Eliminate redundant work and control parallelism.

git_tools.py

 #!/usr/bin/python
 # Copyright (c) 2012 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 cStringIO
 import os
 import re
 import subprocess
 import sys
 import threading
 
+try:
+  import git_cache
+except ImportError:
+  for p in os.environ['PATH'].split(os.pathsep):
+    if (os.path.basename(p) == 'depot_tools' and
+        os.path.exists(os.path.join(p, 'git_cache.py'))):
+      sys.path.append(p)
+  import git_cache
+
 
 # Show more information about the commands being executed.
 VERBOSE = False
 
 # The longest any single subprocess will be allowed to run.
 TIMEOUT = 40 * 60
 
 class AbnormalExit(Exception):
   pass
 
@@ skipping 96 matching lines @@
   # For Abnormal Exit, Windows returns -1, Posix returns 128.
   if status in [-1, 128]:
     raise AbnormalExit('Failed to run %s. Exited Abnormally. output %s' %
                        (cmd, output))
   elif status != 0:
     raise Exception('Failed to run %s. error %d. output %s' % (cmd, status,
                                                                output))
   return (status, output)
 
 
-def GetCacheRepoDir(git_url, cache_dir):
-  """Assuming we used git_cache to populate a cache, get the repo directory."""
-  _, out = Git(None, 'cache exists --cache-dir=%s %s' % (cache_dir, git_url),
-               is_mirror=True)
-  return out.strip()
-
-
-def Clone(git_url, git_repo, is_mirror, out_queue=None, cache_dir=None,
-          shallow=False):
+def Clone(git_url, git_repo, is_mirror, out_buffer=None):
   """Clone a repository."""
-  repo_name = git_url.split('/')[-1]
-  if out_queue:
-    buf = StdioBuffer(repo_name, out_queue)
-  else:
-    buf = None
-
-  if is_mirror == 'bare':
-    if shallow:
-      if 'adobe' in git_url:
-        # --shallow by default checks out 10000 revision, but for really large
-        # repos like adobe ones, we want significantly less than 10000.
-        shallow_arg = '--depth 10 '
-      else:
-        shallow_arg = '--shallow '
-    else:
-      shallow_arg = ''
-    cmd = 'cache populate -v --cache-dir %s %s%s' % (cache_dir, shallow_arg,
-                                                     git_url)
-    return Git(None, cmd, is_mirror=True, out_buffer=buf)
-
   cmd = 'clone'
   if is_mirror:
     cmd += ' --mirror'
   cmd += ' %s %s'  % (git_url, git_repo)
 
   if not is_mirror and not os.path.exists(git_repo):
     os.makedirs(git_repo)
 
-  return Git(None, cmd, is_mirror, buf)
+  return Git(None, cmd, is_mirror=is_mirror, out_buffer=out_buffer)
+
+
+def PopulateCache(git_url, shallow=False):
+  # --shallow by default checks out 10000 revision, but for really large
+  # repos like adobe ones, we want significantly less than 10000.
+  depth = None
+  if shallow and 'adobe' in git_url:
+    depth = 10
+  mirror = git_cache.Mirror(git_url, print_func=lambda *args: None)
+  mirror.populate(depth=depth, shallow=shallow)
+  return mirror.mirror_path
 
 
 def Fetch(git_repo, git_url, is_mirror):
   """Fetch the latest objects for a given git repository."""
   # Always update the upstream url
   Git(git_repo, 'config remote.origin.url %s' % git_url)
   Git(git_repo, 'fetch origin', is_mirror)
 
 
 def Ping(git_repo, verbose=False):
@@ skipping 103 matching lines @@
   # Find the first commit matching the given git-svn-id regex.
   _, output = Git(
       git_repo,
       ('log -E --grep="^git-svn-id: [^@]*@%s [A-Za-z0-9-]*$" '
        '-1 --format="%%H" %s') % (regex, refspec),
       is_mirror)
   output = output.strip()
   if not re.match('^[0-9a-fA-F]{40}$', output):
     raise SearchError('Cannot find revision %s in %s:%s' % (svn_rev, git_repo,
                                                             refspec))
-
-  # Check if it actually matched the svn_rev that was requested.
-  found_rev = _FindRevForCommitish(git_repo, output, is_mirror)
-  found_msg = svn_rev
-  if found_rev != int(svn_rev):
-    found_msg = '%s [actual: %s]' % (svn_rev, found_rev)
-  print '%s: %s <-> %s' % (git_repo, output, found_msg)

[Michael Moss, 2014/05/15 16:26:40]

Why are you getting rid of this? Just to cut down on verbosity?

[szager1, 2014/05/19 21:40:07]

This is the only noise printed to stdout, other than the contents of .DEPS.git,
and it's not really helpful.

[Michael Moss, 2014/05/19 22:09:11]

I have actually found it helpful in the past (though not recently) when tracking
down some git conversion issues. Maybe just wrap with "if VERBOSE:"?

   return output
 
 
 def SearchExact(git_repo, svn_rev, is_mirror, refspec='FETCH_HEAD',
     fetch_url=None):
   """Return the Git commit id exactly matching the given SVN revision.
 
   If fetch_url is not None, will update repo if revision is newer."""
   regex = str(svn_rev)
   return _SearchImpl(git_repo, svn_rev, is_mirror, refspec, fetch_url, regex)
 
 
 def Search(git_repo, svn_rev, is_mirror, refspec='FETCH_HEAD', fetch_url=None):
   """Return the Git commit id fuzzy matching the given SVN revision.
 
   If fetch_url is not None, will update repo if revision is newer."""
   regex = CreateLessThanOrEqualRegex(svn_rev)
   return _SearchImpl(git_repo, svn_rev, is_mirror, refspec, fetch_url, regex)