Properly parse event_fields in filters (#15607)

The event_fields property in filters should use the proper
escape rules, namely backslashes can be escaped with
an additional backslash.

This adds tests (adapted from matrix-js-sdk) and implements
the logic to properly split the event_fields strings.

5 files changed, 99 insertions, 34 deletions

diff --git a/changelog.d/15607.bugfix b/changelog.d/15607.bugfix
new file mode 100644
index 0000000000..a2767adbe2
--- /dev/null
+++ b/changelog.d/15607.bugfix
@@ -0,0 +1 @@
+Fix a long-standing bug where filters with multiple backslashes were rejected.
diff --git a/synapse/api/filtering.py b/synapse/api/filtering.py
index de7c56bc0f..82aeef8d19 100644
--- a/synapse/api/filtering.py
+++ b/synapse/api/filtering.py
@@ -128,20 +128,7 @@ USER_FILTER_SCHEMA = {
         "account_data": {"$ref": "#/definitions/filter"},
         "room": {"$ref": "#/definitions/room_filter"},
         "event_format": {"type": "string", "enum": ["client", "federation"]},
-        "event_fields": {
-            "type": "array",
-            "items": {
-                "type": "string",
-                # Don't allow '\\' in event field filters. This makes matching
-                # events a lot easier as we can then use a negative lookbehind
-                # assertion to split '\.' If we allowed \\ then it would
-                # incorrectly split '\\.' See synapse.events.utils.serialize_event
-                #
-                # Note that because this is a regular expression, we have to escape
-                # each backslash in the pattern.
-                "pattern": r"^((?!\\\\).)*$",
-            },
-        },
+        "event_fields": {"type": "array", "items": {"type": "string"}},
     },
     "additionalProperties": True,  # Allow new fields for forward compatibility
 }
diff --git a/synapse/events/utils.py b/synapse/events/utils.py
index e6d040176b..e7b7b78b84 100644
--- a/synapse/events/utils.py
+++ b/synapse/events/utils.py
@@ -22,6 +22,7 @@ from typing import (
     Iterable,
     List,
     Mapping,
+    Match,
     MutableMapping,
     Optional,
     Union,
@@ -46,12 +47,10 @@ if TYPE_CHECKING:
     from synapse.handlers.relations import BundledAggregations
 
 
-# Split strings on "." but not "\." This uses a negative lookbehind assertion for '\'
-# (?<!stuff) matches if the current position in the string is not preceded
-# by a match for 'stuff'.
-# TODO: This is fast, but fails to handle "foo\\.bar" which should be treated as
-#       the literal fields "foo\" and "bar" but will instead be treated as "foo\\.bar"
-SPLIT_FIELD_REGEX = re.compile(r"(?<!\\)\.")
+# Split strings on "." but not "\." (or "\\\.").
+SPLIT_FIELD_REGEX = re.compile(r"\\*\.")
+# Find escaped characters, e.g. those with a \ in front of them.
+ESCAPE_SEQUENCE_PATTERN = re.compile(r"\\(.)")
 
 CANONICALJSON_MAX_INT = (2**53) - 1
 CANONICALJSON_MIN_INT = -CANONICALJSON_MAX_INT
@@ -253,6 +252,57 @@ def _copy_field(src: JsonDict, dst: JsonDict, field: List[str]) -> None:
     sub_out_dict[key_to_move] = sub_dict[key_to_move]
 
 
+def _escape_slash(m: Match[str]) -> str:
+    """
+    Replacement function; replace a backslash-backslash or backslash-dot with the
+    second character. Leaves any other string alone.
+    """
+    if m.group(1) in ("\\", "."):
+        return m.group(1)
+    return m.group(0)
+
+
+def _split_field(field: str) -> List[str]:
+    """
+    Splits strings on unescaped dots and removes escaping.
+
+    Args:
+        field: A string representing a path to a field.
+
+    Returns:
+        A list of nested fields to traverse.
+    """
+
+    # Convert the field and remove escaping:
+    #
+    # 1. "content.body.thing\.with\.dots"
+    # 2. ["content", "body", "thing\.with\.dots"]
+    # 3. ["content", "body", "thing.with.dots"]
+
+    # Find all dots (and their preceding backslashes). If the dot is unescaped
+    # then emit a new field part.
+    result = []
+    prev_start = 0
+    for match in SPLIT_FIELD_REGEX.finditer(field):
+        # If the match is an *even* number of characters than the dot was escaped.
+        if len(match.group()) % 2 == 0:
+            continue
+
+        # Add a new part (up to the dot, exclusive) after escaping.
+        result.append(
+            ESCAPE_SEQUENCE_PATTERN.sub(
+                _escape_slash, field[prev_start : match.end() - 1]
+            )
+        )
+        prev_start = match.end()
+
+    # Add any part of the field after the last unescaped dot. (Note that if the
+    # character is a dot this correctly adds a blank string.)
+    result.append(re.sub(r"\\(.)", _escape_slash, field[prev_start:]))
+
+    return result
+
+
 def only_fields(dictionary: JsonDict, fields: List[str]) -> JsonDict:
     """Return a new dict with only the fields in 'dictionary' which are present
     in 'fields'.
@@ -260,7 +310,7 @@ def only_fields(dictionary: JsonDict, fields: List[str]) -> JsonDict:
     If there are no event fields specified then all fields are included.
     The entries may include '.' characters to indicate sub-fields.
     So ['content.body'] will include the 'body' field of the 'content' object.
-    A literal '.' character in a field name may be escaped using a '\'.
+    A literal '.' or '\' character in a field name may be escaped using a '\'.
 
     Args:
         dictionary: The dictionary to read from.
@@ -275,13 +325,7 @@ def only_fields(dictionary: JsonDict, fields: List[str]) -> JsonDict:
 
     # for each field, convert it:
     # ["content.body.thing\.with\.dots"] => [["content", "body", "thing\.with\.dots"]]
-    split_fields = [SPLIT_FIELD_REGEX.split(f) for f in fields]
-
-    # for each element of the output array of arrays:
-    # remove escaping so we can use the right key names.
-    split_fields[:] = [
-        [f.replace(r"\.", r".") for f in field_array] for field_array in split_fields
-    ]
+    split_fields = [_split_field(f) for f in fields]
 
     output: JsonDict = {}
     for field_array in split_fields:
diff --git a/tests/api/test_filtering.py b/tests/api/test_filtering.py
index 222449baac..aa6af5ad7b 100644
--- a/tests/api/test_filtering.py
+++ b/tests/api/test_filtering.py
@@ -48,8 +48,6 @@ class FilteringTestCase(unittest.HomeserverTestCase):
         invalid_filters: List[JsonDict] = [
             # `account_data` must be a dictionary
             {"account_data": "Hello World"},
-            # `event_fields` entries must not contain backslashes
-            {"event_fields": [r"\\foo"]},
             # `event_format` must be "client" or "federation"
             {"event_format": "other"},
             # `not_rooms` must contain valid room IDs
@@ -114,10 +112,6 @@ class FilteringTestCase(unittest.HomeserverTestCase):
                 "event_format": "client",
                 "event_fields": ["type", "content", "sender"],
             },
-            # a single backslash should be permitted (though it is debatable whether
-            # it should be permitted before anything other than `.`, and what that
-            # actually means)
-            #
             # (note that event_fields is implemented in
             # synapse.events.utils.serialize_event, and so whether this actually works
             # is tested elsewhere. We just want to check that it is allowed through the
diff --git a/tests/events/test_utils.py b/tests/events/test_utils.py
index e40eac2eb0..c9a610db9a 100644
--- a/tests/events/test_utils.py
+++ b/tests/events/test_utils.py
@@ -16,6 +16,7 @@ import unittest as stdlib_unittest
 from typing import Any, List, Mapping, Optional
 
 import attr
+from parameterized import parameterized
 
 from synapse.api.constants import EventContentFields
 from synapse.api.room_versions import RoomVersions
@@ -23,6 +24,7 @@ from synapse.events import EventBase, make_event_from_dict
 from synapse.events.utils import (
     PowerLevelsContent,
     SerializeEventConfig,
+    _split_field,
     copy_and_fixup_power_levels_contents,
     maybe_upsert_event_field,
     prune_event,
@@ -794,3 +796,40 @@ class CopyPowerLevelsContentTestCase(stdlib_unittest.TestCase):
     def test_invalid_nesting_raises_type_error(self) -> None:
         with self.assertRaises(TypeError):
             copy_and_fixup_power_levels_contents({"a": {"b": {"c": 1}}})  # type: ignore[dict-item]
+
+
+class SplitFieldTestCase(stdlib_unittest.TestCase):
+    @parameterized.expand(
+        [
+            # A field with no dots.
+            ["m", ["m"]],
+            # Simple dotted fields.
+            ["m.foo", ["m", "foo"]],
+            ["m.foo.bar", ["m", "foo", "bar"]],
+            # Backslash is used as an escape character.
+            [r"m\.foo", ["m.foo"]],
+            [r"m\\.foo", ["m\\", "foo"]],
+            [r"m\\\.foo", [r"m\.foo"]],
+            [r"m\\\\.foo", ["m\\\\", "foo"]],
+            [r"m\foo", [r"m\foo"]],
+            [r"m\\foo", [r"m\foo"]],
+            [r"m\\\foo", [r"m\\foo"]],
+            [r"m\\\\foo", [r"m\\foo"]],
+            # Ensure that escapes at the end don't cause issues.
+            ["m.foo\\", ["m", "foo\\"]],
+            ["m.foo\\", ["m", "foo\\"]],
+            [r"m.foo\.", ["m", "foo."]],
+            [r"m.foo\\.", ["m", "foo\\", ""]],
+            [r"m.foo\\\.", ["m", r"foo\."]],
+            # Empty parts (corresponding to properties which are an empty string) are allowed.
+            [".m", ["", "m"]],
+            ["..m", ["", "", "m"]],
+            ["m.", ["m", ""]],
+            ["m..", ["m", "", ""]],
+            ["m..foo", ["m", "", "foo"]],
+            # Invalid escape sequences.
+            [r"\m", [r"\m"]],
+        ]
+    )
+    def test_split_field(self, input: str, expected: str) -> None:
+        self.assertEqual(_split_field(input), expected)