From 16090a077f9f387dcd42edabda58063e3df6b771 Mon Sep 17 00:00:00 2001
From: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com>
Date: Fri, 15 May 2020 17:17:42 +0100
Subject: [PATCH] Prevent 0-member/null room_version rooms from appearing in
 group room queries (#7465)

---
 changelog.d/7465.bugfix                       |  1 +
 .../storage/data_stores/main/group_server.py  | 92 ++++++++++++++++---
 2 files changed, 79 insertions(+), 14 deletions(-)
 create mode 100644 changelog.d/7465.bugfix

diff --git a/changelog.d/7465.bugfix b/changelog.d/7465.bugfix
new file mode 100644
index 000000000..1cbe50caa
--- /dev/null
+++ b/changelog.d/7465.bugfix
@@ -0,0 +1 @@
+Prevent rooms with 0 members or with invalid version strings from breaking group queries.
\ No newline at end of file
diff --git a/synapse/storage/data_stores/main/group_server.py b/synapse/storage/data_stores/main/group_server.py
index 0963e6c25..fb1361f1c 100644
--- a/synapse/storage/data_stores/main/group_server.py
+++ b/synapse/storage/data_stores/main/group_server.py
@@ -68,24 +68,78 @@ class GroupServerWorkerStore(SQLBaseStore):
             desc="get_invited_users_in_group",
         )
 
-    def get_rooms_in_group(self, group_id, include_private=False):
+    def get_rooms_in_group(self, group_id: str, include_private: bool = False):
+        """Retrieve the rooms that belong to a given group. Does not return rooms that
+        lack members.
+
+        Args:
+            group_id: The ID of the group to query for rooms
+            include_private: Whether to return private rooms in results
+
+        Returns:
+            Deferred[List[Dict[str, str|bool]]]: A list of dictionaries, each in the
+            form of:
+
+            {
+              "room_id": "!a_room_id:example.com",  # The ID of the room
+              "is_public": False                    # Whether this is a public room or not
+            }
+        """
         # TODO: Pagination
 
-        keyvalues = {"group_id": group_id}
-        if not include_private:
-            keyvalues["is_public"] = True
+        def _get_rooms_in_group_txn(txn):
+            sql = """
+            SELECT room_id, is_public FROM group_rooms
+                WHERE group_id = ?
+                AND room_id IN (
+                    SELECT group_rooms.room_id FROM group_rooms
+                    LEFT JOIN room_stats_current ON
+                        group_rooms.room_id = room_stats_current.room_id
+                        AND joined_members > 0
+                        AND local_users_in_room > 0
+                    LEFT JOIN rooms ON
+                        group_rooms.room_id = rooms.room_id
+                        AND (room_version <> '') = ?
+                )
+            """
+            args = [group_id, False]
 
-        return self.db.simple_select_list(
-            table="group_rooms",
-            keyvalues=keyvalues,
-            retcols=("room_id", "is_public"),
-            desc="get_rooms_in_group",
-        )
+            if not include_private:
+                sql += " AND is_public = ?"
+                args += [True]
 
-    def get_rooms_for_summary_by_category(self, group_id, include_private=False):
+            txn.execute(sql, args)
+
+            return [
+                {"room_id": room_id, "is_public": is_public}
+                for room_id, is_public in txn
+            ]
+
+        return self.db.runInteraction("get_rooms_in_group", _get_rooms_in_group_txn)
+
+    def get_rooms_for_summary_by_category(
+        self, group_id: str, include_private: bool = False,
+    ):
         """Get the rooms and categories that should be included in a summary request
 
-        Returns ([rooms], [categories])
+        Args:
+            group_id: The ID of the group to query the summary for
+            include_private: Whether to return private rooms in results
+
+        Returns:
+            Deferred[Tuple[List, Dict]]: A tuple containing:
+
+                * A list of dictionaries with the keys:
+                    * "room_id": str, the room ID
+                    * "is_public": bool, whether the room is public
+                    * "category_id": str|None, the category ID if set, else None
+                    * "order": int, the sort order of rooms
+
+                * A dictionary with the key:
+                    * category_id (str): a dictionary with the keys:
+                        * "is_public": bool, whether the category is public
+                        * "profile": str, the category profile
+                        * "order": int, the sort order of rooms in this category
         """
 
         def _get_rooms_for_summary_txn(txn):
@@ -97,13 +151,23 @@ class GroupServerWorkerStore(SQLBaseStore):
                 SELECT room_id, is_public, category_id, room_order
                 FROM group_summary_rooms
                 WHERE group_id = ?
+                AND room_id IN (
+                    SELECT group_rooms.room_id FROM group_rooms
+                    LEFT JOIN room_stats_current ON
+                        group_rooms.room_id = room_stats_current.room_id
+                        AND joined_members > 0
+                        AND local_users_in_room > 0
+                    LEFT JOIN rooms ON
+                        group_rooms.room_id = rooms.room_id
+                        AND (room_version <> '') = ?
+                )
             """
 
             if not include_private:
                 sql += " AND is_public = ?"
-                txn.execute(sql, (group_id, True))
+                txn.execute(sql, (group_id, False, True))
             else:
-                txn.execute(sql, (group_id,))
+                txn.execute(sql, (group_id, False))
 
             rooms = [
                 {