Parse json validation (#16923)

Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com>

3 files changed, 106 insertions, 47 deletions

diff --git a/synapse/http/servlet.py b/synapse/http/servlet.py
index 0ca08038f4..ab12951da8 100644
--- a/synapse/http/servlet.py
+++ b/synapse/http/servlet.py
@@ -23,6 +23,7 @@
 
 import enum
 import logging
+import urllib.parse as urlparse
 from http import HTTPStatus
 from typing import (
     TYPE_CHECKING,
@@ -450,6 +451,87 @@ def parse_string(
     )
 
 
+def parse_json(
+    request: Request,
+    name: str,
+    default: Optional[dict] = None,
+    required: bool = False,
+    encoding: str = "ascii",
+) -> Optional[JsonDict]:
+    """
+    Parse a JSON parameter from the request query string.
+
+    Args:
+        request: the twisted HTTP request.
+        name: the name of the query parameter.
+        default: value to use if the parameter is absent,
+            defaults to None.
+        required: whether to raise a 400 SynapseError if the
+           parameter is absent, defaults to False.
+        encoding: The encoding to decode the string content with.
+
+    Returns:
+        A JSON value, or `default` if the named query parameter was not found
+        and `required` was False.
+
+    Raises:
+        SynapseError if the parameter is absent and required, or if the
+            parameter is present and not a JSON object.
+    """
+    args: Mapping[bytes, Sequence[bytes]] = request.args  # type: ignore
+    return parse_json_from_args(
+        args,
+        name,
+        default,
+        required=required,
+        encoding=encoding,
+    )
+
+
+def parse_json_from_args(
+    args: Mapping[bytes, Sequence[bytes]],
+    name: str,
+    default: Optional[dict] = None,
+    required: bool = False,
+    encoding: str = "ascii",
+) -> Optional[JsonDict]:
+    """
+    Parse a JSON parameter from the request query string.
+
+    Args:
+        args: a mapping of request args as bytes to a list of bytes (e.g. request.args).
+        name: the name of the query parameter.
+        default: value to use if the parameter is absent,
+            defaults to None.
+        required: whether to raise a 400 SynapseError if the
+            parameter is absent, defaults to False.
+        encoding: the encoding to decode the string content with.
+
+        A JSON value, or `default` if the named query parameter was not found
+        and `required` was False.
+
+    Raises:
+        SynapseError if the parameter is absent and required, or if the
+            parameter is present and not a JSON object.
+    """
+    name_bytes = name.encode("ascii")
+
+    if name_bytes not in args:
+        if not required:
+            return default
+
+        message = f"Missing required integer query parameter {name}"
+        raise SynapseError(HTTPStatus.BAD_REQUEST, message, errcode=Codes.MISSING_PARAM)
+
+    json_str = parse_string_from_args(args, name, required=True, encoding=encoding)
+
+    try:
+        return json_decoder.decode(urlparse.unquote(json_str))
+    except Exception:
+        message = f"Query parameter {name} must be a valid JSON object"
+        raise SynapseError(HTTPStatus.BAD_REQUEST, message, errcode=Codes.NOT_JSON)
+
+
 EnumT = TypeVar("EnumT", bound=enum.Enum)
 
 
diff --git a/synapse/rest/admin/rooms.py b/synapse/rest/admin/rooms.py
index 4252f98a6c..0d86a4e15f 100644
--- a/synapse/rest/admin/rooms.py
+++ b/synapse/rest/admin/rooms.py
@@ -21,7 +21,6 @@
 import logging
 from http import HTTPStatus
 from typing import TYPE_CHECKING, List, Optional, Tuple, cast
-from urllib import parse as urlparse
 
 import attr
 
@@ -38,6 +37,7 @@ from synapse.http.servlet import (
     assert_params_in_dict,
     parse_enum,
     parse_integer,
+    parse_json,
     parse_json_object_from_request,
     parse_string,
 )
@@ -51,7 +51,6 @@ from synapse.storage.databases.main.room import RoomSortOrder
 from synapse.streams.config import PaginationConfig
 from synapse.types import JsonDict, RoomID, ScheduledTask, UserID, create_requester
 from synapse.types.state import StateFilter
-from synapse.util import json_decoder
 
 if TYPE_CHECKING:
     from synapse.api.auth import Auth
@@ -776,14 +775,8 @@ class RoomEventContextServlet(RestServlet):
         limit = parse_integer(request, "limit", default=10)
 
         # picking the API shape for symmetry with /messages
-        filter_str = parse_string(request, "filter", encoding="utf-8")
-        if filter_str:
-            filter_json = urlparse.unquote(filter_str)
-            event_filter: Optional[Filter] = Filter(
-                self._hs, json_decoder.decode(filter_json)
-            )
-        else:
-            event_filter = None
+        filter_json = parse_json(request, "filter", encoding="utf-8")
+        event_filter = Filter(self._hs, filter_json) if filter_json else None
 
         event_context = await self.room_context_handler.get_event_context(
             requester,
@@ -914,21 +907,16 @@ class RoomMessagesRestServlet(RestServlet):
         )
         # Twisted will have processed the args by now.
         assert request.args is not None
+
+        filter_json = parse_json(request, "filter", encoding="utf-8")
+        event_filter = Filter(self._hs, filter_json) if filter_json else None
+
         as_client_event = b"raw" not in request.args
-        filter_str = parse_string(request, "filter", encoding="utf-8")
-        if filter_str:
-            filter_json = urlparse.unquote(filter_str)
-            event_filter: Optional[Filter] = Filter(
-                self._hs, json_decoder.decode(filter_json)
-            )
-            if (
-                event_filter
-                and event_filter.filter_json.get("event_format", "client")
-                == "federation"
-            ):
-                as_client_event = False
-        else:
-            event_filter = None
+        if (
+            event_filter
+            and event_filter.filter_json.get("event_format", "client") == "federation"
+        ):
+            as_client_event = False
 
         msgs = await self._pagination_handler.get_messages(
             room_id=room_id,
diff --git a/synapse/rest/client/room.py b/synapse/rest/client/room.py
index 4eeadf8779..e4c7dd1a58 100644
--- a/synapse/rest/client/room.py
+++ b/synapse/rest/client/room.py
@@ -52,6 +52,7 @@ from synapse.http.servlet import (
     parse_boolean,
     parse_enum,
     parse_integer,
+    parse_json,
     parse_json_object_from_request,
     parse_string,
     parse_strings_from_args,
@@ -65,7 +66,6 @@ from synapse.rest.client.transactions import HttpTransactionCache
 from synapse.streams.config import PaginationConfig
 from synapse.types import JsonDict, Requester, StreamToken, ThirdPartyInstanceID, UserID
 from synapse.types.state import StateFilter
-from synapse.util import json_decoder
 from synapse.util.cancellation import cancellable
 from synapse.util.stringutils import parse_and_validate_server_name, random_string
 
@@ -703,21 +703,16 @@ class RoomMessageListRestServlet(RestServlet):
         )
         # Twisted will have processed the args by now.
         assert request.args is not None
+
+        filter_json = parse_json(request, "filter", encoding="utf-8")
+        event_filter = Filter(self._hs, filter_json) if filter_json else None
+
         as_client_event = b"raw" not in request.args
-        filter_str = parse_string(request, "filter", encoding="utf-8")
-        if filter_str:
-            filter_json = urlparse.unquote(filter_str)
-            event_filter: Optional[Filter] = Filter(
-                self._hs, json_decoder.decode(filter_json)
-            )
-            if (
-                event_filter
-                and event_filter.filter_json.get("event_format", "client")
-                == "federation"
-            ):
-                as_client_event = False
-        else:
-            event_filter = None
+        if (
+            event_filter
+            and event_filter.filter_json.get("event_format", "client") == "federation"
+        ):
+            as_client_event = False
 
         msgs = await self.pagination_handler.get_messages(
             room_id=room_id,
@@ -898,14 +893,8 @@ class RoomEventContextServlet(RestServlet):
         limit = parse_integer(request, "limit", default=10)
 
         # picking the API shape for symmetry with /messages
-        filter_str = parse_string(request, "filter", encoding="utf-8")
-        if filter_str:
-            filter_json = urlparse.unquote(filter_str)
-            event_filter: Optional[Filter] = Filter(
-                self._hs, json_decoder.decode(filter_json)
-            )
-        else:
-            event_filter = None
+        filter_json = parse_json(request, "filter", encoding="utf-8")
+        event_filter = Filter(self._hs, filter_json) if filter_json else None
 
         event_context = await self.room_context_handler.get_event_context(
             requester, room_id, event_id, limit, event_filter