int64_t file_size = -1;
 
   // If no previous error occurred the boundary should have been provided.
   if (save_status == MhtmlSaveStatus::SUCCESS) {
     TRACE_EVENT0("page-serialization",
                  "MHTMLGenerationManager::Job MHTML footer writing");
     DCHECK(!boundary.empty());
+
+    // Write the extra data into a section of its own, if we have any.
+    if (!WriteExtraDataSectionsOnFileThread(boundary, file, extra_sections))
+      save_status = MhtmlSaveStatus::FILE_WRITTING_ERROR;

[Dmitry Titov, 2017/03/18 00:58:34]

there was a comment about early return here and you replied "Done" but no change
in the file.

[Pete Williamson, 2017/03/20 18:26:53]

Ah, there are now two cases, and only one of them got fixed.  Both fixed now.

+
     std::string footer = base::StringPrintf("--%s--\r\n", boundary.c_str());
     DCHECK(base::IsStringASCII(footer));
-    if (file.WriteAtCurrentPos(footer.data(), footer.size()) < 0)
+    if (file.WriteAtCurrentPos(footer.data(), footer.size()) < 0) {
       save_status = MhtmlSaveStatus::FILE_WRITTING_ERROR;
+      return std::make_tuple(save_status, file_size);
+    }
   }
 
   // If the file is still valid try to close it. Only update the status if that
   // won't hide an earlier error.
-  int64_t file_size = -1;
   if (file.IsValid()) {
     file_size = file.GetLength();
     file.Close();
@@ -430,10 +461,70 @@ MHTMLGenerationManager::Job::CloseFileOnFileThread(MhtmlSaveStatus save_status,
   return std::make_tuple(save_status, file_size);
 }
 
+// static
+bool MHTMLGenerationManager::Job::WriteExtraDataSectionsOnFileThread(
+    const std::string& boundary,
+    base::File& file,
+    const std::vector<MHTMLExtraSection>* extra_sections) {
+  // Don't write an extra data section if there is none.
+  if (extra_sections->empty())
+    return true;
+
+  std::string extra_sections_serialized;
+
+  // For each extra section, serialize that section and add to our accumulator
+  // string.
+  for (auto section : *extra_sections) {
+    // Write a newline, then a boundary, another newline, then the content
+    // location, another newline, the content type, another newline, the extra
+    // data string, and end with a newline.
+    std::string extra_data_section = base::StringPrintf(
+        "--%s--\r\n%s%s\r\n%s%s\r\n%s\r\n", boundary.c_str(), kContentLocation,
+        section.content_location.c_str(), kContentType,
+        section.content_type.c_str(), section.body.c_str());
+    DCHECK(base::IsStringASCII(extra_data_section));
+
+    extra_sections_serialized += extra_data_section;
+  }
+
+  // Write the string into the file.  Returns false if we failed the write.
+  return (file.WriteAtCurrentPos(extra_sections_serialized.data(),
+                                 extra_sections_serialized.size()) >= 0);
+}
+
 MHTMLGenerationManager* MHTMLGenerationManager::GetInstance() {
   return base::Singleton<MHTMLGenerationManager>::get();
 }
 
+MHTMLGenerationManager::MHTMLExtraData::MHTMLExtraData() {}
+
+MHTMLGenerationManager::MHTMLExtraData::~MHTMLExtraData() {}
+
+std::vector<MHTMLGenerationManager::MHTMLExtraSection>*
+MHTMLGenerationManager::MHTMLExtraData::FromWebContents(
+    content::WebContents* contents) {
+  MHTMLExtraData* extra_data =
+      static_cast<MHTMLExtraData*>(contents->GetUserData(&kMHTMLExtraDataKey));
+  return extra_data->sections();
+}
+
+void MHTMLGenerationManager::MHTMLExtraData::AddToWebContents(
+    content::WebContents* contents,
+    MHTMLGenerationManager::MHTMLExtraSection& section) {
+  MHTMLExtraData* extra_data =
+      static_cast<MHTMLExtraData*>(contents->GetUserData(&kMHTMLExtraDataKey));
+  // If we don't already have any extra data, add it now.
+  if (extra_data == nullptr) {
+    extra_data = new MHTMLExtraData();
+    // Note that we hold onto extra_data until the end of this function, even
+    // though we transfer it to the web contents user data.
+    contents->SetUserData(&kMHTMLExtraDataKey,
+                          std::unique_ptr<MHTMLExtraData>(extra_data));
+  }
+  // Add this section to the list of sections to be saved out.
+  extra_data->sections()->push_back(section);
+}
+
 MHTMLGenerationManager::MHTMLGenerationManager() : next_job_id_(0) {}
 
 MHTMLGenerationManager::~MHTMLGenerationManager() {