`hermes update` started in an interpreter that had imported the PRE-pull tree, then kept running every post-swap phase (dependency sync, Node/web/Desktop builds, maintenance, config migration, fleet restart, verification, receipt) in that same process, lazily importing NEW source into an OLD `sys.modules` graph. Any rename between the two commits surfaced as an ImportError/AttributeError inside the updater after the code swap had already succeeded (#87134, #111271, #112465, #112558, #112604). Each incident added another purge, reload list or per-step isolation, and each moved the crash to the next module nobody had listed. The pre-pull process now stops at the swap: it writes the open receipt, the pre-update fleet plan, the pre-update version/active features and the Windows pause token to a hand-off file and re-executes `hermes update <same flags> --post-swap <file>` under the venv interpreter. The child imports exclusively from the pulled tree, resumes the receipt and owns the rest of the run; the parent relays its exit code. Git and ZIP paths both hand off. On Windows, when the updater runs from `hermes.exe`, the child is spawned detached exactly as the shim hand-off already did (the shim cannot be awaited while the sync must replace it). With no pulled code ever executing in a pre-pull interpreter, the stale-module layer is dead and removed: `_purge_stale_hermes_modules`, `_stale_purge_prefixes`, `_evict_module`, `_STALE_PURGE_*`, `_reload_updated_runtime_modules`, `_reload_process_scan_modules`, `_reload_config_modules`, `_UPDATE_RUNTIME_RELOAD_MODULES` and their tests. `_run_config_check_fresh` / `_run_migrate_config_fresh` keep their names and simply call the config API. Tests: the hand-off boundary (child argv/env, detached receipt + plan in the payload, exit-code relay) and the child side (receipt resumed with its history, plan rebuilt, pre-update snapshots taken from the payload). Mocked updater flows run the tail in-process through the same payload round-trip (autouse fixture; opt out with `@pytest.mark.real_post_swap_handoff`).
158 lines
7.2 KiB
Python
158 lines
7.2 KiB
Python
"""Fixtures shared across hermes_cli tests."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import pytest
|
|
|
|
|
|
@pytest.fixture
|
|
def all_assignees_spawnable(monkeypatch):
|
|
"""Pretend every assignee maps to a real Hermes profile.
|
|
|
|
Most dispatcher tests use synthetic assignees ("alice", "bob") that
|
|
don't correspond to actual profile directories on disk. Without this
|
|
patch, the dispatcher's profile-exists guard (PR #20105) routes
|
|
those tasks into ``skipped_nonspawnable`` instead of spawning, which
|
|
would break tests that assert spawn behavior.
|
|
"""
|
|
from hermes_cli import profiles
|
|
monkeypatch.setattr(profiles, "profile_exists", lambda name: True)
|
|
|
|
|
|
@pytest.fixture(autouse=True)
|
|
def _suppress_concurrent_hermes_gate(request, monkeypatch):
|
|
"""Default ``_detect_concurrent_hermes_instances`` to ``[]`` for every test.
|
|
|
|
The Windows update path now refuses to proceed when another
|
|
``hermes.exe`` is detected (issue #26670). On a developer's Windows
|
|
machine running the test suite via ``hermes`` itself, this would
|
|
flag the running agent as a concurrent instance and abort every
|
|
``cmd_update`` test. Tests that want to exercise the gate explicitly
|
|
re-patch ``_detect_concurrent_hermes_instances`` with their own
|
|
return value — autouse here gives a clean default without touching
|
|
the rest of the suite.
|
|
|
|
Tests that need to call the REAL function (e.g. unit tests for the
|
|
helper itself) opt out with ``@pytest.mark.real_concurrent_gate``.
|
|
"""
|
|
if request.node.get_closest_marker("real_concurrent_gate"):
|
|
return
|
|
try:
|
|
from hermes_cli import main as _cli_main
|
|
except Exception:
|
|
return
|
|
# raising=False: under pytest's per-test spawn isolation, a concurrent
|
|
# xdist worker importing a module that transitively touches hermes_cli.main
|
|
# can briefly expose a partially-initialized module object here — one where
|
|
# _detect_concurrent_hermes_instances isn't defined yet. A bare setattr
|
|
# would raise AttributeError and error the (unrelated) test. The attribute
|
|
# always exists once main.py finishes importing, so a no-op when it's
|
|
# transiently absent is the correct, race-free default.
|
|
monkeypatch.setattr(
|
|
_cli_main,
|
|
"_detect_concurrent_hermes_instances",
|
|
lambda *_a, **_k: [],
|
|
raising=False,
|
|
)
|
|
|
|
|
|
@pytest.fixture(autouse=True)
|
|
def _inline_post_swap_handoff(request, monkeypatch):
|
|
"""Run the post-swap tail in-process instead of re-executing ``hermes update --post-swap``.
|
|
|
|
``_apply_pulled_update`` / ``_update_via_zip`` hand the rest of the run to a child
|
|
interpreter on the pulled tree. A mocked updater flow must not spawn that child (it would
|
|
run a real dependency sync against the worktree), so the tail runs here through the same
|
|
payload round-trip — every step stays patchable and the payload shape is still exercised.
|
|
Tests of the hand-off itself opt out with ``@pytest.mark.real_post_swap_handoff``.
|
|
"""
|
|
if request.node.get_closest_marker("real_post_swap_handoff"):
|
|
return
|
|
try:
|
|
from hermes_cli import update_cmd, update_receipt
|
|
except Exception:
|
|
return
|
|
|
|
def _inline(args, **payload_kwargs):
|
|
payload = update_cmd._post_swap_payload(**payload_kwargs)
|
|
if payload["receipt"]:
|
|
update_receipt.resume_update_receipt(payload["receipt"])
|
|
update_cmd._execute_post_swap(payload, args, payload_kwargs["gateway_mode"])
|
|
|
|
monkeypatch.setattr(update_cmd, "_hand_off_post_swap", _inline, raising=False)
|
|
|
|
|
|
@pytest.fixture
|
|
def isolated_update_runtime(monkeypatch, tmp_path, request):
|
|
"""Keep mocked updater flows off the host checkout and runtime fleet."""
|
|
from hermes_cli import gateway, main, update_cmd, update_cmd_fleet
|
|
from hermes_cli import update_inventory, update_receipt
|
|
|
|
checkout = tmp_path / "isolated-update-checkout"
|
|
(checkout / ".git").mkdir(parents=True)
|
|
(checkout / "apps" / "desktop").mkdir(parents=True)
|
|
monkeypatch.setattr(main, "PROJECT_ROOT", checkout)
|
|
if hasattr(request.module, "PROJECT_ROOT"):
|
|
monkeypatch.setattr(request.module, "PROJECT_ROOT", checkout)
|
|
|
|
monkeypatch.setattr(gateway, "find_gateway_pids", lambda *a, **k: [])
|
|
monkeypatch.setattr(gateway, "find_profile_gateway_processes", lambda *a, **k: [])
|
|
monkeypatch.setattr(gateway, "_get_service_pids", lambda *a, **k: set())
|
|
monkeypatch.setattr(gateway, "supports_systemd_services", lambda: False)
|
|
monkeypatch.setattr(main, "_pause_windows_gateways_for_update", lambda: None)
|
|
monkeypatch.setattr(main, "_resume_windows_gateways_after_update", lambda *a, **k: None)
|
|
monkeypatch.setattr(main, "_detect_venv_python_processes", lambda: [])
|
|
monkeypatch.setattr(main, "_restore_active_tool_dependencies", lambda *a, **k: None)
|
|
monkeypatch.setattr(update_cmd, "_clear_windows_venv_holders_or_exit", lambda *a, **k: None)
|
|
monkeypatch.setattr(update_cmd, "_finish_dashboard_update_cleanup", lambda *a, **k: None)
|
|
monkeypatch.setattr(update_cmd, "_apply_pending_fleet_restart_catchup", lambda *a, **k: None)
|
|
monkeypatch.setattr(update_cmd_fleet, "_restart_macos_launchd_gateways", lambda *a, **k: None)
|
|
monkeypatch.setattr(update_inventory, "collect_runtime_inventory", lambda: None)
|
|
monkeypatch.setattr(update_receipt, "collect_fleet_versions", lambda *a, **k: [])
|
|
|
|
|
|
# ---- prompt_toolkit / capsys isolation ----
|
|
# ``cli._cprint`` renders through ``prompt_toolkit.print_formatted_text``,
|
|
# which — when called with no explicit ``output=`` — lazily creates an
|
|
# ``Output`` from ``sys.stdout`` **and caches it on the process-global default
|
|
# ``AppSession``** (``prompt_toolkit.application.current._current_app_session``,
|
|
# a ``ContextVar`` with a module-level default). The cache is keyed to nothing
|
|
# and never re-reads ``sys.stdout``.
|
|
#
|
|
# Under pytest, ``capsys`` swaps ``sys.stdout`` for a fresh buffer per test.
|
|
# So the first CLI test that emits through ``_cprint`` (e.g. one exercising
|
|
# ``/queue``, which prints a "Queued: …" line) locks prompt_toolkit's cached
|
|
# output onto *its* captured stdout. Every later ``capsys`` test that asserts
|
|
# on ``_cprint`` output then reads an empty buffer, because the render went to
|
|
# the first test's now-dead capture target. That is the mechanism behind the
|
|
# order-dependent ``test_resume_quiet_stderr`` failure: it passes in isolation
|
|
# and in its own file, but fails in a full ``tests/cli`` run.
|
|
#
|
|
# Reset the cached output before every CLI test so each one re-creates a fresh
|
|
# prompt_toolkit ``Output`` bound to its own ``sys.stdout`` on first use. This
|
|
# is a no-op when prompt_toolkit isn't importable and cheap otherwise (the
|
|
# property re-creates lazily).
|
|
|
|
|
|
@pytest.fixture(autouse=True)
|
|
def _reset_prompt_toolkit_output_cache():
|
|
"""Clear prompt_toolkit's cached AppSession output around each CLI test.
|
|
|
|
See the module docstring for the capsys/prompt_toolkit interaction this
|
|
guards against.
|
|
"""
|
|
|
|
def _clear() -> None:
|
|
try:
|
|
from prompt_toolkit.application.current import get_app_session
|
|
|
|
get_app_session()._output = None
|
|
except Exception:
|
|
# prompt_toolkit not importable / internal shape changed — the
|
|
# tests that rely on this simply keep their prior behavior.
|
|
pass
|
|
|
|
_clear()
|
|
yield
|
|
_clear()
|