Use JSON instead of pickle

Use JSON as it is shown to be much faster than pickle.
Also clean up the loading and saving functions.

Change-Id: I45b3dee7b4d59a1c0e0d38d4a83b543ac5839390

git_config.py

@@ -15,8 +15,8 @@
 
 from __future__ import print_function
 
+import json
 import os
-import pickle
 import re
 import subprocess
 import sys
@@ -80,7 +80,7 @@ class GitConfig(object):
     return cls(configfile = os.path.join(gitdir, 'config'),
                defaults = defaults)
 
-  def __init__(self, configfile, defaults=None, pickleFile=None):
+  def __init__(self, configfile, defaults=None, jsonFile=None):
     self.file = configfile
     self.defaults = defaults
     self._cache_dict = None
@@ -88,12 +88,11 @@ class GitConfig(object):
     self._remotes = {}
     self._branches = {}
 
-    if pickleFile is None:
-      self._pickle = os.path.join(
+    self._json = jsonFile
+    if self._json is None:
+      self._json = os.path.join(
         os.path.dirname(self.file),
-        '.repopickle_' + os.path.basename(self.file))
-    else:
-      self._pickle = pickleFile
+        '.repo_' + os.path.basename(self.file) + '.json')
 
   def Has(self, name, include_defaults = True):
     """Return true if this configuration file has the key.
@@ -248,50 +247,41 @@ class GitConfig(object):
     return self._cache_dict
 
   def _Read(self):
-    d = self._ReadPickle()
+    d = self._ReadJson()
     if d is None:
       d = self._ReadGit()
-      self._SavePickle(d)
+      self._SaveJson(d)
     return d
 
-  def _ReadPickle(self):
+  def _ReadJson(self):
     try:
-      if os.path.getmtime(self._pickle) \
+      if os.path.getmtime(self._json) \
       <= os.path.getmtime(self.file):
-        os.remove(self._pickle)
+        os.remove(self._json)
         return None
     except OSError:
       return None
     try:
-      Trace(': unpickle %s', self.file)
-      fd = open(self._pickle, 'rb')
+      Trace(': parsing %s', self.file)
+      fd = open(self._json)
       try:
-        return pickle.load(fd)
+        return json.load(fd)
       finally:
         fd.close()
-    except EOFError:
-      os.remove(self._pickle)
-      return None
-    except IOError:
-      os.remove(self._pickle)
-      return None
-    except pickle.PickleError:
-      os.remove(self._pickle)
+    except (IOError, ValueError):
+      os.remove(self._json)
       return None
 
-  def _SavePickle(self, cache):
+  def _SaveJson(self, cache):
     try:
-      fd = open(self._pickle, 'wb')
+      fd = open(self._json, 'w')
       try:
-        pickle.dump(cache, fd, pickle.HIGHEST_PROTOCOL)
+        json.dump(cache, fd, indent=2)
       finally:
         fd.close()
-    except IOError:
-      if os.path.exists(self._pickle):
-        os.remove(self._pickle)
-    except pickle.PickleError:
-      if os.path.exists(self._pickle):
-        os.remove(self._pickle)
+    except (IOError, TypeError):
+      if os.path.exists(self.json):
+        os.remove(self._json)
 
   def _ReadGit(self):
     """

subcmds/sync.py

@@ -14,10 +14,10 @@
 # limitations under the License.
 
 from __future__ import print_function
+import json
 import netrc
 from optparse import SUPPRESS_HELP
 import os
-import pickle
 import re
 import shutil
 import socket
@@ -760,7 +760,7 @@ class _FetchTimes(object):
   _ALPHA = 0.5
 
   def __init__(self, manifest):
-    self._path = os.path.join(manifest.repodir, '.repopickle_fetchtimes')
+    self._path = os.path.join(manifest.repodir, '.repo_fetchtimes.json')
     self._times = None
     self._seen = set()
 
@@ -779,22 +779,17 @@ class _FetchTimes(object):
   def _Load(self):
     if self._times is None:
       try:
-        f = open(self._path, 'rb')
-      except IOError:
-        self._times = {}
-        return self._times
-      try:
+        f = open(self._path)
         try:
-          self._times = pickle.load(f)
-        except IOError:
-          try:
-            os.remove(self._path)
-          except OSError:
-            pass
-          self._times = {}
-      finally:
-        f.close()
-    return self._times
+          self._times = json.load(f)
+        finally:
+          f.close()
+      except (IOError, ValueError):
+        try:
+          os.remove(self._path)
+        except OSError:
+          pass
+        self._times = {}
 
   def Save(self):
     if self._times is None:
@@ -808,13 +803,13 @@ class _FetchTimes(object):
       del self._times[name]
 
     try:
-      f = open(self._path, 'wb')
+      f = open(self._path, 'w')
       try:
-        pickle.dump(self._times, f)
-      except (IOError, OSError, pickle.PickleError):
-        try:
-          os.remove(self._path)
-        except OSError:
-          pass
-    finally:
-      f.close()
+        json.dump(self._times, f, indent=2)
+      finally:
+        f.close()
+    except (IOError, TypeError):
+      try:
+        os.remove(self._path)
+      except OSError:
+        pass