Importing certain bookmarks from firefox and HTML file as search engines.

chrome/utility/importer/bookmark_html_reader.cc

 // Copyright 2013 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "chrome/utility/importer/bookmark_html_reader.h"
 
 #include "base/callback.h"
 #include "base/files/file_util.h"
 #include "base/i18n/icu_string_conversions.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_split.h"
 #include "base/strings/string_util.h"
+#include "base/strings/utf_string_conversions.h"
 #include "base/time/time.h"
 #include "chrome/common/importer/imported_bookmark_entry.h"
 #include "chrome/common/importer/imported_favicon_usage.h"
 #include "chrome/utility/importer/favicon_reencode.h"
+#include "components/search_engines/search_terms_data.h"
+#include "components/search_engines/template_url.h"
 #include "net/base/data_url.h"
 #include "net/base/escape.h"
 #include "url/gurl.h"
 #include "url/url_constants.h"
 
 namespace {
 
 // Fetches the given |attribute| value from the |attribute_list|. Returns true
 // if successful, and |value| will contain the value.
 bool GetAttribute(const std::string& attribute_list,
@@ skipping 57 matching lines @@
 
 }  // namespace
 
 namespace bookmark_html_reader {
 
 void ImportBookmarksFile(
       const base::Callback<bool(void)>& cancellation_callback,
       const base::Callback<bool(const GURL&)>& valid_url_callback,
       const base::FilePath& file_path,
       std::vector<ImportedBookmarkEntry>* bookmarks,
+      std::vector<importer::SearchEngineInfo>* search_engines,
       std::vector<ImportedFaviconUsage>* favicons) {
   std::string content;
   base::ReadFileToString(file_path, &content);
   std::vector<std::string> lines;
   base::SplitString(content, '\n', &lines);
 
   base::string16 last_folder;
   bool last_folder_on_toolbar = false;
   bool last_folder_is_empty = true;
   bool has_subfolder = false;
@@ skipping 39 matching lines @@
     base::string16 post_data;
     bool is_bookmark;
     // TODO(jcampan): http://b/issue?id=1196285 we do not support POST based
     //                keywords yet.
     is_bookmark =
         internal::ParseBookmarkFromLine(line, charset, &title,
                                         &url, &favicon, &shortcut,
                                         &add_date, &post_data) ||
         internal::ParseMinimumBookmarkFromLine(line, charset, &title, &url);
 
+    std::string search_engine_url;

[Peter Kasting, 2014/11/17 21:30:36]

Nit: Declare this below the comment instead of above it

[Tapu Ghose, 2014/12/07 02:49:54]

Done.

+
+    // If bookmark contains a valid replaceable url and a keyword then import
+    // it as search engine.
+    if (is_bookmark && post_data.empty() &&
+        CanImportURLAsSearchEngine(url, &search_engine_url) &&
+            !shortcut.empty()) {
+      importer::SearchEngineInfo search_engine_info;
+      search_engine_info.url.assign(base::UTF8ToUTF16(search_engine_url));
+      search_engine_info.keyword = shortcut;
+      search_engine_info.display_name = title;
+      search_engines->push_back(search_engine_info);
+      continue;
+    }
+
     if (is_bookmark)
       last_folder_is_empty = false;
 
     if (is_bookmark &&
         post_data.empty() &&
         (valid_url_callback.is_null() || valid_url_callback.Run(url))) {
       if (toolbar_folder_index > path.size() && !path.empty()) {
         NOTREACHED();  // error in parsing.
         break;
       }
@@ skipping 68 matching lines @@
         // Parent folder include current one, so it's not empty.
         last_folder_is_empty = false;
       }
 
       if (toolbar_folder_index > path.size())
         toolbar_folder_index = 0;
     }
   }
 }
 
+bool CanImportURLAsSearchEngine(const GURL& url,
+                                std::string* search_engine_url) {
+  std::string url_spec = url.possibly_invalid_spec();
+
+  if (url_spec.empty())
+    return false;
+
+  url_spec = net::UnescapeURLComponent(url_spec,
+                                       net::UnescapeRule::URL_SPECIAL_CHARS);

[Peter Kasting, 2014/11/17 21:30:36]

Do you have a testcase that will fail without this unescaping?  If not, please
add one.

[Tapu Ghose, 2014/12/07 02:49:54]

As I mentioned in Patch Set 4, unescaping the url is necessary. For instance,
url "http://example.%s.com" without unescape rule appears as
"http://example.%25s.com". If I skip unescaping the url then
|SupportsReplacement| at line 277 always returns false, because it looks for
"%s" (not %25s). In addition, unescaping is needed to show proper url in the UI
("Manager Search Engines" from chrome://settings).

Hence, all test cases that are supposed to pass would fail, if unescaping the
url is omitted.

+
+  // Replace replacement terms ("%s") in |url_spec| with {searchTerms}.
+  url_spec =
+      TemplateURLRef::DisplayURLToURLRef(base::UTF8ToUTF16(url_spec));
+
+  TemplateURLData data;
+  data.SetURL(url_spec);
+  *search_engine_url = url_spec;
+  return TemplateURL(data).SupportsReplacement(SearchTermsData());
+}
+
 namespace internal {
 
 bool ParseCharsetFromLine(const std::string& line, std::string* charset) {
   const char kCharset[] = "charset=";
   if (StartsWithASCII(line, "<META", false) &&
       (line.find("CONTENT=\"") != std::string::npos ||
           line.find("content=\"") != std::string::npos)) {
     size_t begin = line.find(kCharset);
     if (begin == std::string::npos)
       return false;
@@ skipping 182 matching lines @@
       *url = GURL(value);
     }
   }
 
   return true;
 }
 
 }  // namespace internal
 
 }  // namespace bookmark_html_reader