sync: Fetch after applying bundle and retry after errors

After a $GIT_URL/clone.bundle has been applied to the new local
repository, perform an incremental fetch using `git fetch` to ensure
the local repository is up-to-date. This allows the hosting server
to offer stale /clone.bundle files to bootstrap a new client.

If a single git fetch fails, it may succeed again after a short
delay.  Transient failures are typical in environments where the
remote Git server happens to have limits on how many requests it
can serve at once (the anonymous git daemon, or an HTTP server).
Wait a randomized delay between 30 and 45 seconds and retry the
failed project once.  This delay gives the site time to recover
from a transient traffic spike, and the randomization makes it less
likely that a spike occurs again from all of the same clients.

Change-Id: I97fb0fcb33630fb78ac1a21d1a4a3e2268ab60c0
Signed-off-by: Shawn O. Pearce <sop@google.com>

docs/manifest-format.txt

@@ -38,6 +38,7 @@ following DTD:
     <!ELEMENT default (EMPTY)>
     <!ATTLIST default remote   IDREF #IMPLIED>
     <!ATTLIST default revision CDATA #IMPLIED>
+    <!ATTLIST default sync-j   CDATA #IMPLIED>
 
     <!ELEMENT manifest-server (EMPTY)>
     <!ATTLIST url              CDATA #REQUIRED>

error.py

@@ -57,6 +57,15 @@ class UploadError(Exception):
   def __str__(self):
     return self.reason
 
+class DownloadError(Exception):
+  """Cannot download a repository.
+  """
+  def __init__(self, reason):
+    self.reason = reason
+
+  def __str__(self):
+    return self.reason
+
 class NoSuchProjectError(Exception):
   """A specified project does not exist in the work tree.
   """

git_command.py

@@ -72,6 +72,8 @@ def terminate_ssh_clients():
       pass
   _ssh_clients = []
 
+_git_version = None
+
 class _GitCall(object):
   def version(self):
     p = GitCommand(None, ['--version'], capture_stdout=True)
@@ -79,6 +81,21 @@ class _GitCall(object):
       return p.stdout
     return None
 
+  def version_tuple(self):
+    global _git_version
+
+    if _git_version is None:
+      ver_str = git.version()
+      if ver_str.startswith('git version '):
+        _git_version = tuple(
+          map(lambda x: int(x),
+            ver_str[len('git version '):].strip().split('.')[0:3]
+          ))
+      else:
+        print >>sys.stderr, 'fatal: "%s" unsupported' % ver_str
+        sys.exit(1)
+    return _git_version
+
   def __getattr__(self, name):
     name = name.replace('_','-')
     def fun(*cmdv):
@@ -88,23 +105,9 @@ class _GitCall(object):
     return fun
 git = _GitCall()
 
-_git_version = None
-
 def git_require(min_version, fail=False):
-  global _git_version
-
-  if _git_version is None:
-    ver_str = git.version()
-    if ver_str.startswith('git version '):
-      _git_version = tuple(
-        map(lambda x: int(x),
-          ver_str[len('git version '):].strip().split('.')[0:3]
-        ))
-    else:
-      print >>sys.stderr, 'fatal: "%s" unsupported' % ver_str
-      sys.exit(1)
-
-  if min_version <= _git_version:
+  git_version = git.version_tuple()
+  if min_version <= git_version:
     return True
   if fail:
     need = '.'.join(map(lambda x: str(x), min_version))

git_config.py

@@ -198,6 +198,15 @@ class GitConfig(object):
     except KeyError:
       return False
 
+  def UrlInsteadOf(self, url):
+    """Resolve any url.*.insteadof references.
+    """
+    for new_url in self.GetSubSections('url'):
+      old_url = self.GetString('url.%s.insteadof' % new_url)
+      if old_url is not None and url.startswith(old_url):
+        return new_url + url[len(old_url):]
+    return url
+
   @property
   def _sections(self):
     d = self._section_dict
@@ -482,6 +491,12 @@ def close_ssh():
 URI_SCP = re.compile(r'^([^@:]*@?[^:/]{1,}):')
 URI_ALL = re.compile(r'^([a-z][a-z+]*)://([^@/]*@?[^/]*)/')
 
+def GetSchemeFromUrl(url):
+  m = URI_ALL.match(url)
+  if m:
+    return m.group(1)
+  return None
+
 def _preconnect(url):
   m = URI_ALL.match(url)
   if m:

main.py

@@ -22,17 +22,22 @@ if __name__ == '__main__':
     del sys.argv[-1]
 del magic
 
+import netrc
 import optparse
 import os
 import re
 import sys
+import time
+import urllib2
 
 from trace import SetTrace
+from git_command import git, GitCommand
 from git_config import init_ssh, close_ssh
 from command import InteractiveCommand
 from command import MirrorSafeCommand
 from command import PagedCommand
 from editor import Editor
+from error import DownloadError
 from error import ManifestInvalidRevisionError
 from error import NoSuchProjectError
 from error import RepoChangedException
@@ -53,6 +58,9 @@ global_options.add_option('--no-pager',
 global_options.add_option('--trace',
                           dest='trace', action='store_true',
                           help='trace git command execution')
+global_options.add_option('--time',
+                          dest='time', action='store_true',
+                          help='time repo command execution')
 global_options.add_option('--version',
                           dest='show_version', action='store_true',
                           help='display this version of repo')
@@ -122,7 +130,23 @@ class _Repo(object):
         RunPager(config)
 
     try:
-      cmd.Execute(copts, cargs)
+      start = time.time()
+      try:
+        cmd.Execute(copts, cargs)
+      finally:
+        elapsed = time.time() - start
+        hours, remainder = divmod(elapsed, 3600)
+        minutes, seconds = divmod(remainder, 60)
+        if gopts.time:
+          if hours == 0:
+            print >>sys.stderr, 'real\t%dm%.3fs' \
+              % (minutes, seconds)
+          else:
+            print >>sys.stderr, 'real\t%dh%dm%.3fs' \
+              % (hours, minutes, seconds)
+    except DownloadError, e:
+      print >>sys.stderr, 'error: %s' % str(e)
+      sys.exit(1)
     except ManifestInvalidRevisionError, e:
       print >>sys.stderr, 'error: %s' % str(e)
       sys.exit(1)
@@ -133,6 +157,9 @@ class _Repo(object):
         print >>sys.stderr, 'error: no project in current directory'
       sys.exit(1)
 
+def _MyRepoPath():
+  return os.path.dirname(__file__)
+
 def _MyWrapperPath():
   return os.path.join(os.path.dirname(__file__), 'repo')
 
@@ -199,6 +226,77 @@ def _PruneOptions(argv, opt):
       continue
     i += 1
 
+_user_agent = None
+
+def _UserAgent():
+  global _user_agent
+
+  if _user_agent is None:
+    py_version = sys.version_info
+
+    os_name = sys.platform
+    if os_name == 'linux2':
+      os_name = 'Linux'
+    elif os_name == 'win32':
+      os_name = 'Win32'
+    elif os_name == 'cygwin':
+      os_name = 'Cygwin'
+    elif os_name == 'darwin':
+      os_name = 'Darwin'
+
+    p = GitCommand(
+      None, ['describe', 'HEAD'],
+      cwd = _MyRepoPath(),
+      capture_stdout = True)
+    if p.Wait() == 0:
+      repo_version = p.stdout
+      if len(repo_version) > 0 and repo_version[-1] == '\n':
+        repo_version = repo_version[0:-1]
+      if len(repo_version) > 0 and repo_version[0] == 'v':
+        repo_version = repo_version[1:]
+    else:
+      repo_version = 'unknown'
+
+    _user_agent = 'git-repo/%s (%s) git/%s Python/%d.%d.%d' % (
+      repo_version,
+      os_name,
+      '.'.join(map(lambda d: str(d), git.version_tuple())),
+      py_version[0], py_version[1], py_version[2])
+  return _user_agent
+
+class _UserAgentHandler(urllib2.BaseHandler):
+  def http_request(self, req):
+    req.add_header('User-Agent', _UserAgent())
+    return req
+
+  def https_request(self, req):
+    req.add_header('User-Agent', _UserAgent())
+    return req
+
+def init_http():
+  handlers = [_UserAgentHandler()]
+
+  mgr = urllib2.HTTPPasswordMgrWithDefaultRealm()
+  try:
+    n = netrc.netrc()
+    for host in n.hosts:
+      p = n.hosts[host]
+      mgr.add_password(None, 'http://%s/'  % host, p[0], p[2])
+      mgr.add_password(None, 'https://%s/' % host, p[0], p[2])
+  except netrc.NetrcParseError:
+    pass
+  except IOError:
+    pass
+  handlers.append(urllib2.HTTPBasicAuthHandler(mgr))
+
+  if 'http_proxy' in os.environ:
+    url = os.environ['http_proxy']
+    handlers.append(urllib2.ProxyHandler({'http': url, 'https': url}))
+  if 'REPO_CURL_VERBOSE' in os.environ:
+    handlers.append(urllib2.HTTPHandler(debuglevel=1))
+    handlers.append(urllib2.HTTPSHandler(debuglevel=1))
+  urllib2.install_opener(urllib2.build_opener(*handlers))
+
 def _Main(argv):
   opt = optparse.OptionParser(usage="repo wrapperinfo -- ...")
   opt.add_option("--repo-dir", dest="repodir",
@@ -217,6 +315,7 @@ def _Main(argv):
   try:
     try:
       init_ssh()
+      init_http()
       repo._Run(argv)
     finally:
       close_ssh()

manifest_xml.py

@@ -14,7 +14,9 @@
 # limitations under the License.
 
 import os
+import re
 import sys
+import urlparse
 import xml.dom.minidom
 
 from git_config import GitConfig, IsId
@@ -24,26 +26,36 @@ from error import ManifestParseError
 MANIFEST_FILE_NAME = 'manifest.xml'
 LOCAL_MANIFEST_NAME = 'local_manifest.xml'
 
+urlparse.uses_relative.extend(['ssh', 'git'])
+urlparse.uses_netloc.extend(['ssh', 'git'])
+
 class _Default(object):
   """Project defaults within the manifest."""
 
   revisionExpr = None
   remote = None
+  sync_j = 1
 
 class _XmlRemote(object):
   def __init__(self,
                name,
                fetch=None,
+               manifestUrl=None,
                review=None):
     self.name = name
     self.fetchUrl = fetch
+    self.manifestUrl = manifestUrl
     self.reviewUrl = review
 
   def ToRemoteSpec(self, projectName):
-    url = self.fetchUrl
-    while url.endswith('/'):
-      url = url[:-1]
-    url += '/%s.git' % projectName
+    url = self.fetchUrl.rstrip('/') + '/' + projectName + '.git'
+    manifestUrl = self.manifestUrl.rstrip('/')
+    # urljoin will get confused if there is no scheme in the base url
+    # ie, if manifestUrl is of the form <hostname:port>
+    if manifestUrl.find(':') != manifestUrl.find('/') - 1:
+        manifestUrl = 'gopher://' + manifestUrl
+    url = urlparse.urljoin(manifestUrl, url)
+    url = re.sub(r'^gopher://', '', url)
     return RemoteSpec(self.name, url, self.reviewUrl)
 
 class XmlManifest(object):
@@ -133,6 +145,9 @@ class XmlManifest(object):
     if d.revisionExpr:
       have_default = True
       e.setAttribute('revision', d.revisionExpr)
+    if d.sync_j > 1:
+      have_default = True
+      e.setAttribute('sync-j', '%d' % d.sync_j)
     if have_default:
       root.appendChild(e)
       root.appendChild(doc.createTextNode(''))
@@ -362,7 +377,8 @@ class XmlManifest(object):
 
     if name is None:
       s = m_url.rindex('/') + 1
-      remote = _XmlRemote('origin', m_url[:s])
+      manifestUrl = self.manifestProject.config.GetString('remote.origin.url')
+      remote = _XmlRemote('origin', m_url[:s], manifestUrl)
       name = m_url[s:]
 
     if name.endswith('.git'):
@@ -390,7 +406,8 @@ class XmlManifest(object):
     review = node.getAttribute('review')
     if review == '':
       review = None
-    return _XmlRemote(name, fetch, review)
+    manifestUrl = self.manifestProject.config.GetString('remote.origin.url')
+    return _XmlRemote(name, fetch, manifestUrl, review)
 
   def _ParseDefault(self, node):
     """
@@ -401,6 +418,11 @@ class XmlManifest(object):
     d.revisionExpr = node.getAttribute('revision')
     if d.revisionExpr == '':
       d.revisionExpr = None
+    sync_j = node.getAttribute('sync-j')
+    if sync_j == '' or sync_j is None:
+      d.sync_j = 1
+    else:
+      d.sync_j = int(sync_j)
     return d
 
   def _ParseNotice(self, node):

progress.py

@@ -21,13 +21,14 @@ from trace import IsTrace
 _NOT_TTY = not os.isatty(2)
 
 class Progress(object):
-  def __init__(self, title, total=0):
+  def __init__(self, title, total=0, units=''):
     self._title = title
     self._total = total
     self._done = 0
     self._lastp = -1
     self._start = time()
     self._show = False
+    self._units = units
 
   def update(self, inc=1):
     self._done += inc
@@ -51,11 +52,11 @@ class Progress(object):
 
       if self._lastp != p:
         self._lastp = p
-        sys.stderr.write('\r%s: %3d%% (%d/%d)  ' % (
+        sys.stderr.write('\r%s: %3d%% (%d%s/%d%s)  ' % (
           self._title,
           p,
-          self._done,
-          self._total))
+          self._done, self._units,
+          self._total, self._units))
         sys.stderr.flush()
 
   def end(self):
@@ -69,9 +70,9 @@ class Progress(object):
       sys.stderr.flush()
     else:
       p = (100 * self._done) / self._total
-      sys.stderr.write('\r%s: %3d%% (%d/%d), done.  \n' % (
+      sys.stderr.write('\r%s: %3d%% (%d%s/%d%s), done.  \n' % (
         self._title,
         p,
-        self._done,
-        self._total))
+        self._done, self._units,
+        self._total, self._units))
       sys.stderr.flush()

project.py

@@ -16,17 +16,21 @@ import traceback
 import errno
 import filecmp
 import os
+import random
 import re
 import shutil
 import stat
 import sys
+import time
 import urllib2
 
 from color import Coloring
 from git_command import GitCommand
-from git_config import GitConfig, IsId
+from git_config import GitConfig, IsId, GetSchemeFromUrl
+from error import DownloadError
 from error import GitError, HookError, ImportError, UploadError
 from error import ManifestInvalidRevisionError
+from progress import Progress
 
 from git_refs import GitRefs, HEAD, R_HEADS, R_TAGS, R_PUB, R_M
 
@@ -884,19 +888,33 @@ class Project(object):
 
 ## Sync ##
 
-  def Sync_NetworkHalf(self, quiet=False):
+  def Sync_NetworkHalf(self, quiet=False, is_new=None):
     """Perform only the network IO portion of the sync process.
        Local working directory/branch state is not affected.
     """
-    is_new = not self.Exists
+    if is_new is None:
+      is_new = not self.Exists
     if is_new:
-      if not quiet:
-        print >>sys.stderr
-        print >>sys.stderr, 'Initializing project %s ...' % self.name
       self._InitGitDir()
-
     self._InitRemote()
-    if not self._RemoteFetch(initial=is_new, quiet=quiet):
+
+    if is_new:
+      alt = os.path.join(self.gitdir, 'objects/info/alternates')
+      try:
+        fd = open(alt, 'rb')
+        try:
+          alt_dir = fd.readline().rstrip()
+        finally:
+          fd.close()
+      except IOError:
+        alt_dir = None
+    else:
+      alt_dir = None
+
+    if alt_dir is None and self._ApplyCloneBundle(initial=is_new, quiet=quiet):
+      is_new = False
+
+    if not self._RemoteFetch(initial=is_new, quiet=quiet, alt_dir=alt_dir):
       return False
 
     #Check that the requested ref was found after fetch
@@ -1307,29 +1325,19 @@ class Project(object):
 
   def _RemoteFetch(self, name=None, tag=None,
                    initial=False,
-                   quiet=False):
+                   quiet=False,
+                   alt_dir=None):
     if not name:
       name = self.remote.name
 
     ssh_proxy = False
-    if self.GetRemote(name).PreConnectFetch():
+    remote = self.GetRemote(name)
+    if remote.PreConnectFetch():
       ssh_proxy = True
 
     if initial:
-      alt = os.path.join(self.gitdir, 'objects/info/alternates')
-      try:
-        fd = open(alt, 'rb')
-        try:
-          ref_dir = fd.readline()
-          if ref_dir and ref_dir.endswith('\n'):
-            ref_dir = ref_dir[:-1]
-        finally:
-          fd.close()
-      except IOError, e:
-        ref_dir = None
-
-      if ref_dir and 'objects' == os.path.basename(ref_dir):
-        ref_dir = os.path.dirname(ref_dir)
+      if alt_dir and 'objects' == os.path.basename(alt_dir):
+        ref_dir = os.path.dirname(alt_dir)
         packed_refs = os.path.join(self.gitdir, 'packed-refs')
         remote = self.GetRemote(name)
 
@@ -1365,9 +1373,8 @@ class Project(object):
             old_packed += line
 
         _lwrite(packed_refs, tmp_packed)
-
       else:
-        ref_dir = None
+        alt_dir = None
 
     cmd = ['fetch']
 
@@ -1386,21 +1393,149 @@ class Project(object):
       cmd.append('tag')
       cmd.append(tag)
 
-    ok = GitCommand(self,
-                    cmd,
-                    bare = True,
-                    ssh_proxy = ssh_proxy).Wait() == 0
+    ok = False
+    for i in range(2):
+      if GitCommand(self, cmd, bare=True, ssh_proxy=ssh_proxy).Wait() == 0:
+        ok = True
+        break
+      time.sleep(random.randint(30, 45))
 
     if initial:
-      if ref_dir:
+      if alt_dir:
         if old_packed != '':
           _lwrite(packed_refs, old_packed)
         else:
           os.remove(packed_refs)
       self.bare_git.pack_refs('--all', '--prune')
-
     return ok
 
+  def _ApplyCloneBundle(self, initial=False, quiet=False):
+    if initial and self.manifest.manifestProject.config.GetString('repo.depth'):
+      return False
+
+    remote = self.GetRemote(self.remote.name)
+    bundle_url = remote.url + '/clone.bundle'
+    bundle_url = GitConfig.ForUser().UrlInsteadOf(bundle_url)
+    if GetSchemeFromUrl(bundle_url) not in ('http', 'https'):
+      return False
+
+    bundle_dst = os.path.join(self.gitdir, 'clone.bundle')
+    bundle_tmp = os.path.join(self.gitdir, 'clone.bundle.tmp')
+
+    exist_dst = os.path.exists(bundle_dst)
+    exist_tmp = os.path.exists(bundle_tmp)
+
+    if not initial and not exist_dst and not exist_tmp:
+      return False
+
+    if not exist_dst:
+      exist_dst = self._FetchBundle(bundle_url, bundle_tmp, bundle_dst, quiet)
+    if not exist_dst:
+      return False
+
+    cmd = ['fetch']
+    if quiet:
+      cmd.append('--quiet')
+    if not self.worktree:
+      cmd.append('--update-head-ok')
+    cmd.append(bundle_dst)
+    for f in remote.fetch:
+      cmd.append(str(f))
+    cmd.append('refs/tags/*:refs/tags/*')
+
+    ok = GitCommand(self, cmd, bare=True).Wait() == 0
+    if os.path.exists(bundle_dst):
+      os.remove(bundle_dst)
+    if os.path.exists(bundle_tmp):
+      os.remove(bundle_tmp)
+    return ok
+
+  def _FetchBundle(self, srcUrl, tmpPath, dstPath, quiet):
+    keep = True
+    done = False
+    dest = open(tmpPath, 'a+b')
+    try:
+      dest.seek(0, os.SEEK_END)
+      pos = dest.tell()
+
+      req = urllib2.Request(srcUrl)
+      if pos > 0:
+        req.add_header('Range', 'bytes=%d-' % pos)
+
+      try:
+        r = urllib2.urlopen(req)
+      except urllib2.HTTPError, e:
+        if e.code == 404:
+          keep = False
+          return False
+        elif e.info()['content-type'] == 'text/plain':
+          try:
+            msg = e.read()
+            if len(msg) > 0 and msg[-1] == '\n':
+              msg = msg[0:-1]
+            msg = ' (%s)' % msg
+          except:
+            msg = ''
+        else:
+          try:
+            from BaseHTTPServer import BaseHTTPRequestHandler
+            res = BaseHTTPRequestHandler.responses[e.code]
+            msg = ' (%s: %s)' % (res[0], res[1])
+          except:
+            msg = ''
+        raise DownloadError('HTTP %s%s' % (e.code, msg))
+      except urllib2.URLError, e:
+        raise DownloadError('%s (%s)' % (e.reason, req.get_host()))
+
+      p = None
+      try:
+        size = r.headers['content-length']
+        unit = 1 << 10
+
+        if size and not quiet:
+          if size > 1024 * 1.3:
+            unit = 1 << 20
+            desc = 'MB'
+          else:
+            desc = 'KB'
+          p = Progress(
+            'Downloading %s' % self.relpath,
+            int(size) / unit,
+            units=desc)
+          if pos > 0:
+            p.update(pos / unit)
+
+        s = 0
+        while True:
+          d = r.read(8192)
+          if d == '':
+            done = True
+            return True
+          dest.write(d)
+          if p:
+            s += len(d)
+            if s >= unit:
+              p.update(s / unit)
+              s = s % unit
+        if p:
+          if s >= unit:
+            p.update(s / unit)
+          else:
+            p.update(1)
+      finally:
+        r.close()
+        if p:
+          p.end()
+    finally:
+      dest.close()
+
+      if os.path.exists(dstPath):
+        os.remove(dstPath)
+      if done:
+        os.rename(tmpPath, dstPath)
+      elif not keep:
+        os.remove(tmpPath)
+
   def _Checkout(self, rev, quiet=False):
     cmd = ['checkout']
     if quiet:
@@ -1484,10 +1619,13 @@ class Project(object):
     for stock_hook in _ProjectHooks():
       name = os.path.basename(stock_hook)
 
-      if name in ('commit-msg',) and not self.remote.review:
+      if name in ('commit-msg',) and not self.remote.review \
+            and not self is self.manifest.manifestProject:
         # Don't install a Gerrit Code Review hook if this
         # project does not appear to use it for reviews.
         #
+        # Since the manifest project is one of those, but also
+        # managed through gerrit, it's excluded
         continue
 
       dst = os.path.join(hooks, name)

repo

@@ -2,7 +2,7 @@
 
 ## repo default configuration
 ##
-REPO_URL='git://android.git.kernel.org/tools/repo.git'
+REPO_URL='https://code.google.com/p/git-repo/'
 REPO_REV='stable'
 
 # Copyright (C) 2008 Google Inc.
@@ -28,7 +28,7 @@ if __name__ == '__main__':
 del magic
 
 # increment this whenever we make important changes to this script
-VERSION = (1, 10)
+VERSION = (1, 13)
 
 # increment this if the MAINTAINER_KEYS block is modified
 KEYRING_VERSION = (1,0)
@@ -91,6 +91,7 @@ import re
 import readline
 import subprocess
 import sys
+import urllib2
 
 home_dot_repo = os.path.expanduser('~/.repoconfig')
 gpg_dir = os.path.join(home_dot_repo, 'gnupg')
@@ -121,6 +122,10 @@ group.add_option('--mirror',
 group.add_option('--reference',
                  dest='reference',
                  help='location of mirror directory', metavar='DIR')
+group.add_option('--depth', type='int', default=None,
+                 dest='depth',
+                 help='create a shallow clone with given depth; see git clone')
+
 
 # Tool
 group = init_optparse.add_option_group('repo Version options')
@@ -183,10 +188,6 @@ def _Init(args):
     else:
       can_verify = True
 
-    if not opt.quiet:
-      print >>sys.stderr, 'Getting repo ...'
-      print >>sys.stderr, '   from %s' % url
-
     dst = os.path.abspath(os.path.join(repodir, S_repo))
     _Clone(url, dst, opt.quiet)
 
@@ -296,15 +297,42 @@ def _SetConfig(local, name, value):
     raise CloneFailure()
 
 
-def _Fetch(local, quiet, *args):
+def _InitHttp():
+  handlers = []
+
+  mgr = urllib2.HTTPPasswordMgrWithDefaultRealm()
+  try:
+    import netrc
+    n = netrc.netrc()
+    for host in n.hosts:
+      p = n.hosts[host]
+      mgr.add_password(None, 'http://%s/'  % host, p[0], p[2])
+      mgr.add_password(None, 'https://%s/' % host, p[0], p[2])
+  except:
+    pass
+  handlers.append(urllib2.HTTPBasicAuthHandler(mgr))
+
+  if 'http_proxy' in os.environ:
+    url = os.environ['http_proxy']
+    handlers.append(urllib2.ProxyHandler({'http': url, 'https': url}))
+  if 'REPO_CURL_VERBOSE' in os.environ:
+    handlers.append(urllib2.HTTPHandler(debuglevel=1))
+    handlers.append(urllib2.HTTPSHandler(debuglevel=1))
+  urllib2.install_opener(urllib2.build_opener(*handlers))
+
+def _Fetch(url, local, src, quiet):
+  if not quiet:
+    print >>sys.stderr, 'Get %s' % url
+
   cmd = [GIT, 'fetch']
   if quiet:
     cmd.append('--quiet')
     err = subprocess.PIPE
   else:
     err = None
-  cmd.extend(args)
-  cmd.append('origin')
+  cmd.append(src)
+  cmd.append('+refs/heads/*:refs/remotes/origin/*')
+  cmd.append('refs/tags/*:refs/tags/*')
 
   proc = subprocess.Popen(cmd, cwd = local, stderr = err)
   if err:
@@ -313,6 +341,62 @@ def _Fetch(local, quiet, *args):
   if proc.wait() != 0:
     raise CloneFailure()
 
+def _DownloadBundle(url, local, quiet):
+  if not url.endswith('/'):
+    url += '/'
+  url += 'clone.bundle'
+
+  proc = subprocess.Popen(
+    [GIT, 'config', '--get-regexp', 'url.*.insteadof'],
+    cwd = local,
+    stdout = subprocess.PIPE)
+  for line in proc.stdout:
+    m = re.compile(r'^url\.(.*)\.insteadof (.*)$').match(line)
+    if m:
+      new_url = m.group(1)
+      old_url = m.group(2)
+      if url.startswith(old_url):
+        url = new_url + url[len(old_url):]
+        break
+  proc.stdout.close()
+  proc.wait()
+
+  if not url.startswith('http:') and not url.startswith('https:'):
+    return False
+
+  dest = open(os.path.join(local, '.git', 'clone.bundle'), 'w+b')
+  try:
+    try:
+      r = urllib2.urlopen(url)
+    except urllib2.HTTPError, e:
+      if e.code == 404:
+        return False
+      print >>sys.stderr, 'fatal: Cannot get %s' % url
+      print >>sys.stderr, 'fatal: HTTP error %s' % e.code
+      raise CloneFailure()
+    except urllib2.URLError, e:
+      print >>sys.stderr, 'fatal: Cannot get %s' % url
+      print >>sys.stderr, 'fatal: error %s' % e.reason
+      raise CloneFailure()
+    try:
+      if not quiet:
+        print >>sys.stderr, 'Get %s' % url
+      while True:
+        buf = r.read(8192)
+        if buf == '':
+          return True
+        dest.write(buf)
+    finally:
+      r.close()
+  finally:
+    dest.close()
+
+def _ImportBundle(local):
+  path = os.path.join(local, '.git', 'clone.bundle')
+  try:
+    _Fetch(local, local, path, True)
+  finally:
+    os.remove(path)
 
 def _Clone(url, local, quiet):
   """Clones a git repository to a new subdirectory of repodir
@@ -340,11 +424,14 @@ def _Clone(url, local, quiet):
     print >>sys.stderr, 'fatal: could not create %s' % local
     raise CloneFailure()
 
+  _InitHttp()
   _SetConfig(local, 'remote.origin.url', url)
   _SetConfig(local, 'remote.origin.fetch',
                     '+refs/heads/*:refs/remotes/origin/*')
-  _Fetch(local, quiet)
-  _Fetch(local, quiet, '--tags')
+  if _DownloadBundle(url, local, quiet):
+    _ImportBundle(local)
+  else:
+    _Fetch(url, local, 'origin', quiet)
 
 
 def _Verify(cwd, branch, quiet):
@@ -596,4 +683,3 @@ def main(orig_args):
 
 if __name__ == '__main__':
   main(sys.argv[1:])
-

subcmds/init.py

@@ -21,6 +21,7 @@ from color import Coloring
 from command import InteractiveCommand, MirrorSafeCommand
 from error import ManifestParseError
 from project import SyncBuffer
+from git_config import GitConfig
 from git_command import git_require, MIN_GIT_VERSION
 
 class Init(InteractiveCommand, MirrorSafeCommand):
@@ -108,8 +109,8 @@ to update the working directory files.
         sys.exit(1)
 
       if not opt.quiet:
-        print >>sys.stderr, 'Getting manifest ...'
-        print >>sys.stderr, '   from %s' % opt.manifest_url
+        print >>sys.stderr, 'Get %s' \
+          % GitConfig.ForUser().UrlInsteadOf(opt.manifest_url)
       m._InitGitDir()
 
       if opt.manifest_branch:
@@ -138,7 +139,7 @@ to update the working directory files.
         print >>sys.stderr, 'fatal: --mirror not supported on existing client'
         sys.exit(1)
 
-    if not m.Sync_NetworkHalf():
+    if not m.Sync_NetworkHalf(is_new=is_new):
       r = m.GetRemote(m.remote.name)
       print >>sys.stderr, 'fatal: cannot obtain manifest %s' % r.url
 

subcmds/sync.py

@@ -28,6 +28,14 @@ try:
 except ImportError:
   import dummy_threading as _threading
 
+try:
+  import resource
+  def _rlimit_nofile():
+    return resource.getrlimit(resource.RLIMIT_NOFILE)
+except ImportError:
+  def _rlimit_nofile():
+    return (256, 256)
+
 from git_command import GIT
 from git_refs import R_HEADS
 from project import HEAD
@@ -72,7 +80,8 @@ revision is temporarily needed.
 
 The -s/--smart-sync option can be used to sync to a known good
 build as specified by the manifest-server element in the current
-manifest.
+manifest. The -t/--smart-tag option is similar and allows you to
+specify a custom tag/label.
 
 The -f/--force-broken option can be used to proceed with syncing
 other projects if a project sync fails.
@@ -108,6 +117,8 @@ later is required to fix a server side protocol bug.
 """
 
   def _Options(self, p, show_smart=True):
+    self.jobs = self.manifest.default.sync_j
+
     p.add_option('-f', '--force-broken',
                  dest='force_broken', action='store_true',
                  help="continue sync even if a project fails to sync")
@@ -125,11 +136,14 @@ later is required to fix a server side protocol bug.
                  help='be more quiet')
     p.add_option('-j','--jobs',
                  dest='jobs', action='store', type='int',
-                 help="number of projects to fetch simultaneously")
+                 help="projects to fetch simultaneously (default %d)" % self.jobs)
     if show_smart:
       p.add_option('-s', '--smart-sync',
                    dest='smart_sync', action='store_true',
                    help='smart sync using manifest from a known good build')
+      p.add_option('-t', '--smart-tag',
+                   dest='smart_tag', action='store',
+                   help='smart sync using manifest from a known tag')
 
     g = p.add_option_group('repo Version options')
     g.add_option('--no-repo-verify',
@@ -308,6 +322,10 @@ uncommitted changes are present' % project.relpath
   def Execute(self, opt, args):
     if opt.jobs:
       self.jobs = opt.jobs
+    if self.jobs > 1:
+      soft_limit, _ = _rlimit_nofile()
+      self.jobs = min(self.jobs, (soft_limit - 5) / 3)
+
     if opt.network_only and opt.detach_head:
       print >>sys.stderr, 'error: cannot combine -n and -d'
       sys.exit(1)
@@ -315,27 +333,31 @@ uncommitted changes are present' % project.relpath
       print >>sys.stderr, 'error: cannot combine -n and -l'
       sys.exit(1)
 
-    if opt.smart_sync:
+    if opt.smart_sync or opt.smart_tag:
       if not self.manifest.manifest_server:
         print >>sys.stderr, \
             'error: cannot smart sync: no manifest server defined in manifest'
         sys.exit(1)
       try:
         server = xmlrpclib.Server(self.manifest.manifest_server)
-        p = self.manifest.manifestProject
-        b = p.GetBranch(p.CurrentBranch)
-        branch = b.merge
-        if branch.startswith(R_HEADS):
-          branch = branch[len(R_HEADS):]
+        if opt.smart_sync:
+          p = self.manifest.manifestProject
+          b = p.GetBranch(p.CurrentBranch)
+          branch = b.merge
+          if branch.startswith(R_HEADS):
+            branch = branch[len(R_HEADS):]
 
-        env = os.environ.copy()
-        if (env.has_key('TARGET_PRODUCT') and
-            env.has_key('TARGET_BUILD_VARIANT')):
-          target = '%s-%s' % (env['TARGET_PRODUCT'],
-                              env['TARGET_BUILD_VARIANT'])
-          [success, manifest_str] = server.GetApprovedManifest(branch, target)
+          env = os.environ.copy()
+          if (env.has_key('TARGET_PRODUCT') and
+              env.has_key('TARGET_BUILD_VARIANT')):
+            target = '%s-%s' % (env['TARGET_PRODUCT'],
+                                env['TARGET_BUILD_VARIANT'])
+            [success, manifest_str] = server.GetApprovedManifest(branch, target)
+          else:
+            [success, manifest_str] = server.GetApprovedManifest(branch)
         else:
-          [success, manifest_str] = server.GetApprovedManifest(branch)
+          assert(opt.smart_tag)
+          [success, manifest_str] = server.GetManifest(opt.smart_tag)
 
         if success:
           manifest_name = "smart_sync_override.xml"
@@ -378,6 +400,8 @@ uncommitted changes are present' % project.relpath
       if not syncbuf.Finish():
         sys.exit(1)
       self.manifest._Unload()
+      if opt.jobs is None:
+        self.jobs = self.manifest.default.sync_j
     all = self.GetProjects(args, missing_ok=True)
 
     if not opt.local_only: