Files
deer-flow/backend/tests/test_subagent_executor.py
DanielWalnut 76803b826f refactor: split backend into harness (deerflow.*) and app (app.*) (#1131)
* refactor: extract shared utils to break harness→app cross-layer imports

Move _validate_skill_frontmatter to src/skills/validation.py and
CONVERTIBLE_EXTENSIONS + convert_file_to_markdown to src/utils/file_conversion.py.
This eliminates the two reverse dependencies from client.py (harness layer)
into gateway/routers/ (app layer), preparing for the harness/app package split.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* refactor: split backend/src into harness (deerflow.*) and app (app.*)

Physically split the monolithic backend/src/ package into two layers:

- **Harness** (`packages/harness/deerflow/`): publishable agent framework
  package with import prefix `deerflow.*`. Contains agents, sandbox, tools,
  models, MCP, skills, config, and all core infrastructure.

- **App** (`app/`): unpublished application code with import prefix `app.*`.
  Contains gateway (FastAPI REST API) and channels (IM integrations).

Key changes:
- Move 13 harness modules to packages/harness/deerflow/ via git mv
- Move gateway + channels to app/ via git mv
- Rename all imports: src.* → deerflow.* (harness) / app.* (app layer)
- Set up uv workspace with deerflow-harness as workspace member
- Update langgraph.json, config.example.yaml, all scripts, Docker files
- Add build-system (hatchling) to harness pyproject.toml
- Add PYTHONPATH=. to gateway startup commands for app.* resolution
- Update ruff.toml with known-first-party for import sorting
- Update all documentation to reflect new directory structure

Boundary rule enforced: harness code never imports from app.
All 429 tests pass. Lint clean.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* chore: add harness→app boundary check test and update docs

Add test_harness_boundary.py that scans all Python files in
packages/harness/deerflow/ and fails if any `from app.*` or
`import app.*` statement is found. This enforces the architectural
rule that the harness layer never depends on the app layer.

Update CLAUDE.md to document the harness/app split architecture,
import conventions, and the boundary enforcement test.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* feat: add config versioning with auto-upgrade on startup

When config.example.yaml schema changes, developers' local config.yaml
files can silently become outdated. This adds a config_version field and
auto-upgrade mechanism so breaking changes (like src.* → deerflow.*
renames) are applied automatically before services start.

- Add config_version: 1 to config.example.yaml
- Add startup version check warning in AppConfig.from_file()
- Add scripts/config-upgrade.sh with migration registry for value replacements
- Add `make config-upgrade` target
- Auto-run config-upgrade in serve.sh and start-daemon.sh before starting services
- Add config error hints in service failure messages

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix comments

* fix: update src.* import in test_sandbox_tools_security to deerflow.*

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix: handle empty config and search parent dirs for config.example.yaml

Address Copilot review comments on PR #1131:
- Guard against yaml.safe_load() returning None for empty config files
- Search parent directories for config.example.yaml instead of only
  looking next to config.yaml, fixing detection in common setups

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix: correct skills root path depth and config_version type coercion

- loader.py: fix get_skills_root_path() to use 5 parent levels (was 3)
  after harness split, file lives at packages/harness/deerflow/skills/
  so parent×3 resolved to backend/packages/harness/ instead of backend/
- app_config.py: coerce config_version to int() before comparison in
  _check_config_version() to prevent TypeError when YAML stores value
  as string (e.g. config_version: "1")
- tests: add regression tests for both fixes

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix: update test imports from src.* to deerflow.*/app.* after harness refactor

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-14 22:55:52 +08:00

774 lines
26 KiB
Python

"""Tests for subagent executor async/sync execution paths.
Covers:
- SubagentExecutor.execute() synchronous execution path
- SubagentExecutor._aexecute() asynchronous execution path
- asyncio.run() properly executes async workflow within thread pool context
- Error handling in both sync and async paths
- Async tool support (MCP tools)
Note: Due to circular import issues in the main codebase, conftest.py mocks
deerflow.subagents.executor. This test file uses delayed import via fixture to test
the real implementation in isolation.
"""
import asyncio
import sys
from datetime import datetime
from unittest.mock import MagicMock, patch
import pytest
# Module names that need to be mocked to break circular imports
_MOCKED_MODULE_NAMES = [
"deerflow.agents",
"deerflow.agents.thread_state",
"deerflow.agents.middlewares",
"deerflow.agents.middlewares.thread_data_middleware",
"deerflow.sandbox",
"deerflow.sandbox.middleware",
"deerflow.models",
]
@pytest.fixture(scope="session", autouse=True)
def _setup_executor_classes():
"""Set up mocked modules and import real executor classes.
This fixture runs once per session and yields the executor classes.
It handles module cleanup to avoid affecting other test files.
"""
# Save original modules
original_modules = {name: sys.modules.get(name) for name in _MOCKED_MODULE_NAMES}
original_executor = sys.modules.get("deerflow.subagents.executor")
# Remove mocked executor if exists (from conftest.py)
if "deerflow.subagents.executor" in sys.modules:
del sys.modules["deerflow.subagents.executor"]
# Set up mocks
for name in _MOCKED_MODULE_NAMES:
sys.modules[name] = MagicMock()
# Import real classes inside fixture
from langchain_core.messages import AIMessage, HumanMessage
from deerflow.subagents.config import SubagentConfig
from deerflow.subagents.executor import (
SubagentExecutor,
SubagentResult,
SubagentStatus,
)
# Store classes in a dict to yield
classes = {
"AIMessage": AIMessage,
"HumanMessage": HumanMessage,
"SubagentConfig": SubagentConfig,
"SubagentExecutor": SubagentExecutor,
"SubagentResult": SubagentResult,
"SubagentStatus": SubagentStatus,
}
yield classes
# Cleanup: Restore original modules
for name in _MOCKED_MODULE_NAMES:
if original_modules[name] is not None:
sys.modules[name] = original_modules[name]
elif name in sys.modules:
del sys.modules[name]
# Restore executor module (conftest.py mock)
if original_executor is not None:
sys.modules["deerflow.subagents.executor"] = original_executor
elif "deerflow.subagents.executor" in sys.modules:
del sys.modules["deerflow.subagents.executor"]
# Helper classes that wrap real classes for testing
class MockHumanMessage:
"""Mock HumanMessage for testing - wraps real class from fixture."""
def __init__(self, content, _classes=None):
self._content = content
self._classes = _classes
def _get_real(self):
return self._classes["HumanMessage"](content=self._content)
class MockAIMessage:
"""Mock AIMessage for testing - wraps real class from fixture."""
def __init__(self, content, msg_id=None, _classes=None):
self._content = content
self._msg_id = msg_id
self._classes = _classes
def _get_real(self):
msg = self._classes["AIMessage"](content=self._content)
if self._msg_id:
msg.id = self._msg_id
return msg
async def async_iterator(items):
"""Helper to create an async iterator from a list."""
for item in items:
yield item
# -----------------------------------------------------------------------------
# Fixtures
# -----------------------------------------------------------------------------
@pytest.fixture
def classes(_setup_executor_classes):
"""Provide access to executor classes."""
return _setup_executor_classes
@pytest.fixture
def base_config(classes):
"""Return a basic subagent config for testing."""
return classes["SubagentConfig"](
name="test-agent",
description="Test agent",
system_prompt="You are a test agent.",
max_turns=10,
timeout_seconds=60,
)
@pytest.fixture
def mock_agent():
"""Return a properly configured mock agent with async stream."""
agent = MagicMock()
agent.astream = MagicMock()
return agent
# Helper to create real message objects
class _MsgHelper:
"""Helper to create real message objects from fixture classes."""
def __init__(self, classes):
self.classes = classes
def human(self, content):
return self.classes["HumanMessage"](content=content)
def ai(self, content, msg_id=None):
msg = self.classes["AIMessage"](content=content)
if msg_id:
msg.id = msg_id
return msg
@pytest.fixture
def msg(classes):
"""Provide message factory."""
return _MsgHelper(classes)
# -----------------------------------------------------------------------------
# Async Execution Path Tests
# -----------------------------------------------------------------------------
class TestAsyncExecutionPath:
"""Test _aexecute() async execution path."""
@pytest.mark.anyio
async def test_aexecute_success(self, classes, base_config, mock_agent, msg):
"""Test successful async execution returns completed result."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
final_message = msg.ai("Task completed successfully", "msg-1")
final_state = {
"messages": [
msg.human("Do something"),
final_message,
]
}
mock_agent.astream = lambda *args, **kwargs: async_iterator([final_state])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
trace_id="test-trace",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Do something")
assert result.status == SubagentStatus.COMPLETED
assert result.result == "Task completed successfully"
assert result.error is None
assert result.started_at is not None
assert result.completed_at is not None
@pytest.mark.anyio
async def test_aexecute_collects_ai_messages(self, classes, base_config, mock_agent, msg):
"""Test that AI messages are collected during streaming."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
msg1 = msg.ai("First response", "msg-1")
msg2 = msg.ai("Second response", "msg-2")
chunk1 = {"messages": [msg.human("Task"), msg1]}
chunk2 = {"messages": [msg.human("Task"), msg1, msg2]}
mock_agent.astream = lambda *args, **kwargs: async_iterator([chunk1, chunk2])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
assert result.status == SubagentStatus.COMPLETED
assert len(result.ai_messages) == 2
assert result.ai_messages[0]["id"] == "msg-1"
assert result.ai_messages[1]["id"] == "msg-2"
@pytest.mark.anyio
async def test_aexecute_handles_duplicate_messages(self, classes, base_config, mock_agent, msg):
"""Test that duplicate AI messages are not added."""
SubagentExecutor = classes["SubagentExecutor"]
msg1 = msg.ai("Response", "msg-1")
# Same message appears in multiple chunks
chunk1 = {"messages": [msg.human("Task"), msg1]}
chunk2 = {"messages": [msg.human("Task"), msg1]}
mock_agent.astream = lambda *args, **kwargs: async_iterator([chunk1, chunk2])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
assert len(result.ai_messages) == 1
@pytest.mark.anyio
async def test_aexecute_handles_list_content(self, classes, base_config, mock_agent, msg):
"""Test handling of list-type content in AIMessage."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
final_message = msg.ai([{"text": "Part 1"}, {"text": "Part 2"}])
final_state = {
"messages": [
msg.human("Task"),
final_message,
]
}
mock_agent.astream = lambda *args, **kwargs: async_iterator([final_state])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
assert result.status == SubagentStatus.COMPLETED
assert "Part 1" in result.result
assert "Part 2" in result.result
@pytest.mark.anyio
async def test_aexecute_handles_agent_exception(self, classes, base_config, mock_agent):
"""Test that exceptions during execution are caught and returned as FAILED."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
mock_agent.astream.side_effect = Exception("Agent error")
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
assert result.status == SubagentStatus.FAILED
assert "Agent error" in result.error
assert result.completed_at is not None
@pytest.mark.anyio
async def test_aexecute_no_final_state(self, classes, base_config, mock_agent):
"""Test handling when no final state is returned."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
mock_agent.astream = lambda *args, **kwargs: async_iterator([])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
assert result.status == SubagentStatus.COMPLETED
assert result.result == "No response generated"
@pytest.mark.anyio
async def test_aexecute_no_ai_message_in_state(self, classes, base_config, mock_agent, msg):
"""Test fallback when no AIMessage found in final state."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
final_state = {"messages": [msg.human("Task")]}
mock_agent.astream = lambda *args, **kwargs: async_iterator([final_state])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
# Should fallback to string representation of last message
assert result.status == SubagentStatus.COMPLETED
assert "Task" in result.result
# -----------------------------------------------------------------------------
# Sync Execution Path Tests
# -----------------------------------------------------------------------------
class TestSyncExecutionPath:
"""Test execute() synchronous execution path with asyncio.run()."""
def test_execute_runs_async_in_event_loop(self, classes, base_config, mock_agent, msg):
"""Test that execute() runs _aexecute() in a new event loop via asyncio.run()."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
final_message = msg.ai("Sync result", "msg-1")
final_state = {
"messages": [
msg.human("Task"),
final_message,
]
}
mock_agent.astream = lambda *args, **kwargs: async_iterator([final_state])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = executor.execute("Task")
assert result.status == SubagentStatus.COMPLETED
assert result.result == "Sync result"
def test_execute_in_thread_pool_context(self, classes, base_config, msg):
"""Test that execute() works correctly when called from a thread pool.
This simulates the real-world usage where execute() is called from
_execution_pool in execute_async().
"""
from concurrent.futures import ThreadPoolExecutor
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
final_message = msg.ai("Thread pool result", "msg-1")
final_state = {
"messages": [
msg.human("Task"),
final_message,
]
}
def run_in_thread():
mock_agent = MagicMock()
mock_agent.astream = lambda *args, **kwargs: async_iterator([final_state])
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
return executor.execute("Task")
# Execute in thread pool (simulating _execution_pool usage)
with ThreadPoolExecutor(max_workers=1) as pool:
future = pool.submit(run_in_thread)
result = future.result(timeout=5)
assert result.status == SubagentStatus.COMPLETED
assert result.result == "Thread pool result"
def test_execute_handles_asyncio_run_failure(self, classes, base_config):
"""Test handling when asyncio.run() itself fails."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_aexecute") as mock_aexecute:
mock_aexecute.side_effect = Exception("Asyncio run error")
result = executor.execute("Task")
assert result.status == SubagentStatus.FAILED
assert "Asyncio run error" in result.error
assert result.completed_at is not None
def test_execute_with_result_holder(self, classes, base_config, mock_agent, msg):
"""Test execute() updates provided result_holder in real-time."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
msg1 = msg.ai("Step 1", "msg-1")
chunk1 = {"messages": [msg.human("Task"), msg1]}
mock_agent.astream = lambda *args, **kwargs: async_iterator([chunk1])
# Pre-create result holder (as done in execute_async)
result_holder = SubagentResult(
task_id="predefined-id",
trace_id="test-trace",
status=SubagentStatus.RUNNING,
started_at=datetime.now(),
)
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = executor.execute("Task", result_holder=result_holder)
# Should be the same object
assert result is result_holder
assert result.task_id == "predefined-id"
assert result.status == SubagentStatus.COMPLETED
# -----------------------------------------------------------------------------
# Async Tool Support Tests (MCP Tools)
# -----------------------------------------------------------------------------
class TestAsyncToolSupport:
"""Test that async-only tools (like MCP tools) work correctly."""
@pytest.mark.anyio
async def test_async_tool_called_in_astream(self, classes, base_config, msg):
"""Test that async tools are properly awaited in astream.
This verifies the fix for: async MCP tools not being executed properly
because they were being called synchronously.
"""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
async_tool_calls = []
async def mock_async_tool(*args, **kwargs):
async_tool_calls.append("called")
await asyncio.sleep(0.01) # Simulate async work
return {"result": "async tool result"}
mock_agent = MagicMock()
# Simulate agent that calls async tools during streaming
async def mock_astream(*args, **kwargs):
await mock_async_tool()
yield {
"messages": [
msg.human("Task"),
msg.ai("Done", "msg-1"),
]
}
mock_agent.astream = mock_astream
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = await executor._aexecute("Task")
assert len(async_tool_calls) == 1
assert result.status == SubagentStatus.COMPLETED
def test_sync_execute_with_async_tools(self, classes, base_config, msg):
"""Test that sync execute() properly runs async tools via asyncio.run()."""
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
async_tool_calls = []
async def mock_async_tool():
async_tool_calls.append("called")
await asyncio.sleep(0.01)
return {"result": "async result"}
mock_agent = MagicMock()
async def mock_astream(*args, **kwargs):
await mock_async_tool()
yield {
"messages": [
msg.human("Task"),
msg.ai("Done", "msg-1"),
]
}
mock_agent.astream = mock_astream
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id="test-thread",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
result = executor.execute("Task")
assert len(async_tool_calls) == 1
assert result.status == SubagentStatus.COMPLETED
# -----------------------------------------------------------------------------
# Thread Safety Tests
# -----------------------------------------------------------------------------
class TestThreadSafety:
"""Test thread safety of executor operations."""
def test_multiple_executors_in_parallel(self, classes, base_config, msg):
"""Test multiple executors running in parallel via thread pool."""
from concurrent.futures import ThreadPoolExecutor, as_completed
SubagentExecutor = classes["SubagentExecutor"]
SubagentStatus = classes["SubagentStatus"]
results = []
def execute_task(task_id: int):
def make_astream(*args, **kwargs):
return async_iterator(
[
{
"messages": [
msg.human(f"Task {task_id}"),
msg.ai(f"Result {task_id}", f"msg-{task_id}"),
]
}
]
)
mock_agent = MagicMock()
mock_agent.astream = make_astream
executor = SubagentExecutor(
config=base_config,
tools=[],
thread_id=f"thread-{task_id}",
)
with patch.object(executor, "_create_agent", return_value=mock_agent):
return executor.execute(f"Task {task_id}")
# Execute multiple tasks in parallel
with ThreadPoolExecutor(max_workers=3) as pool:
futures = [pool.submit(execute_task, i) for i in range(5)]
for future in as_completed(futures):
results.append(future.result())
assert len(results) == 5
for result in results:
assert result.status == SubagentStatus.COMPLETED
assert "Result" in result.result
# -----------------------------------------------------------------------------
# Cleanup Background Task Tests
# -----------------------------------------------------------------------------
class TestCleanupBackgroundTask:
"""Test cleanup_background_task function for race condition prevention."""
@pytest.fixture
def executor_module(self, _setup_executor_classes):
"""Import the executor module with real classes."""
# Re-import to get the real module with cleanup_background_task
import importlib
from deerflow.subagents import executor
return importlib.reload(executor)
def test_cleanup_removes_terminal_completed_task(self, executor_module, classes):
"""Test that cleanup removes a COMPLETED task."""
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
# Add a completed task
task_id = "test-completed-task"
result = SubagentResult(
task_id=task_id,
trace_id="test-trace",
status=SubagentStatus.COMPLETED,
result="done",
completed_at=datetime.now(),
)
executor_module._background_tasks[task_id] = result
# Cleanup should remove it
executor_module.cleanup_background_task(task_id)
assert task_id not in executor_module._background_tasks
def test_cleanup_removes_terminal_failed_task(self, executor_module, classes):
"""Test that cleanup removes a FAILED task."""
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
task_id = "test-failed-task"
result = SubagentResult(
task_id=task_id,
trace_id="test-trace",
status=SubagentStatus.FAILED,
error="error",
completed_at=datetime.now(),
)
executor_module._background_tasks[task_id] = result
executor_module.cleanup_background_task(task_id)
assert task_id not in executor_module._background_tasks
def test_cleanup_removes_terminal_timed_out_task(self, executor_module, classes):
"""Test that cleanup removes a TIMED_OUT task."""
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
task_id = "test-timedout-task"
result = SubagentResult(
task_id=task_id,
trace_id="test-trace",
status=SubagentStatus.TIMED_OUT,
error="timeout",
completed_at=datetime.now(),
)
executor_module._background_tasks[task_id] = result
executor_module.cleanup_background_task(task_id)
assert task_id not in executor_module._background_tasks
def test_cleanup_skips_running_task(self, executor_module, classes):
"""Test that cleanup does NOT remove a RUNNING task.
This prevents race conditions where task_tool calls cleanup
while the background executor is still updating the task.
"""
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
task_id = "test-running-task"
result = SubagentResult(
task_id=task_id,
trace_id="test-trace",
status=SubagentStatus.RUNNING,
started_at=datetime.now(),
)
executor_module._background_tasks[task_id] = result
executor_module.cleanup_background_task(task_id)
# Should still be present because it's RUNNING
assert task_id in executor_module._background_tasks
def test_cleanup_skips_pending_task(self, executor_module, classes):
"""Test that cleanup does NOT remove a PENDING task."""
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
task_id = "test-pending-task"
result = SubagentResult(
task_id=task_id,
trace_id="test-trace",
status=SubagentStatus.PENDING,
)
executor_module._background_tasks[task_id] = result
executor_module.cleanup_background_task(task_id)
assert task_id in executor_module._background_tasks
def test_cleanup_handles_unknown_task_gracefully(self, executor_module):
"""Test that cleanup doesn't raise for unknown task IDs."""
# Should not raise
executor_module.cleanup_background_task("nonexistent-task")
def test_cleanup_removes_task_with_completed_at_even_if_running(self, executor_module, classes):
"""Test that cleanup removes task if completed_at is set, even if status is RUNNING.
This is a safety net: if completed_at is set, the task is considered done
regardless of status.
"""
SubagentResult = classes["SubagentResult"]
SubagentStatus = classes["SubagentStatus"]
task_id = "test-completed-at-task"
result = SubagentResult(
task_id=task_id,
trace_id="test-trace",
status=SubagentStatus.RUNNING, # Status not terminal
completed_at=datetime.now(), # But completed_at is set
)
executor_module._background_tasks[task_id] = result
executor_module.cleanup_background_task(task_id)
# Should be removed because completed_at is set
assert task_id not in executor_module._background_tasks