[Findit] findit algorithms and a cluster-fuzz implementation.

tools/findit/chromium_deps.py

Index: tools/findit/chromium_deps.py
diff --git a/tools/findit/chromium_deps.py b/tools/findit/chromium_deps.py
index 0de6721d125428b1c712053735d8aa6305907328..37e16f534918934996f60d42831a05c7a6feecca 100644
--- a/tools/findit/chromium_deps.py
+++ b/tools/findit/chromium_deps.py
@@ -2,14 +2,18 @@
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
+import base64
 import https
 import utils
 
 
-DEPS_FILE_URL = 'https://src.chromium.org/chrome/trunk/src/DEPS?p=%s'
+DEPS_FILE_URL_SVN = 'https://src.chromium.org/chrome/trunk/src/DEPS?p=%s'
+DEPS_FILE_URL_GIT = (
+    'https://chromium.googlesource.com/chromium/src/+/%s/DEPS?format=text')
 
 
 class _VarImpl(object):
+
   def __init__(self, local_scope):
     self._local_scope = local_scope
 
@@ -70,7 +74,11 @@ def _GetComponentName(path):
 
 
 def _GetContentOfDEPS(chromium_revision):
-  return https.SendRequest(DEPS_FILE_URL % chromium_revision)
+  if utils.IsGitHash(chromium_revision):
+    url = DEPS_FILE_URL_GIT
+  else:
+    url = DEPS_FILE_URL_SVN
+  return https.SendRequest(url % chromium_revision)
 
 
 def GetChromiumComponents(chromium_revision,
@@ -85,13 +93,23 @@ def GetChromiumComponents(chromium_revision,
                           and returns the content of the DEPS file. The returned
                           content is assumed to be trusted input and will be
                           evaluated as python code.
+
+  Returns:
+    A map from component path to parsed component name, repository URL,
+    repository type and revision number.
   """
   if os_platform.lower() == 'linux':
     os_platform = 'unix'
 
+  is_git_hash = utils.IsGitHash(chromium_revision)
+
   # Download the content of DEPS file in chromium.
   deps_content = deps_file_downloader(chromium_revision)
 
+  # Googlesource git returns text file encoded in base64, so decode it.
+  if is_git_hash:
+    deps_content = base64.b64decode(deps_content)
+
   all_deps = {}
 
   # Parse the content of DEPS file.
@@ -102,7 +120,7 @@ def GetChromiumComponents(chromium_revision,
 
   # Figure out components based on the dependencies.
   components = {}
-  for component_path in all_deps.keys():
+  for component_path in all_deps:
     name = _GetComponentName(component_path)
     repository, revision = all_deps[component_path].split('@')
     is_git_hash = utils.IsGitHash(revision)
@@ -123,47 +141,51 @@ def GetChromiumComponents(chromium_revision,
         'revision': revision
     }
 
-  # Add chromium as a component.
-  # TODO(stgao): Move to git.
+  # Add chromium as a component, depending on the repository type.
+  if is_git_hash:
+    repository = 'https://chromium.googlesource.com/chromium/src/'
+    repository_type = 'git'
+  else:
+    repository = 'https://src.chromium.org/chrome/trunk'
+    repository_type = 'svn'
+
   components['src/'] = {
       'path': 'src/',
       'name': 'chromium',
-      'repository': 'https://src.chromium.org/chrome/trunk',
-      'repository_type': 'svn',
+      'repository': repository,
+      'repository_type': repository_type,
       'revision': chromium_revision
   }
 
   return components
 
 
-def GetChromiumComponentRange(chromium_revision1,
-                              chromium_revision2,
+def GetChromiumComponentRange(old_revision,
+                              new_revision,
                               os_platform='unix',
                               deps_file_downloader=_GetContentOfDEPS):
   """Return a list of components with their revision ranges.
 
   Args:
-    chromium_revision1: The revision of a Chrome build.
-    chromium_revision2: The revision of another Chrome build.
+    old_revision: The old revision of a Chrome build.
+    new_revision: The new revision of a Chrome build.
     os_platform: The target platform of the Chrome build, eg. win, mac, etc.
     deps_file_downloader: A function that takes the chromium_revision as input,
                           and returns the content of the DEPS file. The returned
                           content is assumed to be trusted input and will be
                           evaluated as python code.
-  """
-  # TODO(stgao): support git.
-  chromium_revision1 = int(chromium_revision1)
-  chromium_revision2 = int(chromium_revision2)
-  old_revision = str(min(chromium_revision1, chromium_revision2))
-  new_revision = str(max(chromium_revision1, chromium_revision2))
 
+  Returns:
+    A map from component path to its parsed regression and other information.
+  """
+  # Assume first revision is the old revision.
   old_components = GetChromiumComponents(old_revision, os_platform,
                                          deps_file_downloader)
   new_components = GetChromiumComponents(new_revision, os_platform,
                                          deps_file_downloader)
 
   components = {}
-  for path in new_components.keys():
+  for path in new_components:
     new_component = new_components[path]
     old_revision = None
     if path in old_components:
@@ -184,4 +206,5 @@ def GetChromiumComponentRange(chromium_revision1,
 
 if __name__ == '__main__':
   import json
-  print json.dumps(GetChromiumComponents(284750), sort_keys=True, indent=2)
+  print json.dumps(GetChromiumComponents(
+      'b4b1aea80b25a3b2f7952c9d95585e880421ef2b'), sort_keys=True, indent=2)