Files
roboco/tests/unit/runtime/test_war_room_dispatch.py
a36a180b73 fix(orchestrator): let a failed Board Program exploration retry (#706)
`_board_dispatched` is an in-memory, never-expiring set of
(agent_slug, task_id). The solo exploration dispatchers consulted it, so an
exploration whose propose verb rejected was never retried for the life of the
process — Periscope, Sentinel, Scales and Barfly each spawned once on
2026-07-25, failed, and sat PENDING until the stack restarted.

The guard was written for the two-reviewer board REVIEW pass, where it is
correct: a reviewer has no verb to advance the task, so a respawn can only
loop. An explorer is the opposite — `propose_*` is exactly such a verb, so a
respawn can and should advance it.

Drop it from the 15 `_dispatch_*_exploration` functions. Bounding falls to
`_pm_respawn_should_gate`, which is what actually bounds a loop: DB-persisted,
reset by a status change, and cooled down so a deploy that fixes the cause
lets the work resume. `_dispatch_board_reviewer` keeps the set (its
`_board_review_complete` reads it as a has-run signal) as does vault
curation (same-process race guard behind its own durable marker).

The 14 `*_dispatch_is_one_shot` tests asserted the old contract on the false
rationale that board roles have no progression verb; they now assert a second
tick re-attempts.

Co-authored-by: Renn F <rennf93@users.noreply.github.com>
2026-07-26 15:52:33 +02:00

171 lines
6.1 KiB
Python

"""War Room campaign-planning dispatch — Head-of-Marketing-solo,
EVENT-triggered, never the two-reviewer board-review gate, never the dev/PM
delivery dispatchers. Mirrors test_coroner_dispatch.py: complete-at-propose
(no "already authored" marker pre-check — propose_campaign completes the
task atomically), and no extra async context-gathering helper (the release
brief, or {} for a blank on-demand cycle, is already ON the task's own
``war_room_brief`` marker at origination time — no separate DB read needed).
"""
from __future__ import annotations
from typing import Any, cast
from unittest.mock import AsyncMock, MagicMock, patch
from uuid import uuid4
import pytest
from roboco.runtime.orchestrator import AgentOrchestrator
from roboco.services.task import WAR_ROOM_SOURCE
def _make_orch() -> AgentOrchestrator:
orch = AgentOrchestrator.__new__(AgentOrchestrator)
cast("Any", orch)._pm_respawn_tracker = {}
cast("Any", orch)._schedule_respawn_persist = lambda *_a, **_k: None
orch._instances = {}
orch._board_dispatched = set()
return orch
def _war_room_task(
*, orchestration_markers: dict[str, Any] | None = None
) -> dict[str, Any]:
return {
"id": str(uuid4()),
"status": "pending",
"team": "board",
"title": "War Room campaign planning",
"description": "Design one campaign and propose it.",
"assigned_to": "head-marketing",
"source": WAR_ROOM_SOURCE,
"orchestration_markers": orchestration_markers
if orchestration_markers is not None
else {
"war_room_brief": {
"version": "0.30.0",
"highlights": ["MegaTask v2", "Findings ledger"],
}
},
}
@pytest.mark.asyncio
async def test_war_room_dispatch_spawns_only_hom() -> None:
orch = _make_orch()
task = _war_room_task()
with (
patch.object(orch, "_is_agent_active", return_value=False),
patch.object(orch, "_task_git_context", return_value=None),
patch.object(orch, "spawn_agent", new=AsyncMock()) as spawn,
):
await orch._dispatch_war_room_exploration(task)
spawn.assert_awaited_once()
calls = list(spawn.await_args_list)
assert calls[0].kwargs["agent_id"] == "head-marketing"
assert calls[0].kwargs["task_id"] == task["id"]
@pytest.mark.asyncio
async def test_war_room_dispatch_retries_until_breaker() -> None:
"""A failed exploration must be retried, not abandoned.
The explorer has a progression verb (``propose_*``), so a respawn CAN
advance the task — unlike the two-reviewer review pass this guard was
originally written for. Bounding belongs to
``_pm_respawn_should_gate`` (DB-persisted, reset by a status change),
not to a never-expiring in-memory set.
"""
orch = _make_orch()
task = _war_room_task()
with (
patch.object(orch, "_is_agent_active", return_value=False),
patch.object(orch, "_task_git_context", return_value=None),
patch.object(orch, "spawn_agent", new=AsyncMock()) as spawn,
):
await orch._dispatch_war_room_exploration(task)
await orch._dispatch_war_room_exploration(task)
ticks = 2
assert spawn.await_count == ticks, (
"a second tick must re-attempt a failed exploration"
)
@pytest.mark.asyncio
async def test_war_room_dispatch_skips_active_hom() -> None:
orch = _make_orch()
task = _war_room_task()
with (
patch.object(orch, "_is_agent_active", return_value=True),
patch.object(orch, "spawn_agent", new=AsyncMock()) as spawn,
):
await orch._dispatch_war_room_exploration(task)
spawn.assert_not_awaited()
@pytest.mark.asyncio
async def test_dispatch_pm_work_routes_war_room_away_from_board_handler() -> None:
"""A board_war_room task must ride the dedicated dispatcher, never the
two-reviewer ``_handle_board_assigned_task`` or the generic PM-assigned
handler."""
task = _war_room_task()
stub = MagicMock()
stub._fetch_tasks = AsyncMock(return_value=[task])
stub._is_task_handled_this_tick = MagicMock(return_value=False)
stub._resolve_agent_slug = MagicMock(return_value="head-marketing")
stub._BOARD_AGENTS = frozenset({"product-owner", "head-marketing"})
stub._dispatch_war_room_exploration = AsyncMock()
stub._handle_board_assigned_task = AsyncMock()
stub._handle_pm_assigned_task = AsyncMock()
stub._route_unassigned_pm_task = AsyncMock()
client: Any = MagicMock()
await AgentOrchestrator._dispatch_pm_work(cast("AgentOrchestrator", stub), client)
stub._dispatch_war_room_exploration.assert_awaited_once()
stub._handle_board_assigned_task.assert_not_awaited()
stub._handle_pm_assigned_task.assert_not_awaited()
@pytest.mark.asyncio
async def test_war_room_tasks_are_never_routed_by_dev_dispatch() -> None:
tasks = [_war_room_task()]
stub = MagicMock()
stub._fetch_tasks = AsyncMock(return_value=tasks)
stub._is_task_handled_this_tick = MagicMock(return_value=False)
stub._dev_dispatch_one = AsyncMock()
client: Any = MagicMock()
await AgentOrchestrator._dispatch_dev_work(cast("AgentOrchestrator", stub), client)
stub._dev_dispatch_one.assert_not_awaited()
def test_war_room_prompt_names_solo_hom_and_real_verbs() -> None:
orch = _make_orch()
task = _war_room_task()
prompt = orch._build_war_room_prompt(task)
assert "triage()" in prompt
assert "propose_campaign(" in prompt
assert "i_am_idle()" in prompt
assert "v0.30.0" in prompt
assert "MegaTask v2" in prompt
def test_war_room_prompt_renders_blank_brief_message_when_absent() -> None:
orch = _make_orch()
task = _war_room_task(orchestration_markers={"war_room_brief": {}})
prompt = orch._build_war_room_prompt(task)
assert "No release triggered this cycle" in prompt
def test_war_room_prompt_renders_blank_brief_message_with_no_marker() -> None:
"""A malformed/absent marker degrades to the same blank-brief branch,
never a KeyError."""
orch = _make_orch()
task = _war_room_task(orchestration_markers={})
prompt = orch._build_war_room_prompt(task)
assert "No release triggered this cycle" in prompt