Merge pull request #169 from matrix-org/erikj/ultrajson

Use ultrajson when possible. Add option to turn off freezing of events.

synapse/app/homeserver.py

@@ -54,6 +54,8 @@ from synapse.rest.client.v1 import ClientV1RestResource
 from synapse.rest.client.v2_alpha import ClientV2AlphaRestResource
 from synapse.metrics.resource import MetricsResource, METRICS_PREFIX
 
+from synapse import events
+
 from daemonize import Daemonize
 import twisted.manhole.telnet
 
@@ -415,6 +417,8 @@ def setup(config_options):
     logger.info("Server hostname: %s", config.server_name)
     logger.info("Server version: %s", version_string)
 
+    events.USE_FROZEN_DICTS = config.use_frozen_dicts
+
     if re.search(":[0-9]+$", config.server_name):
         domain_with_port = config.server_name
     else:

synapse/config/server.py

@@ -28,6 +28,7 @@ class ServerConfig(Config):
         self.web_client = config["web_client"]
         self.soft_file_limit = config["soft_file_limit"]
         self.daemonize = config.get("daemonize")
+        self.use_frozen_dicts = config.get("use_frozen_dicts", True)
 
         # Attempt to guess the content_addr for the v0 content repostitory
         content_addr = config.get("content_addr")

synapse/events/__init__.py

@@ -16,6 +16,12 @@
 from synapse.util.frozenutils import freeze
 
 
+# Whether we should use frozen_dict in FrozenEvent. Using frozen_dicts prevents
+# bugs where we accidentally share e.g. signature dicts. However, converting
+# a dict to frozen_dicts is expensive.
+USE_FROZEN_DICTS = True
+
+
 class _EventInternalMetadata(object):
     def __init__(self, internal_metadata_dict):
         self.__dict__ = dict(internal_metadata_dict)
@@ -122,7 +128,10 @@ class FrozenEvent(EventBase):
 
         unsigned = dict(event_dict.pop("unsigned", {}))
 
+        if USE_FROZEN_DICTS:
             frozen_dict = freeze(event_dict)
+        else:
+            frozen_dict = dict(event_dict)
 
         super(FrozenEvent, self).__init__(
             frozen_dict,

synapse/http/server.py

@@ -19,9 +19,10 @@ from synapse.api.errors import (
 )
 from synapse.util.logcontext import LoggingContext, PreserveLoggingContext
 import synapse.metrics
+import synapse.events
 
 from syutil.jsonutil import (
-    encode_canonical_json, encode_pretty_printed_json
+    encode_canonical_json, encode_pretty_printed_json, encode_json
 )
 
 from twisted.internet import defer
@@ -168,9 +169,10 @@ class JsonResource(HttpServer, resource.Resource):
 
     _PathEntry = collections.namedtuple("_PathEntry", ["pattern", "callback"])
 
-    def __init__(self, hs):
+    def __init__(self, hs, canonical_json=True):
         resource.Resource.__init__(self)
 
+        self.canonical_json = canonical_json
         self.clock = hs.get_clock()
         self.path_regexs = {}
         self.version_string = hs.version_string
@@ -256,6 +258,7 @@ class JsonResource(HttpServer, resource.Resource):
             response_code_message=response_code_message,
             pretty_print=_request_user_agent_is_curl(request),
             version_string=self.version_string,
+            canonical_json=self.canonical_json,
         )
 
 
@@ -277,11 +280,17 @@ class RootRedirect(resource.Resource):
 
 def respond_with_json(request, code, json_object, send_cors=False,
                       response_code_message=None, pretty_print=False,
-                      version_string=""):
+                      version_string="", canonical_json=True):
     if pretty_print:
         json_bytes = encode_pretty_printed_json(json_object) + "\n"
     else:
+        if canonical_json:
             json_bytes = encode_canonical_json(json_object)
+        else:
+            json_bytes = encode_json(
+                json_object, using_frozen_dicts=synapse.events.USE_FROZEN_DICTS
+            )
+            logger.debug("json_bytes: %r %r", json_object, json_bytes)
 
     return respond_with_json_bytes(
         request, code, json_bytes,

synapse/python_dependencies.py

@@ -18,7 +18,7 @@ from distutils.version import LooseVersion
 logger = logging.getLogger(__name__)
 
 REQUIREMENTS = {
-    "syutil>=0.0.6": ["syutil>=0.0.6"],
+    "syutil>=0.0.7": ["syutil>=0.0.7"],
     "Twisted==14.0.2": ["twisted==14.0.2"],
     "service_identity>=1.0.0": ["service_identity>=1.0.0"],
     "pyopenssl>=0.14": ["OpenSSL>=0.14"],
@@ -30,6 +30,7 @@ REQUIREMENTS = {
     "frozendict>=0.4": ["frozendict"],
     "pillow": ["PIL"],
     "pydenticon": ["pydenticon"],
+    "ujson": ["ujson"],
 }
 CONDITIONAL_REQUIREMENTS = {
     "web_client": {
@@ -52,8 +53,8 @@ def github_link(project, version, egg):
 DEPENDENCY_LINKS = [
     github_link(
         project="matrix-org/syutil",
-        version="v0.0.6",
-        egg="syutil-0.0.6",
+        version="v0.0.7",
+        egg="syutil-0.0.7",
     ),
     github_link(
         project="matrix-org/matrix-angular-sdk",

synapse/rest/client/v1/__init__.py

@@ -25,7 +25,7 @@ class ClientV1RestResource(JsonResource):
     """A resource for version 1 of the matrix client API."""
 
     def __init__(self, hs):
-        JsonResource.__init__(self, hs)
+        JsonResource.__init__(self, hs, canonical_json=False)
         self.register_servlets(self, hs)
 
     @staticmethod

synapse/rest/client/v2_alpha/__init__.py

@@ -28,7 +28,7 @@ class ClientV2AlphaRestResource(JsonResource):
     """A resource for version 2 alpha of the matrix client API."""
 
     def __init__(self, hs):
-        JsonResource.__init__(self, hs)
+        JsonResource.__init__(self, hs, canonical_json=False)
         self.register_servlets(self, hs)
 
     @staticmethod

synapse/storage/events.py

@@ -17,7 +17,7 @@ from _base import SQLBaseStore, _RollbackButIsFineException
 
 from twisted.internet import defer, reactor
 
-from synapse.events import FrozenEvent
+from synapse.events import FrozenEvent, USE_FROZEN_DICTS
 from synapse.events.utils import prune_event
 
 from synapse.util.logcontext import preserve_context_over_deferred
@@ -26,11 +26,11 @@ from synapse.api.constants import EventTypes
 from synapse.crypto.event_signing import compute_event_reference_hash
 
 from syutil.base64util import decode_base64
-from syutil.jsonutil import encode_canonical_json
+from syutil.jsonutil import encode_json
 from contextlib import contextmanager
 
 import logging
-import simplejson as json
+import ujson as json
 
 logger = logging.getLogger(__name__)
 
@@ -166,8 +166,9 @@ class EventsStore(SQLBaseStore):
             allow_none=True,
         )
 
-        metadata_json = encode_canonical_json(
-            event.internal_metadata.get_dict()
+        metadata_json = encode_json(
+            event.internal_metadata.get_dict(),
+            using_frozen_dicts=USE_FROZEN_DICTS
         ).decode("UTF-8")
 
         # If we have already persisted this event, we don't need to do any
@@ -235,12 +236,14 @@ class EventsStore(SQLBaseStore):
                 "event_id": event.event_id,
                 "room_id": event.room_id,
                 "internal_metadata": metadata_json,
-                "json": encode_canonical_json(event_dict).decode("UTF-8"),
+                "json": encode_json(
+                    event_dict, using_frozen_dicts=USE_FROZEN_DICTS
+                ).decode("UTF-8"),
             },
         )
 
-        content = encode_canonical_json(
-            event.content
+        content = encode_json(
+            event.content, using_frozen_dicts=USE_FROZEN_DICTS
         ).decode("UTF-8")
 
         vals = {
@@ -266,8 +269,8 @@ class EventsStore(SQLBaseStore):
             ]
         }
 
-        vals["unrecognized_keys"] = encode_canonical_json(
-            unrec
+        vals["unrecognized_keys"] = encode_json(
+            unrec, using_frozen_dicts=USE_FROZEN_DICTS
         ).decode("UTF-8")
 
         sql = (