Files
AgentCompany/tests/unit/tasks-phase.test.ts
T
ZacharyZhang-NY 54a97f67ac B2 Phase 1 follow-up: plug lease orphan gap in advanceTaskStatus + scope interruptStaleRuns cleanup
Linus review T1 surfaced three real bugs in the B2 Phase 1 shadow
dual-write surface. Fix them in one pass.

1. advanceTaskStatus left task_leases behind on terminal (done /
   cancelled / blocked) and "return to queue" (todo / backlog)
   transitions. The status->phase trigger advanced tasks.phase but the
   lease row lingered as an orphan. dev-mode verifyPhaseConsistency
   could see the divergence; production silently drifted. Wrap the
   status UPDATE and an explicit DELETE FROM task_leases in a single
   transaction so no caller observes the intermediate state.

2. The no-restricted-syntax guard in eslint.config.mjs only matched
   literals shaped "UPDATE tasks SET status" -- advanceTaskStatus's
   dynamic `UPDATE tasks SET ${timestampSets.join(", ")}` slipped past,
   and so did saveTask / recoverStaleLocks (status not first column).
   Tighten the selectors to (a) match any status-write within UPDATE
   tasks regardless of column order, and (b) flag template literals
   whose head ends with `SET ` so dynamic helpers must carry an
   explicit eslint-disable with rationale. Five grandfathered sites
   now carry matching disables.

3. interruptStaleRuns scoped its task+lease cleanup with
   `WHERE finished_at = ?` against the freshly stamped timestamp --
   safe in isolation, vulnerable to ISO-millisecond ties with other
   already-interrupted runs. Capture the stale run ids before
   mutating anything and drive every downstream statement from that
   explicit id list. Original semantics preserved; scoping is exact.

Unit tests added in tests/unit/tasks-phase.test.ts cover the
advanceTaskStatus lease cleanup across done / cancelled / blocked /
todo, and pin the in_review case as "lease stays".
2026-04-20 13:51:53 -04:00

425 lines
16 KiB
TypeScript

import { join } from "node:path";
import { tmpdir } from "node:os";
import { mkdtemp, rm } from "node:fs/promises";
import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
vi.mock("electron", () => ({
safeStorage: {
isEncryptionAvailable: () => false,
getSelectedStorageBackend: () => "basic_text",
encryptString: (value: string) => Buffer.from(value, "utf8"),
decryptString: (value: Buffer) => value.toString("utf8"),
},
}));
import { AppDatabase } from "../../src/main/database";
/**
* B2 Phase 1 step 2/5 — unit coverage for the new phase + lease helpers.
*
* These tests exercise the shadow surface directly against a real
* SQLite database: v50 has run, the `tasks.phase` column and
* `task_leases` table exist, and the new helpers write and read them
* without going through any production code path. The goal is to pin
* the state machine semantics (CAS rejection on wrong `from`, lease
* replacement on staleness, release idempotency) before step 3/5 wires
* the helpers into `checkoutTask` / `finishRun`.
*/
function createTestCompany(db: AppDatabase): string {
return db.company.saveCompany({
name: "Phase Helper Company",
description: "B2 phase + lease unit tests",
status: "active",
});
}
function createTestConnector(db: AppDatabase): void {
db.connectorsCrud.saveConnector({
id: "openai-codex",
model: "gpt-5",
envBindingText: "",
notes: "Test connector",
});
}
function createTestWorkspace(db: AppDatabase, companyId: string): string {
return db.workspacesMod.saveWorkspace({
companyId,
name: "Phase Workspace",
localPath: "/tmp/phase-workspace",
repoUrl: "https://github.com/example/phase",
repoRef: "main",
isPrimary: true,
});
}
function createTestAgent(db: AppDatabase, companyId: string, workspaceId: string, name = "Agent"): string {
return db.agents.saveAgent({
companyId,
name,
role: "IC",
title: name,
status: "active",
reportsTo: null,
connectorId: "openai-codex",
workspaceId,
model: null,
capabilities: "Run task helpers",
budgetMonthlyUsd: 100,
});
}
function createTestTask(
db: AppDatabase,
companyId: string,
agentId: string,
workspaceId: string,
): string {
return db.tasks.saveTask({
companyId,
title: "Phase helper task",
description: "dual-write seed",
assigneeAgentId: agentId,
workspaceId,
priority: "medium",
status: "todo",
});
}
function createTestRun(
db: AppDatabase,
companyId: string,
taskId: string,
agentId: string,
workspaceId: string,
): string {
const { id } = db.runs.createRun({
companyId,
taskId,
agentId,
workspaceId,
connectorId: "openai-codex",
});
return id;
}
describe("tasks-phase helpers", () => {
let tempDir: string;
let db: AppDatabase;
beforeEach(async () => {
tempDir = await mkdtemp(join(tmpdir(), "agentcompany-phase-"));
db = new AppDatabase(tempDir);
db.init();
});
afterEach(async () => {
db.close();
await rm(tempDir, { recursive: true, force: true });
});
describe("advanceTaskPhase", () => {
it("backfill leaves every existing task with a phase matching its status", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
// `saveTask` ran with the v50 migration already applied, so the
// row must carry a phase value immediately.
const row = (db as unknown as {
db: { prepare: (sql: string) => { get: (...args: unknown[]) => unknown } };
}).db
.prepare("SELECT phase FROM tasks WHERE id = ?")
.get(taskId) as { phase: string } | undefined;
expect(row?.phase).toBe("todo");
});
it("CAS transition succeeds when the current phase matches `from`", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const ok = db.tasks.advanceTaskPhase(taskId, "todo", "in_progress");
expect(ok).toBe(true);
const after = db.tasks.getTaskLease(taskId);
expect(after).toBeNull(); // phase transitions do not auto-create leases
const row = (db as unknown as {
db: { prepare: (sql: string) => { get: (...args: unknown[]) => unknown } };
}).db
.prepare("SELECT phase FROM tasks WHERE id = ?")
.get(taskId) as { phase: string } | undefined;
expect(row?.phase).toBe("in_progress");
});
it("CAS transition refuses when the current phase does not match `from`", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
// Row is at phase=todo; pretending it was in_progress must fail.
const ok = db.tasks.advanceTaskPhase(taskId, "in_progress", "completed");
expect(ok).toBe(false);
});
it("CHECK constraint rejects a write to an unknown phase value", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
// Using the raw DB to bypass the TypeScript enum — the CHECK
// constraint installed in v50 must still reject the write.
expect(() => {
(db as unknown as {
db: { prepare: (sql: string) => { run: (...args: unknown[]) => unknown } };
}).db
.prepare("UPDATE tasks SET phase = ? WHERE id = ?")
.run("bogus", taskId);
}).toThrow();
});
});
describe("acquireTaskLease", () => {
it("acquires a lease when none exists and returns the row", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
const lease = db.tasks.acquireTaskLease({
taskId,
runId,
ttlMs: 60_000,
checkoutRunId: runId,
});
expect(lease).not.toBeNull();
expect(lease?.taskId).toBe(taskId);
expect(lease?.runId).toBe(runId);
expect(lease?.checkoutRunId).toBe(runId);
});
it("refreshing the lease for the same run succeeds and moves expires_at forward", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
const first = db.tasks.acquireTaskLease({ taskId, runId, ttlMs: 10_000 });
expect(first).not.toBeNull();
const renewed = db.tasks.acquireTaskLease({ taskId, runId, ttlMs: 60_000 });
expect(renewed).not.toBeNull();
expect(Date.parse(renewed!.expiresAt)).toBeGreaterThanOrEqual(Date.parse(first!.expiresAt));
});
it("refuses to hand the lease to a different live run", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runA = createTestRun(db, companyId, taskId, agentId, workspaceId);
const runB = createTestRun(db, companyId, taskId, agentId, workspaceId);
const ownedByA = db.tasks.acquireTaskLease({ taskId, runId: runA, ttlMs: 60_000 });
expect(ownedByA).not.toBeNull();
const blocked = db.tasks.acquireTaskLease({ taskId, runId: runB, ttlMs: 60_000 });
expect(blocked).toBeNull();
const current = db.tasks.getTaskLease(taskId);
expect(current?.runId).toBe(runA);
});
it("replaces a stale lease when the previous window has expired", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runA = createTestRun(db, companyId, taskId, agentId, workspaceId);
const runB = createTestRun(db, companyId, taskId, agentId, workspaceId);
const ownedByA = db.tasks.acquireTaskLease({ taskId, runId: runA, ttlMs: 60_000 });
expect(ownedByA).not.toBeNull();
// Force the lease into the past by rewriting expires_at directly.
const pastIso = new Date(Date.now() - 1_000).toISOString();
(db as unknown as {
db: { prepare: (sql: string) => { run: (...args: unknown[]) => unknown } };
}).db
.prepare("UPDATE task_leases SET expires_at = ? WHERE task_id = ?")
.run(pastIso, taskId);
const takenByB = db.tasks.acquireTaskLease({ taskId, runId: runB, ttlMs: 60_000 });
expect(takenByB).not.toBeNull();
expect(takenByB?.runId).toBe(runB);
});
it("throws when ttlMs is not a positive finite number", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
expect(() =>
db.tasks.acquireTaskLease({ taskId, runId, ttlMs: 0 }),
).toThrow(/positive finite number/);
expect(() =>
db.tasks.acquireTaskLease({ taskId, runId, ttlMs: Number.NaN }),
).toThrow(/positive finite number/);
});
});
describe("advanceTaskStatus lease cleanup", () => {
// B2 Phase 1 follow-up (Linus review ask #1): advanceTaskStatus
// used to write only the status column, leaving any lease row
// behind as an orphan. The helpers below seed a lease directly,
// then drive status transitions and assert the shadow row lifecycle.
function seedLease(taskId: string, runId: string): void {
const acquired = new Date().toISOString();
const expires = new Date(Date.now() + 60_000).toISOString();
(db as unknown as {
db: { prepare: (sql: string) => { run: (...args: unknown[]) => unknown } };
}).db
.prepare(
`INSERT OR REPLACE INTO task_leases
(task_id, run_id, acquired_at, expires_at, checkout_run_id)
VALUES (?, ?, ?, ?, ?)`,
)
.run(taskId, runId, acquired, expires, runId);
}
function readLease(taskId: string): { run_id: string } | undefined {
return (db as unknown as {
db: { prepare: (sql: string) => { get: (...args: unknown[]) => unknown } };
}).db
.prepare("SELECT run_id FROM task_leases WHERE task_id = ?")
.get(taskId) as { run_id: string } | undefined;
}
it("drops the lease when the task transitions to done", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.advanceTaskStatus(taskId, "in_progress");
seedLease(taskId, runId);
expect(readLease(taskId)?.run_id).toBe(runId);
db.tasks.advanceTaskStatus(taskId, "done");
expect(readLease(taskId)).toBeUndefined();
});
it("drops the lease when the task transitions to cancelled", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.advanceTaskStatus(taskId, "in_progress");
seedLease(taskId, runId);
db.tasks.advanceTaskStatus(taskId, "cancelled");
expect(readLease(taskId)).toBeUndefined();
});
it("drops the lease when the task transitions to blocked", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.advanceTaskStatus(taskId, "in_progress");
seedLease(taskId, runId);
db.tasks.advanceTaskStatus(taskId, "blocked");
expect(readLease(taskId)).toBeUndefined();
});
it("drops the lease when the task is bounced back to todo", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.advanceTaskStatus(taskId, "in_progress");
seedLease(taskId, runId);
db.tasks.advanceTaskStatus(taskId, "todo");
expect(readLease(taskId)).toBeUndefined();
});
it("leaves the lease intact when the task transitions to in_review", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
// in_review guard requires either a reports_to or requires_user_review.
db.tasks.setTaskRequiresUserReview(
createTestTask(db, companyId, agentId, workspaceId),
true,
);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
db.tasks.setTaskRequiresUserReview(taskId, true);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.advanceTaskStatus(taskId, "in_progress");
seedLease(taskId, runId);
db.tasks.advanceTaskStatus(taskId, "in_review");
expect(readLease(taskId)?.run_id).toBe(runId);
});
});
describe("releaseTaskLease", () => {
it("releases the lease when the runId matches", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runId = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.acquireTaskLease({ taskId, runId, ttlMs: 60_000 });
const released = db.tasks.releaseTaskLease(taskId, runId);
expect(released).toBe(true);
expect(db.tasks.getTaskLease(taskId)).toBeNull();
});
it("does nothing when a different runId asks to release", () => {
const companyId = createTestCompany(db);
createTestConnector(db);
const workspaceId = createTestWorkspace(db, companyId);
const agentId = createTestAgent(db, companyId, workspaceId);
const taskId = createTestTask(db, companyId, agentId, workspaceId);
const runA = createTestRun(db, companyId, taskId, agentId, workspaceId);
const runB = createTestRun(db, companyId, taskId, agentId, workspaceId);
db.tasks.acquireTaskLease({ taskId, runId: runA, ttlMs: 60_000 });
const released = db.tasks.releaseTaskLease(taskId, runB);
expect(released).toBe(false);
const current = db.tasks.getTaskLease(taskId);
expect(current?.runId).toBe(runA);
});
});
});