Docserver: Only fetch content versions in the crons, not their contents.

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

Index: chrome/common/extensions/docs/server2/content_provider.py
diff --git a/chrome/common/extensions/docs/server2/content_provider.py b/chrome/common/extensions/docs/server2/content_provider.py
index e04dc7f4489a23ed207bbfb6dbf4049399f3ce88..f12f923937c90653ec6d145110f3170e11e5923d 100644
--- a/chrome/common/extensions/docs/server2/content_provider.py
+++ b/chrome/common/extensions/docs/server2/content_provider.py
@@ -197,16 +197,36 @@ class ContentProvider(object):
             .Then(lambda found: found or path))
 
   def Cron(self):
-    futures = [self._path_canonicalizer.Cron()]
-    for root, _, files in self.file_system.Walk(''):
-      for f in files:
-        futures.append(self.GetContentAndType(Join(root, f)))
-        # Also cache the extension-less version of the file if needed.
-        base, ext = posixpath.splitext(f)
-        if f != SITE_VERIFICATION_FILE and ext in self._default_extensions:
-          futures.append(self.GetContentAndType(Join(root, base)))
-      # TODO(kalman): Cache .zip files for each directory (if supported).
-    return All(futures, except_pass=Exception, except_pass_log=True)
+    def map_cron_paths(op):
+      results = []
+      for root, _, files in self.file_system.Walk(''):
+        for f in files:
+          results.append(op(Join(root, f)))
+          # Also cache the extension-less version of the file if needed.
+          base, ext = posixpath.splitext(f)
+          if f != SITE_VERIFICATION_FILE and ext in self._default_extensions:
+            results.append(op(Join(root, base)))
+        # TODO(kalman): Cache .zip files for each directory (if supported).
+      return results
+
+    # XXX(kalman): Need to do this stuff in APIModels as well - basically
+    # anywhere the _patch logic has been implemented. Err, sort of.
+    # Has it been implemented in here? Does it need to be?
+
+    # Immediately stat everything so that files are guaranteed to be eventually
+    # up to date. See http://crbug.com/398042 for background.
+    All(map_cron_paths(self.GetVersion)).Get()
+
+    # Update content in the future.
+    futures = [('<path_canonicalizer>',  # semi-arbitrary string since there is
+                                         # no path associated with this Future.
+                self._path_canonicalizer.Cron())]
+    futures += map_cron_paths(lambda path: (path, self.GetContentAndType(path)))
+    def resolve():
+      for label, future in futures:
+        try: future.Get()
+        except: logging.error('%s: %s' % (label, traceback.format_exc()))
+    return Future(callback=resolve)
 
   def __repr__(self):
     return 'ContentProvider of <%s>' % repr(self.file_system)