Docserver: Persist stat cache for versioned file systems

chrome/common/extensions/docs/server2/refresh_servlet.py

 # Copyright 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 traceback
 
 from app_yaml_helper import AppYamlHelper
-from appengine_wrappers import IsDeadlineExceededError, logservice
+from appengine_wrappers import IsDeadlineExceededError, logservice, taskqueue
 from branch_utility import BranchUtility
+from commit_tracker import CommitTracker
 from compiled_file_system import CompiledFileSystem
 from custom_logger import CustomLogger
 from data_source_registry import CreateDataSource
 from environment import GetAppVersion
 from file_system import IsFileSystemThrottledError
 from future import Future
 from gcs_file_system_provider import CloudStorageFileSystemProvider
 from github_file_system_provider import GithubFileSystemProvider
 from host_file_system_provider import HostFileSystemProvider
 from object_store_creator import ObjectStoreCreator
+from refresh_tracker import RefreshTracker
 from server_instance import ServerInstance
 from servlet import Servlet, Request, Response
 from timer import Timer, TimerClosure
 
 
-
 _log = CustomLogger('refresh')
 
 
 class RefreshServlet(Servlet):
   '''Servlet which refreshes a single data source.
   '''
   def __init__(self, request, delegate_for_test=None):
     Servlet.__init__(self, request)
     self._delegate = delegate_for_test or RefreshServlet.Delegate()
 
@@ skipping 43 matching lines @@
         (source_name, '' if source_path is None else '[%s]' % source_path))
 
     if 'commit' in self._request.arguments:
       commit = self._request.arguments['commit']
     else:
       _log.warning('No commit given; refreshing from master. '
                    'This is probably NOT what you want.')
       commit = None
 
     server_instance = self._CreateServerInstance(commit)
+    commit_tracker = CommitTracker(server_instance.object_store_creator)
+    refresh_tracker = RefreshTracker(server_instance.object_store_creator)
+
     success = True
     try:
       if source_name == 'platform_bundle':
         data_source = server_instance.platform_bundle
       elif source_name == 'content_providers':
         data_source = server_instance.content_providers
       else:
         data_source = CreateDataSource(source_name, server_instance)
 
       class_name = data_source.__class__.__name__
       refresh_future = data_source.Refresh(source_path)
       assert isinstance(refresh_future, Future), (
           '%s.Refresh() did not return a Future' % class_name)
       timer = Timer()
       try:
         refresh_future.Get()
+
+        # Mark this (commit, task) pair as completed and then see if this
+        # concludes the full cache refresh. The list of tasks required to
+        # complete a cache refresh is registered (and keyed on commit ID) by the
+        # CronServlet before kicking off all the refresh tasks.
+        (refresh_tracker.MarkTaskComplete(commit, path)
+          .Then(lambda _: refresh_tracker.GetRefreshComplete(commit))
+          .Then(lambda is_complete:
+                  commit_tracker.Set('master', commit) if is_complete else None)
+          .Get())
       except Exception as e:
         _log.error('%s: error %s' % (class_name, traceback.format_exc()))
         success = False
         if IsFileSystemThrottledError(e):
           return Response.ThrottledError('Throttled')
         raise
       finally:
         _log.info('Refreshing %s took %s' %
             (class_name, timer.Stop().FormatElapsed()))
 
@@ skipping 19 matching lines @@
     github_file_system_provider = self._delegate.CreateGithubFileSystemProvider(
         object_store_creator)
     gcs_file_system_provider = self._delegate.CreateGCSFileSystemProvider(
         object_store_creator)
     return ServerInstance(object_store_creator,
                           CompiledFileSystem.Factory(object_store_creator),
                           branch_utility,
                           host_file_system_provider,
                           github_file_system_provider,
                           gcs_file_system_provider)