Files
odysseus/tests/test_email_list_stale_while_revalidate.py
T
Alexandre Teixeira 0133de6f0e feat(security): apply audited PR #6738 remediations and test hardening
- Enforce vision probe SSRF safe local host validation, SafeProbeBackend TOCTOU elimination, and session owner authorization
- Authorize session titling endpoints against registered owner configuration
- Sanitize sensitive metadata logging from agent loop debug logs
- Add async non-blocking execution for spec router tools and thinking decisions
- Harden email cache invalidation fencing and thread-safe mutation
- Add regression test suites for legacy browser authority, spec contracts, spec router request scoping, and task MCP delivery
2026-10-27 17:25:25 +00:00

218 lines
8.8 KiB
Python

"""Email list stale-while-revalidate (SWR) regression tests.
A plain list view must never block ~2-5 s on IMAP while we already hold a
recent snapshot or a durable index: it should return the snapshot immediately
(with ``sync.stale``) and refresh in the background. Blocking IMAP only remains
for a genuine cold start (no snapshot AND no indexed rows).
"""
import asyncio
import pathlib
from routes.email import email_routes as er
SRC = pathlib.Path(er.__file__).read_text(encoding="utf-8")
def _list_endpoint():
router = er.setup_email_routes()
for route in router.routes:
if getattr(route, "name", "") == "list_emails":
return route.endpoint
raise AssertionError("list_emails route not found")
def _call(endpoint, **over):
kwargs = dict(
folder="INBOX", limit=50, offset=0, filter="all", from_addr=None,
account_id=None, has_attachments=0, date_from=None, date_to=None,
cached_only=0, cache_bust=None, refresh=0, owner="pewds",
)
kwargs.update(over)
return asyncio.run(endpoint(**kwargs))
def test_list_returns_index_snapshot_with_a_stale_marker_without_blocking_on_imap(monkeypatch):
endpoint = _list_endpoint()
monkeypatch.setattr(
er, "_email_index_list",
lambda owner, account_id, folder, filter_, limit, offset, has_att: (
[{"uid": "11", "subject": "Indexed", "is_read": False}], 1, "2026-10-11T00:00:00Z",
),
)
# The synchronous path must never open IMAP; if it did the request would
# raise instead of returning the index snapshot immediately.
monkeypatch.setattr(er, "_imap_connect", lambda *a, **k: (_ for _ in ()).throw(
AssertionError("SWR must not open IMAP when an index snapshot exists")))
result = _call(endpoint, cache_bust="123")
assert result["total"] == 1
assert [e["uid"] for e in result["emails"]] == ["11"]
assert result["sync"]["stale"] is True
assert result["sync"]["source"] in ("index", "cache_stale")
def test_cold_start_with_no_snapshot_and_no_index_is_not_marked_stale(monkeypatch):
# With neither a snapshot nor an index, SWR must NOT fire: the request
# falls through to the real (blocking) path, so the response is not a
# stale-marked snapshot.
endpoint = _list_endpoint()
monkeypatch.setattr(er, "_email_index_list", lambda *a, **k: ([], 0, None))
monkeypatch.setattr(er, "_imap_connect", lambda *a, **k: (_ for _ in ()).throw(
RuntimeError("offline test")))
try:
result = _call(endpoint)
except Exception:
return # the cold path legitimately attempted real IMAP
assert isinstance(result, dict)
assert not (result.get("sync") or {}).get("stale")
def test_swr_contract_is_wired_into_the_list_path():
assert "_LIST_STALE" in SRC
assert "def _list_stale_put" in SRC
assert "def _schedule_list_refresh" in SRC
assert 'sync_meta["source"] = "cache_stale"' in SRC
assert 'sync_meta["stale"] = True' in SRC
# The manual-refresh contract the UI relies on must stay intact.
assert "manual_refresh = bool(refresh)" in SRC
assert "_invalidate_list_cache(account_id, folder)" in SRC
assert "cached_rows = {} if refresh else _email_index_rows" in SRC
def test_manual_refresh_fetches_mailbox_instead_of_returning_stale_index(monkeypatch):
class EmptyMailbox:
def select(self, folder, readonly=False):
return 'OK', [b'0']
def uid(self, command, *args):
assert command == 'SEARCH'
return 'OK', [b'']
def logout(self):
pass
endpoint = _list_endpoint()
monkeypatch.setattr(er, '_start_poller', lambda: None)
monkeypatch.setattr(er, '_imap_connect', lambda *a, **k: EmptyMailbox())
monkeypatch.setattr(er, '_email_index_list', lambda *a, **k: (
[{'uid': 'old', 'subject': 'Old'}], 1, 'yesterday'))
result = _call(endpoint, refresh=1)
assert result['emails'] == []
assert result['total'] == 0
assert not result.get('sync', {}).get('stale')
def test_stale_index_is_scoped_to_authenticated_owner(monkeypatch):
endpoint = _list_endpoint()
monkeypatch.setattr(er, '_start_poller', lambda: None)
monkeypatch.setattr(er, '_email_index_list', lambda owner, *a: (
[{'uid': owner, 'subject': owner}], 1, 'yesterday'))
monkeypatch.setattr(er, '_imap_connect', lambda *a, **k: (_ for _ in ()).throw(RuntimeError('offline')))
alice = _call(endpoint, owner='alice')
bob = _call(endpoint, owner='bob')
assert [mail['uid'] for mail in alice['emails']] == ['alice']
assert [mail['uid'] for mail in bob['emails']] == ['bob']
def test_inflight_refresh_cannot_restore_old_index_flags_after_mark_read(monkeypatch, tmp_path):
"""A delayed mailbox snapshot must not undo a completed owner mutation."""
import threading
from contextlib import contextmanager
from routes.email import email_helpers
database = tmp_path / 'email.db'
monkeypatch.setattr(email_helpers, 'SCHEDULED_DB', database)
monkeypatch.setattr(er, 'SCHEDULED_DB', database)
email_helpers._init_scheduled_db()
er._email_index_upsert('alice', 'acct', 'INBOX', [{
'uid': '42', 'subject': 'Example', 'flags': '', 'date_epoch': 1,
}])
fetched = threading.Event()
release = threading.Event()
threads = []
class Mailbox:
def __init__(self, mutation=False):
self.mutation = mutation
def select(self, *args, **kwargs):
return 'OK', [b'1']
def uid(self, command, *args):
if command == 'SEARCH':
return 'OK', [b'42']
if command == 'FETCH':
if self.mutation:
return 'OK', [b'1 (UID 42)']
threads.append(threading.current_thread())
fetched.set()
assert release.wait(5)
return 'OK', [(b'1 (UID 42 FLAGS () RFC822.SIZE 100 RFC822.HEADER {20}',
b'Subject: Example\r\n\r\n')]
if command == 'STORE':
return 'OK', []
raise AssertionError(command)
def logout(self):
pass
@contextmanager
def mailbox(*args, **kwargs):
yield Mailbox(mutation=True)
monkeypatch.setattr(er, '_start_poller', lambda: None)
monkeypatch.setattr(er, '_imap_connect', lambda *a, **kw: Mailbox())
monkeypatch.setattr(er, '_imap', mailbox)
router = er.setup_email_routes()
listing = next(r.endpoint for r in router.routes if getattr(r, 'name', '') == 'list_emails')
mark_read = next(r.endpoint for r in router.routes if getattr(r, 'name', '') == 'mark_read')
try:
assert _call(listing, owner='alice', account_id='acct')['emails'][0]['is_read'] is False
assert fetched.wait(5)
refresh_thread = threads[0]
result = asyncio.run(mark_read('42', folder='INBOX', account_id='acct', owner='alice'))
assert result['success'] is True
finally:
release.set()
refresh_thread.join(timeout=5)
assert not refresh_thread.is_alive()
rows, _, _ = er._email_index_list('alice', 'acct', 'INBOX', 'all', 50, 0)
assert rows[0]['is_read'] is True
def test_mark_read_keeps_other_owners_default_mailbox_unread(monkeypatch, tmp_path):
from contextlib import contextmanager
from routes.email import email_helpers
database = tmp_path / 'email.db'
monkeypatch.setattr(email_helpers, 'SCHEDULED_DB', database)
monkeypatch.setattr(er, 'SCHEDULED_DB', database)
email_helpers._init_scheduled_db()
for owner in ('alice', 'bob'):
er._email_index_upsert(owner, None, 'INBOX', [{
'uid': '42', 'subject': owner, 'flags': '', 'date_epoch': 1,
}])
class Mailbox:
def select(self, *args, **kwargs):
return 'OK', [b'1']
def uid(self, command, *args):
if command == 'FETCH':
return 'OK', [b'1 (UID 42)']
if command == 'STORE':
return 'OK', []
raise AssertionError(command)
@contextmanager
def mailbox(*args, **kwargs):
yield Mailbox()
monkeypatch.setattr(er, '_start_poller', lambda: None)
monkeypatch.setattr(er, '_imap', mailbox)
monkeypatch.setattr(er, '_imap_connect', lambda *a, **kw: (_ for _ in ()).throw(RuntimeError('offline')))
router = er.setup_email_routes()
listing = next(r.endpoint for r in router.routes if getattr(r, 'name', '') == 'list_emails')
mark_read = next(r.endpoint for r in router.routes if getattr(r, 'name', '') == 'mark_read')
assert _call(listing, owner='bob')['emails'][0]['is_read'] is False
assert asyncio.run(mark_read('42', folder='INBOX', account_id=None, owner='alice'))['success'] is True
assert _call(listing, owner='bob')['emails'][0]['is_read'] is False
rows, _, _ = er._email_index_list('alice', None, 'INBOX', 'all', 50, 0)
assert rows[0]['is_read'] is True