Finish redaction algorithm.

5 files changed, 33 insertions, 25 deletions

diff --git a/synapse/api/events/__init__.py b/synapse/api/events/__init__.py
index 8d65c29ac1..f1e53f23ab 100644
--- a/synapse/api/events/__init__.py
+++ b/synapse/api/events/__init__.py
@@ -86,8 +86,8 @@ class SynapseEvent(JsonEncodedObject):
 
     def __init__(self, raises=True, **kwargs):
         super(SynapseEvent, self).__init__(**kwargs)
-        if "content" in kwargs:
-            self.check_json(self.content, raises=raises)
+        # if "content" in kwargs:
+        #     self.check_json(self.content, raises=raises)
 
     def get_content_template(self):
         """ Retrieve the JSON template for this event as a dict.
diff --git a/synapse/api/events/utils.py b/synapse/api/events/utils.py
index 5fc79105b5..802648f8f7 100644
--- a/synapse/api/events/utils.py
+++ b/synapse/api/events/utils.py
@@ -18,24 +18,31 @@ from .room import (
     RoomAliasesEvent, RoomCreateEvent,
 )
 
+
 def prune_event(event):
-    """ Prunes the given event of all keys we don't know about or think could
-    potentially be dodgy.
+    """ Returns a pruned version of the given event, which removes all keys we
+    don't know about or think could potentially be dodgy.
 
     This is used when we "redact" an event. We want to remove all fields that
     the user has specified, but we do want to keep necessary information like
     type, state_key etc.
     """
-    return _prune_event_or_pdu(event.type, event)
-
-def prune_pdu(pdu):
-    """Removes keys that contain unrestricted and non-essential data from a PDU
-    """
-    return _prune_event_or_pdu(pdu.type, pdu)
+    event_type = event.type
 
-def _prune_event_or_pdu(event_type, event):
-    # Remove all extraneous fields.
-    event.unrecognized_keys = {}
+    allowed_keys = [
+        "event_id",
+        "user_id",
+        "room_id",
+        "hashes",
+        "signatures",
+        "content",
+        "type",
+        "state_key",
+        "depth",
+        "prev_events",
+        "prev_state",
+        "auth_events",
+    ]
 
     new_content = {}
 
@@ -65,6 +72,12 @@ def _prune_event_or_pdu(event_type, event):
     elif event_type == RoomAliasesEvent.TYPE:
         add_fields("aliases")
 
-    event.content = new_content
+    allowed_fields = {
+        k: v
+        for k, v in event.get_full_dict().items()
+        if k in allowed_keys
+    }
+
+    allowed_fields["content"] = new_content
 
-    return event
+    return type(event)(**allowed_fields)
diff --git a/synapse/crypto/event_signing.py b/synapse/crypto/event_signing.py
index 7d800615fe..056e8f6ca4 100644
--- a/synapse/crypto/event_signing.py
+++ b/synapse/crypto/event_signing.py
@@ -55,9 +55,7 @@ def _compute_content_hash(event, hash_algorithm):
 
 
 def compute_event_reference_hash(event, hash_algorithm=hashlib.sha256):
-    # FIXME(erikj): GenericEvent!
-    tmp_event = GenericEvent(**event.get_full_dict())
-    tmp_event = prune_event(tmp_event)
+    tmp_event = prune_event(event)
     event_json = tmp_event.get_dict()
     event_json.pop("signatures", None)
     event_json.pop("age_ts", None)
@@ -68,8 +66,7 @@ def compute_event_reference_hash(event, hash_algorithm=hashlib.sha256):
 
 
 def compute_event_signature(event, signature_name, signing_key):
-    tmp_event = copy.deepcopy(event)
-    tmp_event = prune_event(tmp_event)
+    tmp_event = prune_event(event)
     redact_json = tmp_event.get_full_dict()
     redact_json.pop("signatures", None)
     redact_json.pop("age_ts", None)
diff --git a/synapse/federation/units.py b/synapse/federation/units.py
index 2070ffe1e2..d98014cac7 100644
--- a/synapse/federation/units.py
+++ b/synapse/federation/units.py
@@ -56,17 +56,15 @@ class Pdu(JsonEncodedObject):
         "origin_server_ts",
         "type",
         "destinations",
-        "transaction_id",
         "prev_events",
         "depth",
         "content",
-        "outlier",
         "hashes",
+        "user_id",
+        "auth_events",
         "signatures",  # Below this are keys valid only for State Pdus.
         "state_key",
         "prev_state",
-        "required_power_level",
-        "user_id",
     ]
 
     internal_keys = [
diff --git a/synapse/storage/_base.py b/synapse/storage/_base.py
index 9aa404695d..3ab81a78d5 100644
--- a/synapse/storage/_base.py
+++ b/synapse/storage/_base.py
@@ -509,7 +509,7 @@ class SQLBaseStore(object):
                 )
 
                 if del_evs:
-                    prune_event(ev)
+                    ev = prune_event(ev)
                     ev.redacted_because = del_evs[0]
 
         return events