Files
synapse/tests/utils.py
T
Erik JohnstonandClaude Opus 5.5 940c43b8fd Prepare chunks in Rust at the shadow and prepare levels
`Persister.prepare_chunk` runs the steps of the previous commits for a
chunk: the extremities and state delta, then the sliding sync table
changes and sticky events, then the chain cover index. The chain cover
step comes last as the only one that writes, so a failure in an earlier
step leaves nothing for Python to find. Sliding sync and sticky events
now come before the stream orderings are allocated, as nothing there
reads them. Each step reads the clock when it needs the time.

`_persist_event_batch` dispatches on `rust_event_persistence`:

- `shadow`: Rust runs first with its chain cover write rolled back, then
  Python. The results are compared and Python's are used. The comparison
  ignores orders Python leaves to chance and the IDs of new chains. Both
  rejecting the chunk with the same error is a match.
- `prepare` and above: Rust's results are used. A `SynapseError` is
  raised as Python would raise it. Python computes the chunk when Rust
  cannot resolve state in the room, or on any other Rust error.

Trial runs strict by default (`SYNAPSE_TEST_RUST_PERSIST_STRICT`), so a
mismatch or Rust error fails the test. Rust's counts and state res stats
are recorded by Python once it uses the result.

At `prepare`, the `_calculate_state_and_extrem` and
`calculate_chain_cover_index_for_events` timings give way to
`rust_prepare_chunk`. Python's state resolution and state group caches
are not filled, so creating a room costs one more transaction
(`RoomsCreateTestCase` counts it) and the first local event after a
fork resolves state again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LXMkcZVkcZhi21augBff4a
2026-10-17 10:45:55 +00:00

389 lines
15 KiB
Python

#
# This file is licensed under the Affero General Public License (AGPL) version 3.
#
# Copyright 2014-2016 OpenMarket Ltd
# Copyright (C) 2023 New Vector, Ltd
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.
#
# See the GNU Affero General Public License for more details:
# <https://www.gnu.org/licenses/agpl-3.0.html>.
#
# Originally licensed under the Apache License, Version 2.0:
# <http://www.apache.org/licenses/LICENSE-2.0>.
#
# [This file includes modifications made by New Vector Limited]
#
#
import atexit
import os
import signal
from types import FrameType, TracebackType
from typing import (
Literal,
TypeVar,
overload,
)
from synapse.api.constants import EventTypes
from synapse.api.room_versions import RoomVersions
from synapse.config.homeserver import HomeServerConfig
from synapse.config.server import DEFAULT_ROOM_VERSION
from synapse.server import HomeServer
from synapse.storage.database import LoggingDatabaseConnection
from synapse.storage.engines import create_engine
from synapse.storage.prepare_database import prepare_database
try:
import authlib # noqa: F401
HAS_AUTHLIB = True
except ImportError:
HAS_AUTHLIB = False
# set this to True to run the tests against postgres instead of sqlite.
#
# When running under postgres, we first create a base database with the name
# POSTGRES_BASE_DB and update it to the current schema. Then, for each test case, we
# create another unique database, using the base database as a template.
USE_POSTGRES_FOR_TESTS = os.environ.get("SYNAPSE_POSTGRES", False)
# Run the Postgres tests against the native Rust driver instead of psycopg2.
# Only meaningful together with SYNAPSE_POSTGRES.
USE_RUST_DRIVER_FOR_TESTS = os.environ.get("SYNAPSE_TEST_RUST_DRIVER", False)
# Set `experimental_features.rust_state_res` ("off", "shadow" or "on") for every
# test homeserver, overriding the test's own config. The Rust state resolution
# only runs on the native Rust driver (SYNAPSE_TEST_RUST_DRIVER).
RUST_STATE_RES_FOR_TESTS = os.environ.get("SYNAPSE_TEST_RUST_STATE_RES")
# As above, for `experimental_features.rust_sync`.
RUST_SYNC_FOR_TESTS = os.environ.get("SYNAPSE_TEST_RUST_SYNC")
# A Rust error, or in `shadow` mode a room whose Rust entry differs from the
# Python one, fails the test, so that the suite is a differential test of the
# two. Set to 0 to only count and log them, as production does.
RUST_SYNC_STRICT_FOR_TESTS = os.environ.get("SYNAPSE_TEST_RUST_SYNC_STRICT", "1") != "0"
# As above, for `experimental_features.rust_event_persistence`.
RUST_PERSIST_FOR_TESTS = os.environ.get("SYNAPSE_TEST_RUST_PERSIST")
# As `SYNAPSE_TEST_RUST_SYNC_STRICT`, for the persister.
RUST_PERSIST_STRICT_FOR_TESTS = (
os.environ.get("SYNAPSE_TEST_RUST_PERSIST_STRICT", "1") != "0"
)
# At `rust_federation_sending: transactions`, check every transaction body Rust
# builds against the one Python would build, and fail the test if they differ.
RUST_FED_STRICT_FOR_TESTS = os.environ.get("SYNAPSE_TEST_RUST_FED_STRICT", "1") != "0"
LEAVE_DB = os.environ.get("SYNAPSE_LEAVE_DB", False)
POSTGRES_USER = os.environ.get("SYNAPSE_POSTGRES_USER", None)
POSTGRES_HOST = os.environ.get("SYNAPSE_POSTGRES_HOST", None)
POSTGRES_PASSWORD = os.environ.get("SYNAPSE_POSTGRES_PASSWORD", None)
POSTGRES_PORT = (
int(os.environ["SYNAPSE_POSTGRES_PORT"])
if "SYNAPSE_POSTGRES_PORT" in os.environ
else None
)
POSTGRES_BASE_DB = "_synapse_unit_tests_base_%s" % (os.getpid(),)
# When debugging a specific test, it's occasionally useful to write the
# DB to disk and query it with the sqlite CLI.
SQLITE_PERSIST_DB = os.environ.get("SYNAPSE_TEST_PERSIST_SQLITE_DB") is not None
# the dbname we will connect to in order to create the base database.
POSTGRES_DBNAME_FOR_INITIAL_CREATE = "postgres"
def setupdb() -> None:
# If we're using PostgreSQL, set up the db once
if USE_POSTGRES_FOR_TESTS:
# create a PostgresEngine
db_engine = create_engine({"name": "psycopg2", "args": {}})
# connect to postgres to create the base database.
db_conn = db_engine.module.connect(
user=POSTGRES_USER,
host=POSTGRES_HOST,
port=POSTGRES_PORT,
password=POSTGRES_PASSWORD,
dbname=POSTGRES_DBNAME_FOR_INITIAL_CREATE,
)
db_engine.attempt_to_set_autocommit(db_conn, autocommit=True)
cur = db_conn.cursor()
cur.execute("DROP DATABASE IF EXISTS %s;" % (POSTGRES_BASE_DB,))
cur.execute(
"CREATE DATABASE %s ENCODING 'UTF8' LC_COLLATE='C' LC_CTYPE='C' "
"template=template0;" % (POSTGRES_BASE_DB,)
)
cur.close()
db_conn.close()
# Set up in the db
db_conn = db_engine.module.connect(
dbname=POSTGRES_BASE_DB,
user=POSTGRES_USER,
host=POSTGRES_HOST,
port=POSTGRES_PORT,
password=POSTGRES_PASSWORD,
)
logging_conn = LoggingDatabaseConnection(
conn=db_conn,
engine=db_engine,
default_txn_name="tests",
server_name="test_server",
)
prepare_database(logging_conn, db_engine, None)
logging_conn.close()
def _cleanup() -> None:
db_conn = db_engine.module.connect(
user=POSTGRES_USER,
host=POSTGRES_HOST,
port=POSTGRES_PORT,
password=POSTGRES_PASSWORD,
dbname=POSTGRES_DBNAME_FOR_INITIAL_CREATE,
)
db_engine.attempt_to_set_autocommit(db_conn, autocommit=True)
cur = db_conn.cursor()
cur.execute("DROP DATABASE IF EXISTS %s;" % (POSTGRES_BASE_DB,))
cur.close()
db_conn.close()
atexit.register(_cleanup)
@overload
def default_config(
*, server_name: str, parse: Literal[False] = ...
) -> dict[str, object]: ...
@overload
def default_config(*, server_name: str, parse: Literal[True]) -> HomeServerConfig: ...
def default_config(
*, server_name: str, parse: bool = False
) -> dict[str, object] | HomeServerConfig:
"""
Create a reasonable test config.
Args:
server_name: homeserver name
parse: TODO
"""
config_dict = {
"server_name": server_name,
# Setting this to an empty list turns off federation sending.
"federation_sender_instances": [],
"media_store_path": "media",
# the test signing key is just an arbitrary ed25519 key to keep the config
# parser happy
"signing_key": "ed25519 a_lPym qvioDNmfExFBRPgdTU+wtFYKq4JfwFRv7sYVgWvmgJg",
# Disable trusted key servers, otherwise unit tests might try to actually
# reach out to matrix.org.
"trusted_key_servers": [],
"event_cache_size": 1,
"enable_registration": True,
"enable_registration_captcha": False,
"macaroon_secret_key": "not even a little secret",
"password_providers": [],
"worker_app": None,
"block_non_admin_invites": False,
"federation_domain_whitelist": None,
"filter_timeline_limit": 5000,
"user_directory_search_all_users": False,
"user_consent_server_notice_content": None,
"block_events_without_consent_error": None,
"user_consent_at_registration": False,
"user_consent_policy_name": "Privacy Policy",
"media_storage_providers": [],
"autocreate_auto_join_rooms": True,
"auto_join_rooms": [],
"limit_usage_by_mau": False,
"hs_disabled": False,
"hs_disabled_message": "",
"max_mau_value": 50,
"mau_trial_days": 0,
"mau_stats_only": False,
"mau_limits_reserved_threepids": [],
"admin_contact": None,
"rc_message": {"per_second": 10000, "burst_count": 10000},
"rc_registration": {"per_second": 10000, "burst_count": 10000},
"rc_login": {
"address": {"per_second": 10000, "burst_count": 10000},
"account": {"per_second": 10000, "burst_count": 10000},
"failed_attempts": {"per_second": 10000, "burst_count": 10000},
},
"rc_joins": {
"local": {"per_second": 10000, "burst_count": 10000},
"remote": {"per_second": 10000, "burst_count": 10000},
},
"rc_joins_per_room": {"per_second": 10000, "burst_count": 10000},
"rc_invites": {
"per_room": {"per_second": 10000, "burst_count": 10000},
"per_user": {"per_second": 10000, "burst_count": 10000},
"per_issuer": {"per_second": 10000, "burst_count": 10000},
},
"rc_room_creation": {"per_second": 10000, "burst_count": 10000},
"rc_3pid_validation": {"per_second": 10000, "burst_count": 10000},
"rc_presence": {"per_user": {"per_second": 10000, "burst_count": 10000}},
"saml2_enabled": False,
"public_baseurl": None,
"default_identity_server": None,
"key_refresh_interval": 24 * 60 * 60 * 1000,
"old_signing_keys": {},
"tls_fingerprints": [],
"use_frozen_dicts": False,
# We need a sane default_room_version, otherwise attempts to create
# rooms will fail.
"default_room_version": DEFAULT_ROOM_VERSION,
# disable user directory updates, because they get done in the
# background, which upsets the test runner. Setting this to an
# (obviously) fake worker name disables updating the user directory.
"update_user_directory_from_worker": "does_not_exist_worker_name",
"caches": {"global_factor": 1, "sync_response_cache_duration": 0},
"listeners": [{"port": 0, "type": "http"}],
}
if parse:
config = HomeServerConfig()
config.parse_config_dict(config_dict, "", "")
return config
return config_dict
def mock_getRawHeaders(headers=None): # type: ignore[no-untyped-def]
headers = headers if headers is not None else {}
def getRawHeaders(name, default=None): # type: ignore[no-untyped-def]
# If the requested header is present, the real twisted function returns
# List[str] if name is a str and List[bytes] if name is a bytes.
# This mock doesn't support that behaviour.
# Fortunately, none of the current callers of mock_getRawHeaders() provide a
# headers dict, so we don't encounter this discrepancy in practice.
return headers.get(name, default)
return getRawHeaders
async def create_room(hs: HomeServer, room_id: str, creator_id: str) -> None:
"""Creates and persist a creation event for the given room"""
persistence_store = hs.get_storage_controllers().persistence
assert persistence_store is not None
store = hs.get_datastores().main
event_builder_factory = hs.get_event_builder_factory()
event_creation_handler = hs.get_event_creation_handler()
await store.store_room(
room_id=room_id,
room_creator_user_id=creator_id,
is_public=False,
room_version=RoomVersions.V1,
)
builder = event_builder_factory.for_room_version(
RoomVersions.V1,
{
"type": EventTypes.Create,
"state_key": "",
"sender": creator_id,
"room_id": room_id,
"content": {},
},
)
event, unpersisted_context = await event_creation_handler.create_new_client_event(
builder
)
context = await unpersisted_context.persist(event)
await persistence_store.persist_event(event, context)
T = TypeVar("T")
def checked_cast(type: type[T], x: object) -> T:
"""A version of typing.cast that is checked at runtime.
We have our own function for this for two reasons:
1. typing.cast itself is deliberately a no-op at runtime, see
https://docs.python.org/3/library/typing.html#typing.cast
2. To help workaround a mypy-zope bug https://github.com/Shoobx/mypy-zope/issues/91
where mypy would erroneously consider `isinstance(x, type)` to be false in all
circumstances.
For this to make sense, `T` needs to be something that `isinstance` can check; see
https://docs.python.org/3/library/functions.html?highlight=isinstance#isinstance
https://docs.python.org/3/glossary.html#term-abstract-base-class
https://docs.python.org/3/library/typing.html#typing.runtime_checkable
for more details.
"""
assert isinstance(x, type)
return x
class TestTimeout(Exception):
pass
class test_timeout:
"""
FIXME: This implementation is not robust against other code tight-looping and
preventing the signals propagating and timing out the test. You may need to add
`time.sleep(0.1)` to your code in order to allow this timeout to work correctly.
```py
with test_timeout(3):
while True:
my_checking_func()
time.sleep(0.1)
```
Args:
seconds: How long to allow the block to run for before raising
`TestTimeout`.
error_message: Extra text to append to the `TestTimeout` message.
cpu_time: If `True`, `seconds` is a budget of CPU time (user + system,
across all threads) consumed by the process rather than wall-clock
time. Useful for performance-regression tests, as time spent
blocked on I/O (e.g. waiting on the database) or lost to a loaded
CI machine doesn't count against the budget. Note that a block
which hangs while consuming *no* CPU will never trip this variant.
"""
def __init__(
self,
seconds: float,
error_message: str | None = None,
*,
cpu_time: bool = False,
) -> None:
self.error_message = f"Test timed out after {seconds}s of {'CPU' if cpu_time else 'wall-clock'} time"
if error_message is not None:
self.error_message += f": {error_message}"
self.seconds = seconds
self.cpu_time = cpu_time
def handle_timeout(self, signum: int, frame: FrameType | None) -> None:
raise TestTimeout(self.error_message)
def __enter__(self) -> None:
if self.cpu_time:
# `ITIMER_PROF` counts down against process CPU time (user +
# system) and delivers `SIGPROF` when it expires.
signal.signal(signal.SIGPROF, self.handle_timeout)
signal.setitimer(signal.ITIMER_PROF, self.seconds)
else:
signal.signal(signal.SIGALRM, self.handle_timeout)
signal.setitimer(signal.ITIMER_REAL, self.seconds)
def __exit__(
self,
exc_type: type[BaseException] | None,
exc_val: BaseException | None,
exc_tb: TracebackType | None,
) -> None:
signal.setitimer(signal.ITIMER_PROF if self.cpu_time else signal.ITIMER_REAL, 0)