Read bookmarks from 'NETSCAPE-Bookmark-file's that don't have a beginning <DT> tag.

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 <stddef.h>
 #include <stdint.h>
 
 #include "base/callback.h"
@@ skipping 111 matching lines @@
     // separator in Firefox that Chrome does not support. Note that there can be
     // multiple "<HR>" tags at the beginning of a single line.
     // See http://crbug.com/257474.
     static const char kHrTag[] = "<HR>";
     while (base::StartsWith(line, kHrTag,
                             base::CompareCase::INSENSITIVE_ASCII)) {
       line.erase(0, arraysize(kHrTag) - 1);
       base::TrimString(line, " ", &line);
     }
 
+    // Remove "<DT>" if the line starts with "<DT>".  This may not occur if
+    // "<DT>" was on the previous line.  Liberally accept entries that do not
+    // have an opening "<DT>" at all.
+    static const char kDtTag[] = "<DT>";
+    if (base::StartsWith(line, kDtTag,
+                            base::CompareCase::INSENSITIVE_ASCII)) {

[Ilya Sherman, 2016/04/13 00:11:11]

nit: This indentation looks a bit off to me.  Could you please run clang-format
to double-check?

[Tom (Use chromium acct), 2016/04/14 01:43:56]

Done.

+      line.erase(0, arraysize(kDtTag) - 1);
+      base::TrimString(line, " ", &line);
+    }
+
     // Get the encoding of the bookmark file.
     if (internal::ParseCharsetFromLine(line, &charset))
       continue;
 
     // Get the folder name.
     if (internal::ParseFolderNameFromLine(line,
                                           charset,
                                           &last_folder,
                                           &last_folder_on_toolbar,
                                           &last_folder_add_date)) {
@@ skipping 154 matching lines @@
     return true;
   }
   return false;
 }
 
 bool ParseFolderNameFromLine(const std::string& line,
                              const std::string& charset,
                              base::string16* folder_name,
                              bool* is_toolbar_folder,
                              base::Time* add_date) {
-  const char kFolderOpen[] = "<DT><H3";
+  const char kFolderOpen[] = "<H3";
   const char kFolderClose[] = "</H3>";
   const char kToolbarFolderAttribute[] = "PERSONAL_TOOLBAR_FOLDER";
   const char kAddDateAttribute[] = "ADD_DATE";
 
   if (!base::StartsWith(line, kFolderOpen, base::CompareCase::SENSITIVE))
     return false;
 
   size_t end = line.find(kFolderClose);
   size_t tag_end = line.rfind('>', end) + 1;
   // If no end tag or start tag is broken, we skip to find the folder name.
@@ skipping 27 matching lines @@
 }
 
 bool ParseBookmarkFromLine(const std::string& line,
                            const std::string& charset,
                            base::string16* title,
                            GURL* url,
                            GURL* favicon,
                            base::string16* shortcut,
                            base::Time* add_date,
                            base::string16* post_data) {
-  const char kItemOpen[] = "<DT><A";
+  const char kItemOpen[] = "<A";
   const char kItemClose[] = "</A>";
   const char kFeedURLAttribute[] = "FEEDURL";
   const char kHrefAttribute[] = "HREF";
   const char kIconAttribute[] = "ICON";
   const char kShortcutURLAttribute[] = "SHORTCUTURL";
   const char kAddDateAttribute[] = "ADD_DATE";
   const char kPostDataAttribute[] = "POST_DATA";
 
   title->clear();
   *url = GURL();
@@ skipping 62 matching lines @@
     *post_data = net::UnescapeForHTML(*post_data);
   }
 
   return true;
 }
 
 bool ParseMinimumBookmarkFromLine(const std::string& line,
                                   const std::string& charset,
                                   base::string16* title,
                                   GURL* url) {
-  const char kItemOpen[] = "<DT><A";
+  const char kItemOpen[] = "<A";
   const char kItemClose[] = "</";
   const char kHrefAttributeUpper[] = "HREF";
   const char kHrefAttributeLower[] = "href";
 
   title->clear();
   *url = GURL();
 
   // Case-insensitive check of open tag.
   if (!base::StartsWith(line, kItemOpen, base::CompareCase::INSENSITIVE_ASCII))
     return false;
@@ skipping 27 matching lines @@
       *url = GURL(value);
     }
   }
 
   return true;
 }
 
 }  // namespace internal
 
 }  // namespace bookmark_html_reader