mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-31 19:16:29 +00:00
Second, deeper pass over tools/gateway/hermes_cli plus first pass over the trees wave 1 missed (acp, acp_adapter, skills, computer_use, docker, dashboard, conformance, monitoring, secret_sources, hermes_state, providers). Same rubric as wave 1 (AGENTS.md test policy); security, alternation/caching invariants, issue-number regressions, and E2E kept. Real test-quality fixes found and rooted out along the way: - tests/tools/test_command_guards.py made real auxiliary-LLM HTTPS calls (DEFAULT_CONFIG smart-approval leaked in) — pinned approval mode=manual via autouse fixture: 17.4s → 0.4s. - test_model_switch_custom_providers.py / test_user_providers_model_switch.py silently probed live provider catalogs (~2s/test) — stubbed cached_provider_model_ids/provider_model_ids/fetch_api_models. - test_telegram_noise_filter.py: 15-platform copy-paste matrix over shared gateway.run logic → 3 representative platforms (55s → 3.9s). - test_gateway_shutdown.py: stop()'s 5s interrupt-deadline loop spun on MagicMock agents — interrupt.side_effect now clears _running_agents (22s → 1.0s). - test_gateway_inactivity_timeout.py poll-harness timings shrunk 3-5x (24s → 1.1s); test_mcp_stability.py backoff/SIGTERM-grace sleeps patched (15.4s → 2.5s); test_async_delegation.py negative-drain wait 5s → 0.5s. - test_telegram_init_deadline.py: loop-block margin restored to 1.0s with rationale comment — the watchdog-dump assertion needs the loop blocked well past deadline+grace under parallel load (flaked once in the 40-worker verification run at a 0.2s margin). Verification: full hermetic suite via scripts/run_tests.sh — 2,438 files, 21,718 tests passed, 0 failed, 293.9s wall. Suite totals vs original baseline: 46,820 → 19,757 test functions (−57.8%), wall 583.5s → 293.9s (−50%), subprocess CPU 13,564s → 11,623s.
122 lines
4.3 KiB
Python
122 lines
4.3 KiB
Python
"""Contract tests for the container Dockerfile.
|
|
|
|
These tests assert invariants about how the Dockerfile composes its runtime —
|
|
they deliberately avoid snapshotting specific package versions, line numbers,
|
|
or exact flag choices. What they DO assert is that the Dockerfile maintains
|
|
the properties required for correct production behaviour:
|
|
|
|
- A PID-1 init is installed and wraps the entrypoint, so that orphaned
|
|
subprocesses (MCP stdio servers, git, bun, browser daemons) get reaped
|
|
instead of accumulating as zombies (#15012).
|
|
- Signal forwarding runs through the init so ``docker stop`` triggers
|
|
hermes's own graceful-shutdown path.
|
|
|
|
The init can be any reaper-capable PID-1: the historical lineage was
|
|
``tini``; the current image uses s6-overlay's ``/init`` (which execs
|
|
``s6-svscan`` as PID 1, with the same SIGCHLD-reaping property). The
|
|
checks below accept either family — the contract is behavioural, not
|
|
nominal.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
|
|
REPO_ROOT = Path(__file__).resolve().parents[2]
|
|
DOCKERFILE = REPO_ROOT / "Dockerfile"
|
|
DOCKERIGNORE = REPO_ROOT / ".dockerignore"
|
|
|
|
|
|
# Init-process families this repo accepts as PID 1. ``tini`` /
|
|
# ``dumb-init`` / ``catatonit`` are classic minimal reapers; s6-overlay
|
|
# ships ``/init`` which execs ``s6-svscan`` as PID 1 (same reaper
|
|
# contract, plus supervision of declared services). Either family
|
|
# satisfies the zombie-reaping invariant — see issue #15012.
|
|
_KNOWN_INIT_TOKENS: tuple[str, ...] = (
|
|
"tini",
|
|
"dumb-init",
|
|
"catatonit",
|
|
"s6-overlay",
|
|
"s6-svscan",
|
|
"/init",
|
|
)
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def dockerfile_text() -> str:
|
|
if not DOCKERFILE.exists():
|
|
pytest.skip("Dockerfile not present in this checkout")
|
|
return DOCKERFILE.read_text()
|
|
|
|
|
|
def _dockerfile_instructions(dockerfile_text: str) -> list[str]:
|
|
instructions: list[str] = []
|
|
current = ""
|
|
|
|
for raw_line in dockerfile_text.splitlines():
|
|
line = raw_line.strip()
|
|
if not line or line.startswith("#"):
|
|
continue
|
|
|
|
continued = line.removesuffix("\\").strip()
|
|
current = f"{current} {continued}".strip()
|
|
if not line.endswith("\\"):
|
|
instructions.append(current)
|
|
current = ""
|
|
|
|
return instructions
|
|
|
|
|
|
def _run_steps(dockerfile_text: str) -> list[str]:
|
|
return [
|
|
instruction
|
|
for instruction in _dockerfile_instructions(dockerfile_text)
|
|
if instruction.startswith("RUN ")
|
|
]
|
|
|
|
|
|
def _instruction_text(dockerfile_text: str) -> str:
|
|
"""Join every non-comment Dockerfile instruction into one searchable
|
|
string. Crucially excludes comments — otherwise the historical
|
|
explanation of "we used to use tini" would silently satisfy a
|
|
substring check long after tini was removed from the build.
|
|
"""
|
|
return "\n".join(_dockerfile_instructions(dockerfile_text))
|
|
|
|
|
|
def test_dockerfile_installs_an_init_for_zombie_reaping(dockerfile_text):
|
|
"""Some init (tini, dumb-init, catatonit, s6-overlay) must be installed.
|
|
|
|
Without a PID-1 init that handles SIGCHLD, hermes accumulates zombie
|
|
processes from MCP stdio subprocesses, git operations, browser
|
|
daemons, etc. In long-running Docker deployments this eventually
|
|
exhausts the PID table.
|
|
"""
|
|
# Accept any of the common reapers. The contract is behavioural:
|
|
# something must be installed that reaps orphans.
|
|
#
|
|
# Scan instructions only (no comments) so a stale historical mention
|
|
# in a comment can't masquerade as a current install. Without this,
|
|
# removing tini from the actual build but leaving the word in a
|
|
# comment would silently keep the test green.
|
|
instructions = _instruction_text(dockerfile_text)
|
|
installed = any(name in instructions for name in _KNOWN_INIT_TOKENS)
|
|
assert installed, (
|
|
"No PID-1 init detected in Dockerfile instructions (looked for: "
|
|
f"{', '.join(_KNOWN_INIT_TOKENS)}). Without an init process to "
|
|
"reap orphaned subprocesses, hermes accumulates zombies in Docker "
|
|
"deployments. See issue #15012."
|
|
)
|
|
|
|
|
|
def test_dockerignore_excludes_nested_dependency_dirs():
|
|
if not DOCKERIGNORE.exists():
|
|
pytest.skip(".dockerignore not present in this checkout")
|
|
|
|
text = DOCKERIGNORE.read_text()
|
|
|
|
assert "**/node_modules" in text
|
|
assert "**/.venv" in text
|