[Findit] Bug fixing and implemented some feature requests.

tools/findit/git_repository_parser.py

-# Copyright 2014 The Chromium Authors. All rights reserved.
+# Copyright (c) 2014 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.
 
 import base64
 import logging
-import os
 import xml.dom.minidom as minidom
 from xml.parsers.expat import ExpatError
 
 import crash_utils
 from repository_parser_interface import ParserInterface
 
 
 class GitParser(ParserInterface):
   """Parser for Git repository in googlesource.
 
@@ skipping 68 matching lines @@
 
       # Set url of this CL.
       revision_url_part = self.url_parts_map['revision_url'] % githash
       revision['url'] = base_url + revision_url_part
 
       # Go through changed files, they are in li.
       lis = ul.getElementsByTagName('li')
       for li in lis:
         # Retrieve path and action of the changed file
         file_path = li.getElementsByTagName('a')[0].firstChild.nodeValue
-        file_action = li.getElementsByTagName('span')[0].getAttribute('class')
+        file_change_type = li.getElementsByTagName('span')[
+            0].getAttribute('class')
 
         # Normalize file action so that it is same as SVN parser.
-        if file_action == 'add':
-          file_action = 'A'
-        elif file_action == 'delete':
-          file_action = 'D'
-        elif file_action == 'modify':
-          file_action = 'M'
+        if file_change_type == 'add':

[Martin Barbella, 2014/08/22 02:24:14]

Define a map to handle this and use it in both places where this logic is
needed.

[jeun, 2014/08/22 22:58:43]

Done.

+          file_change_type = 'A'
+        elif file_change_type == 'delete':
+          file_change_type = 'D'
+        elif file_change_type == 'modify':
+          file_change_type = 'M'
 
         # Add the changed file to the map.
-        changed_file = os.path.basename(file_path)
-        if changed_file not in file_to_revision_map:
-          file_to_revision_map[changed_file] = []
-        file_to_revision_map[changed_file].append((githash, file_action,
-                                                   file_path))
+        if file_path not in file_to_revision_map:
+          file_to_revision_map[file_path] = []
+        file_to_revision_map[file_path].append((githash, file_change_type))
 
       # Add this revision object to the map.
       revision_map[githash] = revision
 
     # Parse one revision for the start range, because googlesource does not
     # include the start of the range.
     self.ParseRevision(revision_url, range_start, revision_map,
                        file_to_revision_map)
 
     return (revision_map, file_to_revision_map)
@@ skipping 57 matching lines @@
     githash = json_revision['commit']
 
     # Set author, message and URL of this CL.
     revision['author'] = json_revision['author']['name']
     revision['message'] = json_revision['message']
     revision['url'] = url
 
     # Iterate through the changed files.
     for diff in json_revision['tree_diff']:
       file_path = diff['new_path']
-      file_action = diff['type']
+      file_change_type = diff['type']
 
       # Normalize file action so that it fits with svn_repository_parser.
-      if file_action == 'add':
-        file_action = 'A'
-      elif file_action == 'delete':
-        file_action = 'D'
-      elif file_action == 'modify':
-        file_action = 'M'
+      if file_change_type == 'add':
+        file_change_type = 'A'
+      elif file_change_type == 'delete':
+        file_change_type = 'D'
+      elif file_change_type == 'modify':
+        file_change_type = 'M'
 
       # Add the file to the map.
-      changed_file = os.path.basename(file_path)
-      if changed_file not in file_to_revision_map:
-        file_to_revision_map[changed_file] = []
-      file_to_revision_map[changed_file].append(
-          (githash, file_action, file_path))
+      if file_path not in file_to_revision_map:
+        file_to_revision_map[file_path] = []
+      file_to_revision_map[file_path].append((githash, file_change_type))
 
     # Add this CL to the map.
     revision_map[githash] = revision
 
     return
 
-  def ParseLineDiff(self, path, component, file_action, githash):
+  def ParseLineDiff(self, path, component, file_change_type, githash):
     changed_line_numbers = []
     changed_line_contents = []
     base_url = self.component_to_url_map[component]['repository']
     backup_url = (base_url + self.url_parts_map['revision_url']) % githash
 
     # If the file is added (not modified), treat it as if it is not changed.
-    if file_action == 'A':
+    if file_change_type == 'A':
       return (backup_url, changed_line_numbers, changed_line_contents)
 
     # Retrieves the diff data from URL, and if it fails, return emptry lines.
     url = (base_url + self.url_parts_map['diff_url']) % (githash, path)
     data = crash_utils.GetDataFromURL(url + '?format=text')
     if not data:
       logging.error('Failed to get diff from %s.', url)
       return (backup_url, changed_line_numbers, changed_line_contents)
 
     # Decode the returned object to line diff info
@@ skipping 53 matching lines @@
       # region.
       if start <= line and line <= start + count - 1:
         # If we are in the right region, get the information from the line.
         revision = blame_line['commit']
         author = blame_line['author']['name']
         revision_url_parts = self.url_parts_map['revision_url'] % revision
         revision_url = base_url + revision_url_parts
         # TODO(jeun): Add a way to get content from JSON object.
         content = None
 
-        return (content, revision, author, revision_url)
+        (revision_info, _) = self.ParseChangelog(component, revision, revision)
+        message = revision_info[revision]['message']
+        return (content, revision, author, revision_url, message)
 
     # Return none if the region does not exist.
     return None