runtime: resume review after CI watcher completion

This commit is contained in:
ejclaw
2026-04-13 10:51:17 +09:00
parent 6e4a1a0d82
commit df04ef25eb
4 changed files with 172 additions and 9 deletions

View File

@@ -137,6 +137,7 @@ vi.mock('./db.js', () => {
getOpenWorkItemForChat: vi.fn((chatJid: string) =>
getOpenWorkItem(chatJid),
),
hasActiveCiWatcherForChat: vi.fn(() => false),
getLatestOpenPairedTaskForChat: vi.fn(() => undefined),
getPairedTaskById: vi.fn(() => undefined),
getPairedTurnOutputs: vi.fn(() => []),
@@ -1878,6 +1879,112 @@ If your first line is DONE_WITH_CONCERNS, the system will reopen review instead
);
});
it('defers reviewer enqueue after owner delivery when a CI watcher is still active', async () => {
const chatJid = 'group@test';
const group = makeGroup('codex');
const ownerChannel: Channel = {
...makeChannel(chatJid),
isOwnMessage: vi.fn((msg) => msg.sender === 'owner-bot@test'),
};
const reviewerChannel = makeChannel(chatJid, 'discord-review', false);
const enqueueMessageCheck = vi.fn();
const pairedTask = {
id: 'task-owner-delivery-watcher-deferred',
chat_jid: chatJid,
group_folder: group.folder,
owner_service_id: 'claude',
reviewer_service_id: 'codex-main',
title: null,
source_ref: 'HEAD',
plan_notes: null,
review_requested_at: null,
round_trip_count: 0,
status: 'active',
arbiter_verdict: null,
arbiter_requested_at: null,
completion_reason: null,
created_at: '2026-03-30T00:00:00.000Z',
updated_at: '2026-03-30T00:00:00.000Z',
} as any;
vi.mocked(serviceRouting.hasReviewerLease).mockReturnValue(true);
vi.mocked(db.hasActiveCiWatcherForChat).mockReturnValue(true);
vi.mocked(db.getLatestOpenPairedTaskForChat).mockImplementation(
() => pairedTask,
);
vi.mocked(db.getMessagesSince).mockReturnValue([
{
id: 'human-1',
chat_jid: chatJid,
sender: 'user@test',
sender_name: 'User',
content: '이 구현 진행해줘',
timestamp: '2026-03-30T00:00:00.000Z',
seq: 1,
is_bot_message: false,
} as any,
]);
vi.mocked(agentRunner.runAgentProcess).mockImplementation(
async (_group, _input, _onProcess, onOutput) => {
pairedTask.status = 'review_ready';
pairedTask.review_requested_at = '2026-03-30T00:00:01.000Z';
pairedTask.round_trip_count = 1;
await onOutput?.({
status: 'success',
phase: 'final',
result: 'DONE\nowner complete',
newSessionId: 'session-owner-delivery-watcher-deferred',
});
return {
status: 'success',
result: 'DONE\nowner complete',
newSessionId: 'session-owner-delivery-watcher-deferred',
};
},
);
const runtime = createMessageRuntime({
assistantName: 'Andy',
idleTimeout: 1_000,
pollInterval: 1_000,
timezone: 'UTC',
triggerPattern: /^@Andy\b/i,
channels: [ownerChannel, reviewerChannel],
queue: {
registerProcess: vi.fn(),
closeStdin: vi.fn(),
notifyIdle: vi.fn(),
enqueueMessageCheck,
} as any,
getRoomBindings: () => ({ [chatJid]: group }),
getSessions: () => ({}),
getLastTimestamp: () => '',
setLastTimestamp: vi.fn(),
getLastAgentTimestamps: () => ({}),
saveState: vi.fn(),
persistSession: vi.fn(),
clearSession: vi.fn(),
});
const result = await runtime.processGroupMessages(chatJid, {
runId: 'run-owner-delivery-watcher-deferred',
reason: 'messages',
});
expect(result).toBe(true);
expect(enqueueMessageCheck).not.toHaveBeenCalled();
expect(logger.info).toHaveBeenCalledWith(
expect.objectContaining({
chatJid,
runId: 'run-owner-delivery-watcher-deferred',
completedRole: 'owner',
taskId: 'task-owner-delivery-watcher-deferred',
taskStatus: 'review_ready',
}),
'Deferred paired follow-up after successful owner delivery because CI watcher is still active',
);
});
it('skips a stale reviewer follow-up after owner delivery when the latest persisted turn already belongs to the reviewer', async () => {
const chatJid = 'group@test';
const group = makeGroup('codex');

View File

@@ -1,5 +1,6 @@
import {
createProducedWorkItem,
hasActiveCiWatcherForChat,
getOpenWorkItemForChat,
getMessagesSinceSeq,
getLatestOpenPairedTaskForChat,
@@ -217,6 +218,23 @@ export function createMessageRuntime(deps: MessageRuntimeDeps): {
return;
}
const pendingTaskAfterDelivery = getLatestOpenPairedTaskForChat(chatJid);
if (
deliveryRole === 'owner' &&
pendingTaskAfterDelivery?.status === 'review_ready' &&
hasActiveCiWatcherForChat(chatJid)
) {
logger.info(
{
chatJid,
runId,
completedRole: deliveryRole,
taskId: pendingTaskAfterDelivery.id,
taskStatus: pendingTaskAfterDelivery.status,
},
'Deferred paired follow-up after successful owner delivery because CI watcher is still active',
);
return;
}
const followUpResult = enqueuePairedFollowUpAfterEvent({
chatJid,
runId,

View File

@@ -142,15 +142,6 @@ function maybeAutoTriggerReviewerAfterOwnerCompletion(args: {
logMessage: string;
}): void {
const { task, taskId, now, logMessage } = args;
if (hasActiveCiWatcherForChat(task.chat_jid)) {
logger.info(
{ taskId, chatJid: task.chat_jid },
'Active CI watcher found, deferring auto-review until watcher completes',
);
return;
}
if (task.round_trip_count >= PAIRED_MAX_ROUND_TRIPS) {
logger.info(
{
@@ -177,6 +168,17 @@ function maybeAutoTriggerReviewerAfterOwnerCompletion(args: {
round_trip_count: task.round_trip_count + 1,
},
});
if (hasActiveCiWatcherForChat(task.chat_jid)) {
logger.info(
{
taskId,
chatJid: task.chat_jid,
roundTrip: task.round_trip_count + 1,
},
'Active CI watcher found, marked task review_ready and deferred reviewer enqueue until watcher completes',
);
return;
}
logger.info({ taskId, roundTrip: task.round_trip_count + 1 }, logMessage);
}
}

View File

@@ -757,6 +757,42 @@ describe('paired execution context', () => {
);
});
it('marks review_ready but defers reviewer enqueue when an active CI watcher exists', () => {
vi.mocked(db.getPairedTaskById).mockReturnValue(
buildPairedTask({
status: 'active',
round_trip_count: 0,
}),
);
vi.mocked(db.hasActiveCiWatcherForChat).mockReturnValue(true);
vi.mocked(pairedWorkspaceManager.markPairedTaskReviewReady).mockReturnValue(
{
ownerWorkspace: buildWorkspace('owner', '/tmp/paired/task-1/owner'),
reviewerWorkspace: buildWorkspace(
'reviewer',
'/tmp/paired/task-1/reviewer',
),
},
);
completePairedExecutionContext({
taskId: 'task-1',
role: 'owner',
status: 'succeeded',
summary: 'DONE',
});
expect(
pairedWorkspaceManager.markPairedTaskReviewReady,
).toHaveBeenCalledWith('task-1');
expect(db.updatePairedTask).toHaveBeenCalledWith(
'task-1',
expect.objectContaining({
round_trip_count: 1,
}),
);
});
it('requests arbiter instead of re-reviewing when repeated DONE finalize loops exceed the threshold', () => {
vi.spyOn(config, 'isArbiterEnabled').mockReturnValue(true);