Files
PaperClipAI/server/src/services/remote-workspaces.integration.test.ts
T
DottaandPaperclip e3fb4bc73a fix(native): recover saved listener startup failures (#19479)
## Thinking Path

> - Paperclip manages AI agents and their work.
> - Native recovery must preserve saved work and respect current task
authority.
> - A closed local API listener can stop startup before a provider
starts.
> - #19462 gives new failures a bounded bootstrap retry.
> - Older saved failures still have a generic terminal code and a
recovery hold.
> - This pull request verifies the old producer outcome and uses the
existing safe replacement path.
> - Eligible tasks continue once. Unknown provider outcomes and user
holds remain blocked.

## Linked Issues or Issue Description

Refs: #19462

**What happened?**

A restart can leave an old listener startup failure in
`terminal_failure` with `native_session_interrupted`. Startup recovery
does not claim it after #19462. A completed file recovery can also
retain its no-replay hold.

**Expected behavior**

Continue automatically when durable records prove that no provider
started, the source stopped, and the current files were saved. Keep the
existing attempt budget and normal admission checks.

**Steps to reproduce**

1. Close the actual API listener after a fresh native coordinator claim.
2. Persist the old executor's generic failure. Save the workspace and
confirm source termination.
3. Restart recovery. The baseline schedules no successor. The repair
schedules one successor after all receipts pass.

**Paperclip version or commit**

Baseline: `0252421d5c`. Old failure producer: `2f35eb1080`.

**Deployment mode**

Built from source. Local macOS arm64, Node 26.4.0, real embedded
PostgreSQL, independent filesystem execution. No external AI service is
used by the controlled acceptance.

## What Changed

- Recognize the exact saved local-listener guard through one proof
predicate shared by discovery and the locked transaction.
- Require a fresh first attempt, matching execution bindings, no prior
provider lineage, no provider events or tool calls, no checkpoint or
admission receipt, and no cancellation.
- Reuse current snapshot, exact stop, source block, attempt budget, and
successor idempotency checks.
- Wait for completed source file-recovery receipts. Retire only that
same source's verified preservation hold in the successor transaction.
- Bound discovery to 100 candidates per pass, keep completed history
quiet, and wake through existing transactional producer notifications.
- Retain retries after transient database and final ownership-proof
errors; advance the bounded cursor and force a rescan so a failing page
cannot hide later work.
- Let only the replacement duty discharge a proved-empty notification
during drain through its existing read-only readiness check, once per
notification generation. Keep real work and concurrent changes
outstanding without running replacement during drain.
- Wake replacement at canonical local-stop proof writes and the existing
dispatch eligibility boundary; these hints grant no new execution
authority.
- Add database regressions, scheduler quietness checks, real controller
handoff coverage, and an opt-in controlled Runner continuation.

## Verification

Final production head: `0df41b2fef2644195d1224d36984af3827da3e3b`.

Local environment: macOS arm64, Node 26.4.0, TypeScript 7.0.2, Vitest
5.0.3, frozen-lockfile dependencies. Tests execute freshly;
`--fsModuleCache` reuses only content-validated module transforms.

- Native safe replacement: 62/62 PASS on isolated PostgreSQL 17 (63.18
seconds), including clean per-suite database disposal. All 62 assertions
also passed on embedded PostgreSQL 18 beta, but its Darwin cleanup hook
timed out; that invocation is not counted as a pass. Docker's disk was
full, so the disposable PostgreSQL 17 data used a size-limited RAM
mount, with no host mounts. The owned container was removed afterward.
- Actual stopped-process producer suite: 31/31 PASS. A real exited child
produces the stop receipt, and its replacement notification settles only
after the outer transaction commits.
- Workspace and separate-process drain regressions: 33/33 PASS (75.90
seconds).
- Shared coordinator and execution-control lifecycle checks: 39/39 PASS
(16.43 seconds). Pending work, concurrent changes, cold startup,
shutdown, and passive drain boundaries remain covered.
- Actual suppressed Cloud routine/drain workflow: PASS (34.64 seconds),
using the unchanged test and timeouts. Its matching local RED reproduced
CI's 462-query, zero-pending-query, one-idle-hold timeout after an
admitted write committed during drain (69.46 seconds).
- Actual historical controller handoff: PASS (48.40 seconds). A separate
controller loads the original executor from `2f35eb1080`, writes its
generic listener failure, and exits. Recovery at the final head admits
exactly one successor. The real Rust Runner and controlled Codex process
complete governed `paperclip_finish`; attempt two commits and the saved
file remains intact. Only the child loads the old artifact, and both
isolated artifacts are restored afterward. Rust source subtree
`77961b70f177ec3c38d11abcaec92fd3d0e96ec0` matches the built binary. No
external AI service is used.
- Seven corrected existing fixtures pass: transactional stop-proof calls
(5 cases) and notification inventory (2 cases). Their earlier CI
failures exposed fixture calls outside the documented transaction and an
outdated expected queue set.
- Server typecheck: PASS.

Red/green proofs also cover the original saved failure on baseline
`0252421d5c` (zero successors instead of one), held-history quietness,
bounded forward scanning through 100 parked rows and concurrent wakes,
real snapshot/cleanup producer notifications, actual issue-row lock
timeout, and task/history eligibility notifications. An obsolete
prior-run question can be skipped to request another proof check; actual
skip while the task is paused or the source owns provider tool activity
schedules no successor.

The failed-page regression starts with 25 persistent verifier errors and
a healthy candidate 26. Baseline RED repeats the first page. Repair
GREEN advances the healthy candidate, revisits the original failures
after their fault clears, creates all 26 successors exactly once, and
becomes quiet.

Greptile is SUCCESS, 5/5 on the exact final head (2026-11-06 17:09:21
UTC). Final-head CI is GREEN: 56 successful checks, three intentional
skips, one neutral result, and no pending or failing checks. [Attempt
two and
results](https://github.com/paperclipai/paperclip/actions/runs/42258352876/attempts/2)
include passing build, typecheck, Rust, native server, package, browser
E2E, serialized server, general server/chat tests, and final
verification. Five jobs in the first attempt were interrupted together
at 17:13:12 UTC; the native-server log records an explicit runner
shutdown signal. The single infrastructure retry reran only those five
jobs and dependent verification, preserving the 49 successful checks on
the unchanged head. No timeout was weakened, and setup failures or
interrupted invocations are not counted as passes.

Focused commands from the repository root:

```sh
./server/node_modules/.bin/vitest run --project @paperclipai/server server/src/services/remote-workspaces.integration.test.ts server/src/services/native-runtime/native-runner-restart-preparation.integration.test.ts -t 'native safe-replacement decision|reconciles the drained remote backend' --fsModuleCache --reporter=dot
./server/node_modules/.bin/vitest run --project @paperclipai/server server/src/__tests__/execution-control-work.test.ts server/src/__tests__/delivery-work-coordinator.test.ts -t 'execution-control lifecycle|work coordinator' --fsModuleCache --reporter=dot
node node_modules/typescript/bin/tsc -p server/tsconfig.json --noEmit --pretty false
```

The historical acceptance sets `PAPERCLIP_TEST_HISTORICAL_LISTENER=1`
and uses an original-executor child artifact with a `try/finally`
restore. Normal regressions require no Rust binary.

## Risks

- Compatibility is deliberately limited to positively proved fresh
remote startup failures. Other old ambiguous failures remain held.
- Existing provider work, a changed task block, incomplete preservation,
or revoked authority must prevent execution. Normal retry admission
remains in control.
- This PR does not deploy the change or mutate a hosted task. Hosted
recovery still requires the repaired revision and all current proof
gates.

## Model Used

OpenAI GPT-6 through Codex, with reasoning, tool use, and code
execution. The exact backend model identifier and context window are not
exposed in this session.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-11-06 11:49:28 -06:00

2014 lines
165 KiB
TypeScript

import { parseNativeExecutionInput } from "../vendor/paperclip-runner/index.js";
import { readNativeWorkspaceMigration, migratedNativeExecutionCwd, canRestoreMigratedNativeSession } from "./native-runtime/native-workspace-migration.js";
import { buildNativeHarnessBackupManifest, verifyNativeHarnessBackup } from "./native-runtime/native-session-executor.js";
import { createHash, randomUUID } from "node:crypto";
import { execFile, spawn } from "node:child_process";
import net from "node:net";
import { createServer } from "node:http";
import { promisify } from "node:util";
import fs from "node:fs/promises";
import os from "node:os";
import path from "node:path";
import { pathToFileURL } from "node:url";
import { afterAll, afterEach, beforeAll, beforeEach, describe, expect, it, vi } from "vitest";
import { and, eq, sql } from "drizzle-orm";
import { activityLog, agents, agentTaskSessions, agentWakeupRequests, applyPendingMigrations, closeRegisteredClients, companies, createDb, environmentLeases, environments,
subscribeDatabaseWork, executionWorkspaceRepositories, executionWorkspaces, executionWorkspaceSnapshots, heartbeatRunEvents, heartbeatRuns, issueRecoveryActions, issueThreadInteractions, issues, livenessDb, nativeRunFinalizations, projects, projectWorkspaces, workspaceOperations } from "@paperclipai/db";
import type { AdapterExecutionTarget } from "@paperclipai/adapter-utils/execution-target";
import type { CommandManagedRuntimeRunner } from "@paperclipai/adapter-utils/command-managed-runtime";
import { buildProjectMentionHref, type Environment, type EnvironmentLease } from "@paperclipai/shared";
import { createStorageService } from "../storage/service.js";
import { createLocalDiskStorageProvider } from "../storage/local-disk-provider.js";
import { createS3StorageProvider } from "../storage/s3-provider.js";
import type { StorageService } from "../storage/types.js";
import { createSavedWorkspaceLoader, prepareRemoteWorkspace, readRemoteWorkspaceAuthority,
clearRemoteWorkspaceLocationForReleasedLease, createRemoteWorkspaceGitAuth, prepareRemoteWorkspaceRepository,
reserveRemoteWorkspace, saveRemoteWorkspace, shouldUseRemoteWorkspaces } from "./remote-workspaces.js";
import type { createGitRemoteAuthProvider } from "./git-credentials.js";
import { invalidateNativeReplacementDiscovery, reconcileSafeNativeReplacements } from "./native-runtime/native-safe-replacement.js";
import { remoteTerminationReceipt } from "./remote-execution-termination.js";
import { RemoteWorkspaceRestoreError } from "./remote-workspace-transport.js";
import { readRemoteRunnerState } from "./native-runtime/native-session-executor.js";
import { notFound } from "../errors.js";
import { appendHeartbeatRunEvent } from "./heartbeat-run-events.js";
import { issueThreadInteractionService } from "./issue-thread-interactions.js";
import { createRemoteWorkspaceFiles } from "./remote-workspace-files.js";
import { workspaceFileResourceService } from "./workspace-file-resources.js";
import { nativeWorkspaceExportPendingForLease, preserveRemoteWorkspacePreparationFailure } from "./native-runtime/native-workspace-finalization-state.js";
import { preserveOrphanedNativeWorkspace } from "./native-runtime/native-failed-workspace-recovery.js";
import { terminalizeLegacyExecution } from "./legacy-execution-recovery.js";
import { environmentRuntimeService } from "./environment-runtime.js";
import { legacyControllerClaim, renewLegacyControllerLease } from "./legacy-controller-lease.js";
import { migrateRemoteWorkspace } from "./remote-workspace-migration.js";
import { executionWorkspaceRepositoryService, taskRepositoryRelativePath } from "./execution-workspace-repositories.js";
import { registerRemoteWorkspaceSnapshotRetentionWork, remoteWorkspaceSnapshotRetentionService } from "./remote-workspace-snapshot-retention.js";
import { createDeliveryWorkCoordinator } from "./delivery-work-coordinator.js";
import { issueService } from "./issues.js";
import { instanceSettingsService } from "./instance-settings.js";
import { resolveRemoteWorkspaceSource } from "./heartbeat/remote-workspace-source.js";
import { toExecutionWorkspace } from "./execution-workspaces.js";
import { recoveryPostgresSupport, startRecoveryPostgres } from "../__tests__/helpers/recovery-postgres.js";
import { getExecutionBlocker } from "./execution-blocker.js";
import { executionControlSchedule } from "./execution-control-work.js";
import { DELIVERY_QUEUES } from "./delivery-work-notifications.js";
import { createSystemRecovery } from "./system-recovery-state.js";
const exec = promisify(execFile);
let databaseUrl = process.env.PAPERCLIP_REMOTE_WORKSPACE_TEST_DATABASE_URL;
if (databaseUrl && !/^remote_workspace_[a-z_]*test$/.test(new URL(databaseUrl).pathname.slice(1))) throw new Error("Disposable workspace database required");
const suite = databaseUrl || recoveryPostgresSupport.supported ? describe : describe.skip;
const s3Bucket = process.env.PAPERCLIP_REMOTE_WORKSPACE_TEST_S3_BUCKET;
if (s3Bucket && !s3Bucket.startsWith("paperclip-workspace-")) throw new Error("Disposable workspace S3 bucket required");
suite("remote workspace lifecycle with real Postgres and independent filesystem execution", () => {
let db: ReturnType<typeof createDb>, root: string, storage: StorageService;
let temporary: Awaited<ReturnType<typeof startRecoveryPostgres>> | undefined;
let ids: { company: string; agent: string; issue: string; run: string; environment: string; lease: string };
beforeAll(async () => {
if (!databaseUrl) {
temporary = await startRecoveryPostgres("remote-workspace-lifecycle-");
databaseUrl = temporary.connectionString;
}
await applyPendingMigrations(databaseUrl); db = createDb(databaseUrl);
}, 120_000);
afterAll(async () => {
try { if (databaseUrl) await closeRegisteredClients(databaseUrl); }
finally { await temporary?.cleanup(); }
});
beforeEach(async () => {
vi.stubEnv("DATABASE_URL", databaseUrl!); vi.stubEnv("PAPERCLIP_SECRETS_MASTER_KEY", "62".repeat(32));
root = await fs.realpath(await fs.mkdtemp(path.join(os.tmpdir(), "pc-remote-workspace-lifecycle-")));
storage = s3Bucket ? createStorageService(createS3StorageProvider({ bucket: s3Bucket, region: "us-east-1", prefix: "lifecycle-tests" }))
: { ...createStorageService(createLocalDiskStorageProvider(path.join(root, "objects"))), provider: "s3" };
ids = Object.fromEntries(["company", "agent", "issue", "run", "environment", "lease"].map(key => [key, randomUUID()])) as typeof ids;
await db.insert(companies).values({ id: ids.company, name: "Disposable remote workspace proof", issuePrefix: "RWP" });
await db.insert(agents).values({ id: ids.agent, companyId: ids.company, name: "Native worker", adapterType: "paperclip_runner" });
await db.insert(environments).values({ id: ids.environment, name: `Independent filesystem execution ${ids.environment}`, driver: "sandbox", config: {} });
await db.insert(issues).values({ id: ids.issue, companyId: ids.company, title: "Use a remote workspace", status: "in_progress", assigneeAgentId: ids.agent });
await db.insert(heartbeatRuns).values({ id: ids.run, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue,
runtimeMode: "native", status: "running" });
invalidateNativeReplacementDiscovery(db);
await db.insert(environmentLeases).values({ id: ids.lease, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: ids.run, issueId: ids.issue, provider: "daytona", providerLeaseId: `allocation-${ids.lease}`, status: "active" });
});
afterEach(async () => {
const snapshots = await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.companyId, ids.company));
for (const row of snapshots) for (const key of new Set([row.objectKey, ...(Array.isArray(row.metadata?.uploads)
? (row.metadata.uploads as { objectKey: string }[]).map(item => item.objectKey) : [])].filter((value): value is string => !!value))) {
await storage.deleteObject(ids.company, key);
}
if (temporary && process.env.PAPERCLIP_TEST_HISTORICAL_LISTENER === "1") await db.transaction(async tx => {
await tx.execute(sql`select set_config('client_min_messages', 'warning', true)`);
await tx.execute(sql`TRUNCATE companies CASCADE`);
});
await db.delete(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.companyId, ids.company));
await db.delete(workspaceOperations).where(eq(workspaceOperations.companyId, ids.company));
await db.delete(activityLog).where(eq(activityLog.companyId, ids.company));
await db.delete(environmentLeases).where(eq(environmentLeases.companyId, ids.company));
await db.delete(nativeRunFinalizations).where(eq(nativeRunFinalizations.companyId, ids.company));
await db.delete(issueRecoveryActions).where(eq(issueRecoveryActions.companyId, ids.company));
await db.delete(heartbeatRunEvents).where(eq(heartbeatRunEvents.companyId, ids.company));
await db.delete(agentTaskSessions).where(eq(agentTaskSessions.companyId, ids.company));
await db.delete(heartbeatRuns).where(eq(heartbeatRuns.companyId, ids.company));
await db.delete(agentWakeupRequests).where(eq(agentWakeupRequests.companyId, ids.company));
await db.update(issues).set({ executionWorkspaceId: null }).where(eq(issues.companyId, ids.company));
await db.delete(executionWorkspaces).where(eq(executionWorkspaces.companyId, ids.company));
await db.delete(issues).where(eq(issues.companyId, ids.company));
await db.delete(projectWorkspaces).where(eq(projectWorkspaces.companyId, ids.company));
await db.delete(projects).where(eq(projects.companyId, ids.company));
await db.delete(agents).where(eq(agents.companyId, ids.company));
await db.delete(companies).where(eq(companies.id, ids.company));
await db.delete(environments).where(eq(environments.id, ids.environment));
await fs.rm(root, { recursive: true, force: true }); vi.unstubAllEnvs();
});
async function target(name: string): Promise<Extract<AdapterExecutionTarget, { transport: "sandbox" }>> {
const remoteCwd = path.join(root, name); await fs.mkdir(remoteCwd);
const runner: CommandManagedRuntimeRunner = {
execute: async input => {
try {
const result = await exec(input.command, input.args ?? [], { cwd: input.cwd ?? remoteCwd,
env: { ...process.env, ...input.env }, timeout: input.timeoutMs, maxBuffer: 4 * 1024 * 1024 });
return { exitCode: 0, stdout: result.stdout, stderr: result.stderr, timedOut: false, signal: null, pid: process.pid, startedAt: new Date().toISOString() };
} catch (error) { const failed = error as { code?: number; stdout?: string; stderr?: string };
return { exitCode: typeof failed.code === "number" ? failed.code : 1, stdout: failed.stdout ?? "", stderr: failed.stderr ?? "", timedOut: false, signal: null, pid: process.pid, startedAt: new Date().toISOString() }; }
},
syncIn: async operations => transfer(operations), syncOut: async operations => transfer(operations),
};
return { kind: "remote", transport: "sandbox", remoteCwd, runner, provider: "daytona", leaseId: ids.lease,
sandboxId: `allocation-${ids.lease}`, timeoutMs: 300_000 } as Extract<AdapterExecutionTarget, { transport: "sandbox" }>;
}
async function transfer(operations: Parameters<NonNullable<CommandManagedRuntimeRunner["syncIn"]>>[0]) {
const results = [];
for (const operation of operations) {
let bytes = 0;
for (const file of operation.files) {
await fs.mkdir(path.dirname(file.targetPath), { recursive: true });
await fs.cp(file.sourcePath, file.targetPath, { recursive: file.kind === "directory" });
if (file.kind === "file") bytes += (await fs.stat(file.targetPath)).size;
}
results.push({ operationId: operation.operationId, filesTransferred: operation.files.length, bytesTransferred: bytes });
}
return { operations: results };
}
const lease = async () => (await db.select().from(environmentLeases).where(eq(environmentLeases.id, ids.lease)))[0]! as EnvironmentLease;
const workspace = async (id: string) => (await db.select().from(executionWorkspaces).where(eq(executionWorkspaces.id, id)))[0]!;
const reserve = (source: Parameters<typeof reserveRemoteWorkspace>[0]["source"]) => reserveRemoteWorkspace({ db, storage,
companyId: ids.company, issueId: ids.issue, runId: ids.run, agentId: ids.agent, environmentId: ids.environment,
projectId: null, projectWorkspaceId: null, existingWorkspaceId: null, source });
it("keeps a fresh preparation timeout eligible for bounded bootstrap recovery", async () => {
const reserved = await reserve({ kind: "empty" }), remote = await target("timed-out-bootstrap");
const execute = vi.spyOn(remote.runner!, "execute").mockResolvedValueOnce({
exitCode: null, timedOut: true, stdout: "untrusted command output", stderr: "untrusted command output", signal: null,
pid: null, startedAt: new Date().toISOString(),
});
await expect(prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: remote }))
.rejects.toMatchObject({ code: "remote_workspace_prepare_timeout" });
expect(execute).toHaveBeenCalledOnce();
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toMatchObject({
writerRunId: ids.run, location: null, snapshotId: null, dirty: false,
});
expect(await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.companyId, ids.company))).toHaveLength(0);
});
it.each(["dirty", "owner_changed"] as const)("does not grant bootstrap replay after the workspace becomes %s", async state => {
const reserved = await reserve({ kind: "empty" }), remote = await target(`changed-${state}`);
await fs.writeFile(path.join(remote.remoteCwd, "saved.txt"), "retain existing work\n");
const authority = readRemoteWorkspaceAuthority(reserved.metadata)!;
vi.spyOn(remote.runner!, "execute").mockImplementationOnce(async () => {
await db.update(executionWorkspaces).set({ metadata: { ...reserved.metadata, remoteWorkspace: {
...authority, ...(state === "dirty" ? { dirty: true } : { epoch: authority.epoch + 1, writerRunId: randomUUID() }),
} } }).where(eq(executionWorkspaces.id, reserved.id));
return { exitCode: null, timedOut: true, stdout: "", stderr: "", signal: null, pid: null, startedAt: new Date().toISOString() };
});
await expect(prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: remote }))
.rejects.toMatchObject({ name: "RemoteWorkspaceOperationTimeoutError" });
expect(await fs.readFile(path.join(remote.remoteCwd, "saved.txt"), "utf8")).toBe("retain existing work\n");
});
async function releaseBootstrapLease(receipt: boolean) {
const pluginId = randomUUID();
await db.update(environmentLeases).set({ metadata: { driver: "sandbox", pluginId, provider: "daytona",
sandboxProviderPlugin: true } }).where(eq(environmentLeases.id, ids.lease));
const runtime = environmentRuntimeService(db, { pluginWorkerManager: {
isRunning: () => true,
getWorker: () => ({ supportedMethods: ["environmentReleaseLease"] }),
call: async (_id: string, method: string) => {
expect(method).toBe("environmentReleaseLease");
return receipt ? { providerLeaseId: `allocation-${ids.lease}`, state: "stopped" } : undefined;
},
} as never });
await runtime.releaseRunLeases(ids.run, "failed");
expect(await lease()).toMatchObject({ status: "failed", cleanupStatus: "success" });
expect(Boolean((await lease()).metadata?.remoteExecutionTermination)).toBe(receipt);
}
it.each([["receipt", "fresh"], ["void", "fresh"], ["receipt", "reused"]] as const)(
"continues unpublished preparation after %s cleanup on a %s allocation and retains the resulting work", async (cleanup, allocation) => {
await db.update(heartbeatRuns).set({ runtimeMode: "legacy", nativeIssueId: null, issueId: ids.issue })
.where(eq(heartbeatRuns.id, ids.run));
const reserved = await reserve({ kind: "empty" }), first = await target("lost-bootstrap-reply");
const interrupted = allocation === "reused";
const actual = first.runner!.execute.bind(first.runner!);
const execute = vi.spyOn(first.runner!, "execute").mockImplementationOnce(async input => {
const result = await actual(input);
expect(result.exitCode).toBe(0);
return { ...result, timedOut: !interrupted };
});
let ownershipChecks = 0;
const preparation = prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first,
assertOwned: async () => { if (interrupted && ++ownershipChecks === 2) throw new Error("lease_released_before_terminal"); } });
if (interrupted) await expect(preparation).rejects.toThrow("lease_released_before_terminal");
else await expect(preparation).rejects.toMatchObject({ code: "remote_workspace_prepare_timeout" });
const oldRoot = execute.mock.calls[0][0].args?.[2];
if (!oldRoot) throw new Error("Preparation did not record its directory");
const oldReceipt = await fs.readFile(path.join(oldRoot, ".paperclip-runtime", "workspace-ready.json"), "utf8");
const [failed] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run));
await terminalizeLegacyExecution({ db, run: failed!, status: interrupted ? "interrupted" : "failed", patch: {
errorCode: interrupted ? "lease_released_before_terminal" : "remote_workspace_prepare_timeout",
resultJson: { executionRecovery: { kind: "bootstrap", providerWorkStarted: false } },
} });
await releaseBootstrapLease(cleanup === "receipt");
const continueOn = async (name: string) => {
const runId = randomUUID(), leaseId = randomUUID();
const reused = allocation === "reused" && name === "fresh-bootstrap-retry";
const providerLeaseId = reused ? `allocation-${ids.lease}` : `allocation-${leaseId}`;
await db.insert(heartbeatRuns).values({ id: runId, companyId: ids.company, agentId: ids.agent,
nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
const [nextLease] = await db.insert(environmentLeases).values({ id: leaseId, companyId: ids.company,
environmentId: ids.environment, heartbeatRunId: runId, issueId: ids.issue, provider: "daytona",
providerLeaseId, status: "active" }).returning();
const next = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue,
runId, agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null,
existingWorkspaceId: reserved.id, source: { kind: "empty" } });
const remote = reused ? { ...first, leaseId, sandboxId: providerLeaseId,
runner: { ...first.runner!, execute: actual } } : await target(name);
return prepareRemoteWorkspace({ db, storage, workspace: next, runId, lease: nextLease as EnvironmentLease,
target: remote }).then(prepared => ({ runId, prepared, remote }));
};
const { runId, prepared, remote } = await continueOn("fresh-bootstrap-retry");
expect(prepared.remoteRoot).not.toBe(oldRoot);
await fs.writeFile(path.join(prepared.remoteRoot, "report.md"), "useful recovered result\n");
const { snapshotId } = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId,
lease: prepared.lease, target: remote, assertOwned: async () => undefined });
await db.update(heartbeatRuns).set({ status: "succeeded", finishedAt: new Date() }).where(eq(heartbeatRuns.id, runId));
await fs.rm(prepared.remoteRoot, { recursive: true });
const restored = await continueOn("cold-bootstrap-continuation");
expect(await fs.readFile(path.join(restored.prepared.remoteRoot, "report.md"), "utf8")).toBe("useful recovered result\n");
expect(restored.prepared.reference.snapshotId).not.toBeNull();
expect(snapshotId).toBeTruthy();
expect(execute).toHaveBeenCalledOnce();
expect(await fs.readFile(path.join(oldRoot, ".paperclip-runtime", "workspace-ready.json"), "utf8")).toBe(oldReceipt);
}, 60_000);
it.each(["unclassified", "provider_started", "retained_native", "runner_identity", "native_mode", "saved", "published", "dirty", "cleanup_failed", "active_lease"] as const)(
"keeps receiptless workspace handoff blocked for %s", async state => {
const reserved = await reserve({ kind: "empty" });
let savedAuthority: ReturnType<typeof readRemoteWorkspaceAuthority> = null;
let remoteRoot: string | undefined;
if (state === "saved" || state === "published") {
const remote = await target(`preserved-${state}`);
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: remote });
remoteRoot = prepared.remoteRoot;
await fs.writeFile(path.join(remoteRoot, "saved.txt"), "keep prior files\n");
await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: remote, boundary: "initial", assertOwned: async () => undefined });
savedAuthority = readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata);
}
await db.update(heartbeatRuns).set({ status: "failed", runtimeMode: state === "native_mode" ? "native" : "legacy", finishedAt: new Date(),
errorCode: state === "unclassified" ? "setup_failed" : "remote_workspace_prepare_timeout",
nativeSessionId: state === "retained_native" ? randomUUID() : null,
runnerInstanceId: state === "runner_identity" ? randomUUID() : null,
resultJson: { executionRecovery: { kind: "bootstrap", providerWorkStarted: state === "provider_started" } },
}).where(eq(heartbeatRuns.id, ids.run));
if (state !== "active_lease") await releaseBootstrapLease(false);
if (state === "cleanup_failed") await db.update(environmentLeases).set({ cleanupStatus: "failed" }).where(eq(environmentLeases.id, ids.lease));
if (savedAuthority) await db.update(executionWorkspaces).set({ metadata: { ...reserved.metadata,
remoteWorkspace: { ...savedAuthority, dirty: false, snapshotId: state === "published" ? null : savedAuthority.snapshotId },
} }).where(eq(executionWorkspaces.id, reserved.id));
if (state === "dirty") await db.update(executionWorkspaces).set({ metadata: { ...reserved.metadata,
remoteWorkspace: { ...readRemoteWorkspaceAuthority(reserved.metadata)!, dirty: true },
} }).where(eq(executionWorkspaces.id, reserved.id));
const nextRun = randomUUID();
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent,
nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
await expect(reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue,
runId: nextRun, agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null,
existingWorkspaceId: reserved.id, source: { kind: "empty" } })).rejects.toThrow("unfinished writer");
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.writerRunId).toBe(ids.run);
if (remoteRoot) expect(await fs.readFile(path.join(remoteRoot, "saved.txt"), "utf8")).toBe("keep prior files\n");
});
it.each(["initial", "follow-up"] as const)("edits and cold-restores both an anchor and a referenced project's task copy (%s attachment)", async attachment => {
vi.stubEnv("PAPERCLIP_MULTI_PROJECT_WORKSPACE_SYNC", "true");
vi.stubEnv("PAPERCLIP_MULTI_PROJECT_WORKSPACE_SYNC_REMOTE", "true");
const origins = path.join(root, "origins"); await fs.mkdir(origins);
const git = (cwd: string, ...args: string[]) => exec("git", ["-C", cwd,
"-c", "user.name=Task copies", "-c", "user.email=copies@example.invalid", ...args]);
for (const name of ["main", "skills"]) {
const directory = path.join(origins, name); await fs.mkdir(directory);
await git(directory, "init", "--initial-branch=main");
await fs.writeFile(path.join(directory, "README.md"), `${name} source\n`);
await git(directory, "add", "."); await git(directory, "commit", "-m", "Source");
if (name === "skills") {
await git(directory, "checkout", "-b", "release");
await fs.writeFile(path.join(directory, "README.md"), "skills release\n");
await git(directory, "add", "."); await git(directory, "commit", "-m", "Release");
await git(directory, "checkout", "main");
}
}
const port = await new Promise<number>((resolve, reject) => {
const listener = net.createServer(); listener.on("error", reject);
listener.listen(0, "127.0.0.1", () => {
const port = (listener.address() as net.AddressInfo).port;
listener.close(error => error ? reject(error) : resolve(port));
});
});
const daemon = spawn("git", ["daemon", "--reuseaddr", "--export-all", `--base-path=${origins}`,
"--listen=127.0.0.1", `--port=${port}`, "--verbose", origins], { stdio: ["ignore", "ignore", "pipe"] });
try {
await new Promise<void>((resolve, reject) => {
daemon.once("error", reject); daemon.once("exit", code => reject(new Error(`Git source exited: ${code}`)));
daemon.stderr.on("data", bytes => { if (String(bytes).includes("Ready to rumble")) resolve(); });
});
const [anchor, reference] = await db.insert(projects).values(["Main", "Skills"].map(name => ({ companyId: ids.company, name }))).returning();
const [anchorSource] = await db.insert(projectWorkspaces).values([
{ companyId: ids.company, projectId: anchor.id, name: "Main", repoUrl: `git://127.0.0.1:${port}/main`, sourceType: "git_repo" },
{ companyId: ids.company, projectId: reference.id, name: "Skills", repoUrl: `git://127.0.0.1:${port}/skills`, defaultRef: "release", sourceType: "git_repo" },
]).returning();
await db.update(issues).set({ projectId: anchor.id, projectWorkspaceId: anchorSource.id,
description: attachment === "initial" ? `Edit both projects, including [@Skills](${buildProjectMentionHref(reference.id)}).`
: "Start in the main project." }).where(eq(issues.id, ids.issue));
const input = { companyId: ids.company, issueId: ids.issue, agentId: ids.agent, runId: ids.run,
responsibleUserId: null, projectId: anchor.id, projectWorkspaceId: anchorSource.id };
const resolved = await resolveRemoteWorkspaceSource(db, input);
if (resolved.authority !== "remote") throw new Error("Expected remote task copies");
const reserved = await reserveRemoteWorkspace({ db, storage, ...input, environmentId: ids.environment,
existingWorkspaceId: null, source: resolved.source });
const first = await target("project-copies-A");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const initialCopy = resolved.source.repositories?.[0];
if (attachment === "initial") {
expect(initialCopy?.projectId).toBe(reference.id); expect(initialCopy?.readOnly).not.toBe(true);
expect(await fs.readFile(path.join(prepared.remoteRoot, initialCopy!.relativePath, "README.md"), "utf8")).toBe("skills release\n");
}
for (const relative of ["", ...(initialCopy ? [initialCopy.relativePath] : [])]) {
await fs.writeFile(path.join(prepared.remoteRoot, relative, "result.md"), "saved edits\n");
}
for (const name of ["main", "skills"]) {
await expect(fs.access(path.join(origins, name, "result.md"))).rejects.toMatchObject({ code: "ENOENT" });
}
await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
await db.update(heartbeatRuns).set({ status: "succeeded", finishedAt: new Date() }).where(eq(heartbeatRuns.id, ids.run));
await fs.rm(path.dirname(prepared.remoteRoot), { recursive: true });
await fs.rm(path.join(origins, "main"), { recursive: true });
if (attachment === "initial") await fs.rm(path.join(origins, "skills"), { recursive: true });
else await db.update(issues).set({ description: `Now edit [@Skills](${buildProjectMentionHref(reference.id)}).` }).where(eq(issues.id, ids.issue));
const nextRun = randomUUID(), nextLeaseId = randomUUID();
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent,
nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
const [nextLease] = await db.insert(environmentLeases).values({ id: nextLeaseId, companyId: ids.company,
environmentId: ids.environment, heartbeatRunId: nextRun, issueId: ids.issue,
provider: "daytona", providerLeaseId: `allocation-${nextLeaseId}`, status: "active" }).returning();
const resumed = await resolveRemoteWorkspaceSource(db, { ...input, runId: nextRun,
existingWorkspace: toExecutionWorkspace(await workspace(reserved.id)) });
if (resumed.authority !== "remote") throw new Error("Expected retained remote task copies");
if (attachment === "initial") expect(resumed.source).toEqual(resolved.source);
const copy = resumed.source.repositories![0]!;
expect(copy.projectId).toBe(reference.id);
const reservedAgain = await reserveRemoteWorkspace({ db, storage, ...input, runId: nextRun,
environmentId: ids.environment, existingWorkspaceId: reserved.id, source: resumed.source });
const second = await target("project-copies-B");
const restored = await prepareRemoteWorkspace({ db, storage, workspace: reservedAgain, runId: nextRun,
lease: nextLease as EnvironmentLease, target: second });
for (const relative of ["", copy.relativePath]) {
if (!relative || attachment === "initial") expect(await fs.readFile(path.join(restored.remoteRoot, relative, "result.md"), "utf8")).toBe("saved edits\n");
await fs.writeFile(path.join(restored.remoteRoot, relative, "follow-up.md"), "continued edits\n");
}
expect(await fs.readFile(path.join(restored.remoteRoot, copy.relativePath, "README.md"), "utf8")).toBe("skills release\n");
await saveRemoteWorkspace({ db, storage, workspace: restored.workspace, runId: nextRun,
lease: restored.lease, target: second, assertOwned: async () => undefined });
const final = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(final.metadata)!;
expect(authority).toMatchObject({ dirty: false, writerRunId: null });
expect(final.metadata?.remoteWorkspaceSnapshotRepositories).toEqual({ snapshotId: authority.snapshotId, repositories: resumed.source.repositories });
await fs.rm(origins, { recursive: true, force: true });
const saved = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id, snapshotId: authority.snapshotId! });
try {
for (const relative of ["", copy.relativePath]) expect(await fs.readFile(path.join(saved.root, relative, "follow-up.md"), "utf8")).toBe("continued edits\n");
expect(await fs.readFile(path.join(saved.root, copy.relativePath, "README.md"), "utf8")).toBe("skills release\n");
} finally { await saved.cleanup(); }
} finally {
if (daemon.exitCode === null) {
const stopped = new Promise<void>(resolve => daemon.once("exit", () => resolve()));
daemon.kill("SIGTERM"); await stopped;
}
}
}, 60_000);
it.each(["same root", "different root", "destroyed allocation", "occupied root", "missing object", "transient restore"] as const)(
"preserves the published workspace path and files on a cold restore (%s)", async mode => {
const reserved = await reserve({ kind: "empty" }), first = await target("sandbox-A");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
expect(readRemoteWorkspaceAuthority(prepared.workspace.metadata)?.snapshotId).toBeTruthy();
expect(await fs.readdir(path.join(prepared.remoteRoot, ".paperclip-runtime", "workspace-snapshots"))).toEqual([]);
await fs.writeFile(path.join(prepared.remoteRoot, "report.md"), "useful saved result\n");
await fs.mkdir(path.join(prepared.remoteRoot, "empty-folder"));
await fs.writeFile(path.join(prepared.remoteRoot, "binary.bin"), Buffer.from([0, 1, 255]));
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
const current = await workspace(reserved.id);
expect(readRemoteWorkspaceAuthority(current.metadata)).toMatchObject({ dirty: false, writerRunId: null, snapshotId: saved.snapshotId });
if (mode === "destroyed allocation") {
const destroyed: EnvironmentLease = { ...prepared.lease, status: "released", releasedAt: new Date(), cleanupStatus: "success",
metadata: { ...prepared.lease.metadata, remoteExecutionTermination: remoteTerminationReceipt(prepared.lease,
{ providerLeaseId: prepared.lease.providerLeaseId, state: "destroyed" }) } };
await db.transaction(tx => clearRemoteWorkspaceLocationForReleasedLease({ db: tx, lease: destroyed }));
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.location).toBeNull();
}
await fs.rm(path.join(root, "sandbox-A"), { recursive: true, force: true });
const second = await target(mode === "different root" ? "sandbox-B" : "sandbox-A");
const expectedRemoteRoot = path.join(second.remoteCwd, ".paperclip-workspaces", path.basename(prepared.remoteRoot));
const nextRun = randomUUID(), nextLease = randomUUID();
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
await db.insert(environmentLeases).values({ id: nextLease, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: nextRun, issueId: ids.issue, provider: "daytona", providerLeaseId: `allocation-${nextLease}`, status: "active" });
const reservedAgain = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue, runId: nextRun,
agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null, existingWorkspaceId: reserved.id, source: { kind: "empty" } });
const newLease = (await db.select().from(environmentLeases).where(eq(environmentLeases.id, nextLease)))[0]! as EnvironmentLease;
if (mode === "occupied root") {
await fs.mkdir(expectedRemoteRoot, { recursive: true });
await fs.writeFile(path.join(expectedRemoteRoot, "unpublished.txt"), "preserve unknown preparation\n");
const occupied = await prepareRemoteWorkspace({ db, storage, workspace: reservedAgain, runId: nextRun, lease: newLease, target: second })
.then(() => null, (error: unknown) => error as RemoteWorkspaceRestoreError);
// The restore names its step and the occupied folder; it is not retried.
expect(occupied).toBeInstanceOf(RemoteWorkspaceRestoreError);
expect(occupied).toMatchObject({ step: "sandbox_restore", retryable: false, message: "remote_workspace_restore_failed: step=sandbox_restore; "
+ "remote_workspace_operation_failed: operation=restore exit=1; remote_workspace_restore_failed: stage=inspect_root reason=occupied_root" });
expect(await fs.readFile(path.join(expectedRemoteRoot, "unpublished.txt"), "utf8")).toBe("preserve unknown preparation\n");
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata))
.toEqual(readRemoteWorkspaceAuthority(reservedAgain.metadata));
await fs.rename(expectedRemoteRoot, path.join(root, "retained-unknown-preparation"));
}
if (mode === "transient restore") {
// An unnamed sandbox failure (an I/O error) is retryable; the retry restores the saved files.
const run = second.runner!.execute.bind(second.runner);
vi.spyOn(second.runner!, "execute").mockImplementation(async command => command.args?.[1]?.includes("remote_workspace_restore_failed")
? { exitCode: 1, stdout: "", stderr: "remote_workspace_restore_failed: stage=extract code=EIO exit=1", timedOut: false, signal: null,
pid: process.pid, startedAt: new Date().toISOString() } : run(command));
const transient = await prepareRemoteWorkspace({ db, storage, workspace: reservedAgain, runId: nextRun, lease: newLease, target: second })
.then(() => null, (error: unknown) => error as RemoteWorkspaceRestoreError);
expect(transient).toMatchObject({ step: "sandbox_restore", retryable: true, message: "remote_workspace_restore_failed: step=sandbox_restore; "
+ "remote_workspace_operation_failed: operation=restore exit=1; remote_workspace_restore_failed: stage=extract code=EIO exit=1" });
expect(await fs.lstat(expectedRemoteRoot).catch(() => null)).toBeNull();
vi.mocked(second.runner!.execute).mockRestore();
}
if (mode === "missing object") {
// Storage reports a lost saved object as 404; retrying cannot bring it back.
const lost: StorageService = { ...storage, getObject: async () => { throw notFound("Object not found"); } };
const missing = await prepareRemoteWorkspace({ db, storage: lost, workspace: reservedAgain, runId: nextRun, lease: newLease, target: second })
.then(() => null, (error: unknown) => error as RemoteWorkspaceRestoreError);
expect(missing).toBeInstanceOf(RemoteWorkspaceRestoreError);
expect(missing).toMatchObject({ step: "snapshot_download", retryable: false,
message: "remote_workspace_restore_failed: step=snapshot_download; remote_workspace_snapshot_object_missing" });
expect(await fs.lstat(expectedRemoteRoot).catch(() => null)).toBeNull();
}
const restored = await prepareRemoteWorkspace({ db, storage, workspace: reservedAgain, runId: nextRun, lease: newLease, target: second });
expect(restored.remoteRoot).toBe(expectedRemoteRoot);
if (mode !== "different root") expect(restored.remoteRoot).toBe(prepared.remoteRoot);
expect(await fs.readFile(path.join(restored.remoteRoot, "report.md"), "utf8")).toBe("useful saved result\n");
expect(await fs.readFile(path.join(restored.remoteRoot, "binary.bin"))).toEqual(Buffer.from([0, 1, 255]));
expect((await fs.stat(path.join(restored.remoteRoot, "empty-folder"))).isDirectory()).toBe(true);
const browse = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id, snapshotId: saved.snapshotId });
try { expect(await fs.readFile(path.join(browse.root, "report.md"), "utf8")).toContain("useful saved result"); } finally { await browse.cleanup(); }
}, 60_000);
it.each(["own retired root", "own restored root", "read-only folder", "receipt missing", "forged receipt", "later allocation",
"unsettled save", "newer checkpoint", "never saved here"] as const)(
"restores a newer save over this workspace's own retired root on a reused sandbox (%s)", async mode => {
// A parent prepares on its retained sandbox, then children inherit the
// workspace on other sandboxes and save newer work. When the parent wakes on
// the same sandbox, its old root is still present.
const reserved = await reserve({ kind: "empty" }), parentSandbox = await target("sandbox-A");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: parentSandbox });
await fs.writeFile(path.join(prepared.remoteRoot, "plan.md"), "parent plan\n");
// The parent's retained native session keeps its runner state beside the
// workspace files. Snapshots exclude it; it belongs to this sandbox.
const runnerStateDirectory = path.join(prepared.remoteRoot, ".paperclip-runtime", "paperclip-runner", "sessions",
createHash("sha256").update("parent-session").digest("hex"), "runner");
const runnerState = { lifecycle: "suspended", outbox: [], pendingTerminalDelivery: null, runId: ids.run };
await fs.mkdir(runnerStateDirectory, { recursive: true });
await fs.writeFile(path.join(runnerStateDirectory, "runner-state.json"), JSON.stringify(runnerState));
await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: parentSandbox, assertOwned: async () => undefined });
const writer = async (providerLeaseId: string) => {
const runId = randomUUID(), leaseId = randomUUID();
await db.insert(heartbeatRuns).values({ id: runId, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
await db.insert(environmentLeases).values({ id: leaseId, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: runId, issueId: ids.issue, provider: "daytona", providerLeaseId, status: "active" });
const workspace = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue, runId,
agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null, existingWorkspaceId: reserved.id, source: { kind: "empty" } });
const leased = (await db.select().from(environmentLeases).where(eq(environmentLeases.id, leaseId)))[0]! as EnvironmentLease;
return { runId, workspace, lease: leased };
};
const child = await writer("allocation-child"), childSandbox = await target("sandbox-B");
const childPrepared = await prepareRemoteWorkspace({ db, storage, workspace: child.workspace, runId: child.runId, lease: child.lease, target: childSandbox });
await fs.writeFile(path.join(childPrepared.remoteRoot, "child-result.md"), "child result\n");
const childSaved = await saveRemoteWorkspace({ db, storage, workspace: childPrepared.workspace, runId: child.runId,
lease: childPrepared.lease, target: childSandbox, assertOwned: async () => undefined });
if (mode === "receipt missing") {
// An occupant without a Paperclip preparation or restore receipt is unknown.
await fs.rm(path.join(prepared.remoteRoot, ".paperclip-runtime"), { recursive: true, force: true });
}
if (mode === "own restored root" || mode === "forged receipt") {
// The retired root was itself restored by an earlier preparation of this workspace.
const runtime = path.join(prepared.remoteRoot, ".paperclip-runtime");
await fs.rm(path.join(runtime, "workspace-ready.json"));
await fs.writeFile(path.join(runtime, "workspace-restored.json"), JSON.stringify(mode === "forged receipt"
? { identity: "not-an-identity" } : { identity: "a".repeat(64), manifestDigest: "b".repeat(64) }));
}
if (mode === "read-only folder") {
// A non-root sandbox user must still be able to remove the retired tree.
const locked = path.join(prepared.remoteRoot, "locked");
await fs.mkdir(path.join(locked, "inner"), { recursive: true });
await fs.writeFile(path.join(locked, "inner", "file.txt"), "old\n");
await fs.chmod(path.join(locked, "inner"), 0o555); await fs.chmod(locked, 0o555);
}
if (mode === "unsettled save" || mode === "newer checkpoint") {
// The last use of this root ended in a save that is not a settled final save.
await db.insert(workspaceOperations).values({ companyId: ids.company, executionWorkspaceId: reserved.id, heartbeatRunId: ids.run,
issueId: ids.issue, phase: "remote_workspace_snapshot", status: mode === "unsettled save" ? "running" : "succeeded",
metadata: { epoch: 1, boundary: "checkpoint", generation: 99,
saveWork: { schema: "paperclip.remote-workspace-save-work.v1", activeIds: [], unknown: false, quiescedAt: new Date().toISOString() },
exportCleanup: { ...readRemoteWorkspaceAuthority(prepared.workspace.metadata)!.location, state: "cleaned" } } });
}
if (mode === "later allocation") {
// A later writer used this root on the parent's sandbox after its last final save.
await db.insert(workspaceOperations).values({ companyId: ids.company, executionWorkspaceId: reserved.id, heartbeatRunId: child.runId,
issueId: ids.issue, phase: "remote_workspace_restore", status: "succeeded",
metadata: { epoch: 2, exportCleanup: { ...readRemoteWorkspaceAuthority(prepared.workspace.metadata)!.location, state: "cleaned" } } });
}
// Preparation narrows a target to its root; a later turn reattaches to the sandbox.
const reattach = () => ({ ...parentSandbox, remoteCwd: path.join(root, "sandbox-A") });
let parentTarget = reattach();
if (mode === "never saved here") {
// A different sandbox that holds an unknown folder at this path.
parentTarget = await target("sandbox-C");
await fs.cp(prepared.remoteRoot, path.join(parentTarget.remoteCwd, ".paperclip-workspaces", path.basename(prepared.remoteRoot)), { recursive: true });
}
const parent = await writer(mode === "never saved here" ? "allocation-unrelated" : prepared.lease.providerLeaseId!);
const restoring = prepareRemoteWorkspace({ db, storage, workspace: parent.workspace, runId: parent.runId, lease: parent.lease, target: parentTarget });
if (mode !== "own retired root" && mode !== "own restored root" && mode !== "read-only folder") {
const refused = await restoring.then(() => null, (error: unknown) => error as RemoteWorkspaceRestoreError);
expect(refused).toBeInstanceOf(RemoteWorkspaceRestoreError);
expect(refused).toMatchObject({ step: "sandbox_restore", retryable: false });
expect(refused!.message).toMatch(/stage=inspect_root reason=occupied_root$/);
expect(await fs.readFile(path.join(prepared.remoteRoot, "plan.md"), "utf8")).toBe("parent plan\n");
expect(await fs.lstat(path.join(prepared.remoteRoot, "child-result.md")).catch(() => null)).toBeNull();
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toEqual(readRemoteWorkspaceAuthority(parent.workspace.metadata));
return;
}
const restored = await restoring;
expect(restored.remoteRoot).toBe(prepared.remoteRoot);
expect(await fs.readFile(path.join(restored.remoteRoot, "plan.md"), "utf8")).toBe("parent plan\n");
expect(await fs.readFile(path.join(restored.remoteRoot, "child-result.md"), "utf8")).toBe("child result\n");
// The replaced root and the restore staging are gone; only saved files were replaced.
expect((await fs.readdir(path.dirname(restored.remoteRoot))).filter(name => name.startsWith(path.basename(restored.remoteRoot))))
.toEqual([path.basename(restored.remoteRoot)]);
expect(readRemoteWorkspaceAuthority(restored.workspace.metadata)).toMatchObject({ snapshotId: childSaved.snapshotId, dirty: true,
writerRunId: parent.runId, location: { providerLeaseId: prepared.lease.providerLeaseId, remoteCwd: prepared.remoteRoot } });
// The parent's continuation reads its retained runner state from the restored root.
expect(await readRemoteRunnerState({ runner: parentSandbox.runner!, stateDirectory: runnerStateDirectory })).toEqual(runnerState);
// Only the new restore receipt identifies the root; the retired preparation receipt is gone.
expect(await fs.lstat(path.join(restored.remoteRoot, ".paperclip-runtime", "workspace-ready.json")).catch(() => null)).toBeNull();
expect(JSON.parse(await fs.readFile(path.join(restored.remoteRoot, ".paperclip-runtime", "workspace-restored.json"), "utf8")))
.toMatchObject({ identity: expect.stringMatching(/^[a-f0-9]{64}$/) });
// A lost reply repeats the same preparation without replacing the restored root again.
await fs.writeFile(path.join(restored.remoteRoot, "parent-continued.md"), "unsaved parent edits\n");
// An attempt interrupted after publish left its retired root; the repeat keeps its runtime state.
const { identity } = JSON.parse(await fs.readFile(path.join(restored.remoteRoot, ".paperclip-runtime", "workspace-restored.json"), "utf8"));
const interrupted = `${restored.remoteRoot}.retired-${identity}`;
await fs.mkdir(path.join(interrupted, ".paperclip-runtime", "other-session"), { recursive: true });
await fs.writeFile(path.join(interrupted, ".paperclip-runtime", "other-session", "state.json"), "{}");
await fs.writeFile(path.join(interrupted, "stale.md"), "old\n");
const repeated = await prepareRemoteWorkspace({ db, storage, workspace: parent.workspace, runId: parent.runId, lease: parent.lease, target: reattach() });
expect(repeated.remoteRoot).toBe(prepared.remoteRoot);
expect(await fs.readFile(path.join(restored.remoteRoot, "parent-continued.md"), "utf8")).toBe("unsaved parent edits\n");
expect(await fs.readFile(path.join(restored.remoteRoot, ".paperclip-runtime", "other-session", "state.json"), "utf8")).toBe("{}");
expect(await fs.lstat(interrupted).catch(() => null)).toBeNull();
expect(await fs.lstat(path.join(restored.remoteRoot, "stale.md")).catch(() => null)).toBeNull();
expect(await readRemoteRunnerState({ runner: parentSandbox.runner!, stateDirectory: runnerStateDirectory })).toEqual(runnerState);
}, 60_000);
it.each(["per_turn", "warm", "occupied", "foreign-target", "foreign-allocation", "forged-receipt"])("preserves retained native workspace identity through migration: %s", async scenario => {
const original = path.join(root, "old-controller", "workspace"); await fs.mkdir(original, { recursive: true });
await fs.writeFile(path.join(original, "saved.txt"), "retained work");
const remote = await target("legacy-provider-root");
const [local] = await db.insert(executionWorkspaces).values({ companyId: ids.company, sourceIssueId: ids.issue,
mode: "isolated_workspace", strategyType: "project_primary", providerType: "local_fs", name: "Retained session",
cwd: original, providerRef: original }).returning();
const sessionId = randomUUID(), runnerInstanceId = randomUUID();
const execution = parseNativeExecutionInput({ schema: "paperclip.native-execution-input.v1",
provider: { kind: "codex", model: null }, binding: { companyId: ids.company, agentId: ids.agent,
issueId: ids.issue, runId: ids.run, executionWorkspaceId: local.id },
task: { identifier: "RWP-1", title: "Retained work", description: null, prompt: "Continue.", workMode: "standard" },
workspace: { cwd: original, repoUrl: null, repoRef: null, branchName: null },
session: { normalizedSessionId: sessionId, driverKind: "codex_app_server", protocolVersion: 1,
lifecyclePolicy: { mode: scenario === "warm" ? "warm" : "per_turn", idleTimeoutMs: scenario === "warm" ? 120000 : null } },
completionContract: { id: randomUUID(), sha256: "fixture", schemaVersion: "paperclip.completion-contract.v1",
contract: { revision: "1", objective: "Continue", criteria: [{ id: "objective", requirement: "Save work." }] } },
credentialBindings: [], interactionResponses: [],
});
if (execution.schema !== "paperclip.native-execution-input.v1") throw new Error("Unexpected fixture schema");
await db.update(heartbeatRuns).set({ status: "succeeded", finishedAt: new Date(), nativeSessionId: sessionId,
runnerInstanceId, runnerProfileJson: { nativeExecutionInput: execution, sessionCheckpoint: { sessionId }, nativeWorkspaceSync: {
schema: "paperclip.native-workspace-sync/v3", state: "finalized", descriptorSha256: "a".repeat(64), baselineSha256: "b".repeat(64),
finalHostSha256: "c".repeat(64), workspaceId: local.id, leaseId: ids.lease, providerLeaseId: `allocation-${ids.lease}`,
remoteCwd: remote.remoteCwd, resourceDisposition: "destroy",
} } }).where(eq(heartbeatRuns.id, ids.run));
await db.insert(agentTaskSessions).values({ companyId: ids.company, agentId: ids.agent, adapterType: "paperclip_runner",
taskKey: ids.issue, lastRunId: ids.run });
await db.update(issues).set({ executionWorkspaceId: local.id }).where(eq(issues.id, ids.issue));
const backupRoot = path.join(root, "private-state"), current = path.join(backupRoot, "failover-backups", "current");
await fs.mkdir(path.join(current, "runner"), { recursive: true }); await fs.mkdir(path.join(current, "codex-home"));
await fs.writeFile(path.join(current, "runner", "state.json"), "unchanged opaque provider state");
await fs.writeFile(path.join(current, "manifest.json"), JSON.stringify(buildNativeHarnessBackupManifest({ backupRoot: current,
execution, runnerInstanceId, sourceProviderLeaseId: "retained-allocation", providerSessionIdentity: {
providerSessionId: "retained-thread", providerBackendSessionId: "retained-backend", providerSessionIdentity: null } })));
expect(verifyNativeHarnessBackup({ root: backupRoot, execution, runnerInstanceId })).not.toBeNull();
await migrateRemoteWorkspace({ db, storage, companyId: ids.company, workspaceId: local.id, source: { kind: "empty" }, apply: true });
const migrated = await workspace(local.id);
const receipt = await readNativeWorkspaceMigration(db, migrated);
expect(receipt?.remoteCwd).toBe(remote.remoteCwd);
const continuation = { companyId: ids.company, workspaceId: local.id, agentId: ids.agent, issueId: ids.issue, sessionId, logicalCwd: original, remoteCwd: remote.remoteCwd };
expect(await canRestoreMigratedNativeSession(db, continuation)).toBe(true);
expect(await canRestoreMigratedNativeSession(db, { ...continuation, sessionId: randomUUID() })).toBe(false);
expect(await canRestoreMigratedNativeSession(db, { ...continuation, remoteCwd: "/foreign" })).toBe(false);
expect(migratedNativeExecutionCwd(receipt, { agentId: ids.agent, issueId: ids.issue, sessionId })).toBe(original);
expect(migratedNativeExecutionCwd(receipt, { agentId: randomUUID(), issueId: ids.issue, sessionId })).toBeNull();
await fs.rm(path.join(root, "old-controller"), { recursive: true });
ids.run = randomUUID();
await db.insert(heartbeatRuns).values({ id: ids.run, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
await db.update(environmentLeases).set({ heartbeatRunId: ids.run }).where(eq(environmentLeases.id, ids.lease));
const reserved = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue, runId: ids.run,
agentId: ids.agent, environmentId: ids.environment, existingWorkspaceId: local.id, projectId: null, projectWorkspaceId: null, source: { kind: "empty" } });
const preparation = { db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: remote };
const allocationPath = ".paperclip-runtime/reusable-sandbox-lease.json";
const allocationBytes = JSON.stringify({ version: 1, token: "owned-allocation", companyId: ids.company,
environmentId: ids.environment, agentId: ids.agent, executionWorkspaceId: local.id, provider: "daytona" });
await fs.mkdir(path.join(remote.remoteCwd, ".paperclip-runtime"));
await fs.writeFile(path.join(remote.remoteCwd, allocationPath), allocationBytes);
preparation.lease.metadata = { ...preparation.lease.metadata, workspaceAllocationFiles: {
providerLeaseId: preparation.lease.providerLeaseId, remoteCwd: remote.remoteCwd,
files: [{ relativePath: allocationPath, sha256: createHash("sha256").update(allocationBytes).digest("hex") }],
} };
if (scenario === "forged-receipt") {
await db.update(executionWorkspaces).set({ metadata: { ...reserved.metadata,
nativeWorkspaceMigration: { schema: "paperclip.native-workspace-migration.v1", operationId: randomUUID() } } }).where(eq(executionWorkspaces.id, local.id));
await expect(prepareRemoteWorkspace(preparation)).rejects.toThrow("native_workspace_migration_receipt_invalid"); return;
}
if (scenario === "occupied") {
await fs.writeFile(path.join(remote.remoteCwd, "unrelated.txt"), "do not overwrite");
await expect(prepareRemoteWorkspace(preparation)).rejects.toThrow();
expect(await fs.readFile(path.join(remote.remoteCwd, "unrelated.txt"), "utf8")).toBe("do not overwrite"); return;
}
if (scenario === "foreign-allocation") {
const receipt = preparation.lease.metadata.workspaceAllocationFiles as { providerLeaseId: string };
receipt.providerLeaseId = "another-allocation";
await expect(prepareRemoteWorkspace(preparation)).rejects.toThrow("remote_workspace_allocation_receipt_invalid");
expect(await fs.readFile(path.join(remote.remoteCwd, allocationPath), "utf8")).toBe(allocationBytes); return;
}
if (scenario === "foreign-target") {
remote.remoteCwd = path.join(root, "different-provider-root");
await expect(prepareRemoteWorkspace(preparation)).rejects.toThrow("native_workspace_migration_target_changed"); return;
}
const prepared = await prepareRemoteWorkspace(preparation);
expect(prepared.remoteRoot).toBe(remote.remoteCwd);
expect(await fs.readFile(path.join(prepared.remoteRoot, allocationPath), "utf8")).toBe(allocationBytes);
expect(await fs.readFile(path.join(prepared.remoteRoot, "saved.txt"), "utf8")).toBe("retained work");
const successor = { ...execution, workspace: { ...execution.workspace,
cwd: migratedNativeExecutionCwd(receipt, { agentId: ids.agent, issueId: ids.issue, sessionId }) ?? prepared.remoteRoot } };
expect(verifyNativeHarnessBackup({ root: backupRoot, execution: successor, runnerInstanceId })).not.toBeNull();
expect(verifyNativeHarnessBackup({ root: backupRoot, execution: { ...successor, workspace: { ...successor.workspace,
cwd: prepared.remoteRoot } }, runnerInstanceId })).toBeNull();
expect(await fs.readFile(path.join(current, "runner", "state.json"), "utf8")).toBe("unchanged opaque provider state");
}, 60_000);
it("cold restores a workspace and prepares its preserved pending repository through normal admission", async () => {
const original = path.join(root, "pending-original"), origin = path.join(root, "pending-origin.git");
await fs.mkdir(original); await fs.mkdir(origin);
await fs.writeFile(path.join(original, "saved.txt"), "saved before repository request\n");
const git = (cwd: string, ...args: string[]) => exec("git", ["-C", cwd, "-c", "user.name=Pending fixture",
"-c", "user.email=pending@example.invalid", ...args]);
await git(origin, "init", "--initial-branch=main");
await fs.writeFile(path.join(origin, "code.txt"), "acquired on next admission\n");
await git(origin, "add", "."); await git(origin, "commit", "-m", "Pending baseline");
await git(origin, "update-server-info");
const server = createServer(async (request, response) => {
const target = path.resolve(root, new URL(request.url!, "http://localhost").pathname.slice(1));
if (!target.startsWith(`${origin}/.git/`)) { response.writeHead(403).end(); return; }
try { response.end(await fs.readFile(target)); } catch { response.writeHead(404).end(); }
});
await new Promise<void>(resolve => server.listen(0, "127.0.0.1", resolve));
try {
const address = server.address(); if (!address || typeof address === "string") throw new Error("test server not bound");
const repoUrl = "https://github.com/paperclip-fixtures/pending";
const fixtureOrigin = `http://127.0.0.1:${address.port}/pending-origin.git/.git`, relativePath = taskRepositoryRelativePath(repoUrl);
await db.update(heartbeatRuns).set({ status: "succeeded", finishedAt: new Date() }).where(eq(heartbeatRuns.id, ids.run));
const [local] = await db.insert(executionWorkspaces).values({ companyId: ids.company, sourceIssueId: ids.issue,
mode: "isolated_workspace", strategyType: "project_primary", providerType: "local_fs", name: "Pending request",
cwd: original, providerRef: original }).returning();
await db.update(issues).set({ executionWorkspaceId: local.id }).where(eq(issues.id, ids.issue));
const [repository] = await db.insert(executionWorkspaceRepositories).values({ companyId: ids.company,
executionWorkspaceId: local.id, requestedByIssueId: ids.issue, repoUrl, repositoryIdentity: repoUrl,
relativePath, requestedRef: "HEAD", requestKey: "pending-request", requestKeys: ["pending-request"] }).returning();
vi.stubEnv("PAPERCLIP_EPHEMERAL_MIGRATION", "true");
await migrateRemoteWorkspace({ db, storage, companyId: ids.company, workspaceId: local.id, apply: true, writersStopped: true });
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.source).toEqual({ kind: "empty" });
expect(await db.select().from(executionWorkspaceRepositories).where(eq(executionWorkspaceRepositories.id, repository.id)))
.toEqual([repository]);
await fs.rm(original, { recursive: true });
vi.stubEnv("PAPERCLIP_EPHEMERAL_MIGRATION", "false");
const nextRun = randomUUID(), nextLease = randomUUID();
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue,
responsibleUserId: "local-board", runtimeMode: "native", status: "running" });
const [allocated] = await db.insert(environmentLeases).values({ id: nextLease, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: nextRun, issueId: ids.issue, provider: "daytona", providerLeaseId: `allocation-${nextLease}`, status: "active" }).returning();
const reserved = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue, runId: nextRun,
agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null,
existingWorkspaceId: local.id, source: { kind: "empty" } });
const remote = await target("pending-restored-sandbox");
const execute = remote.runner!.execute;
remote.runner!.execute = async input => {
const raw = input.args?.at(-1);
if (typeof raw === "string" && raw.startsWith("{")) {
const payload = JSON.parse(raw) as { repoUrl?: string; authArgs?: string[] };
if (payload.repoUrl?.startsWith(repoUrl) && Array.isArray(payload.authArgs)) input = { ...input,
args: [...input.args!.slice(0, -1), JSON.stringify({ ...payload,
authArgs: [...payload.authArgs, "-c", `url.${fixtureOrigin}.insteadOf=${payload.repoUrl}`] })] };
}
return execute(input);
};
const restored = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: nextRun,
lease: allocated as EnvironmentLease, target: remote });
expect(await fs.readFile(path.join(restored.remoteRoot, "saved.txt"), "utf8")).toBe("saved before repository request\n");
const secondUrl = `${repoUrl}-two`;
const secondPath = taskRepositoryRelativePath(secondUrl);
await db.insert(executionWorkspaceRepositories).values({ companyId: ids.company,
executionWorkspaceId: local.id, requestedByIssueId: ids.issue, repoUrl: secondUrl, repositoryIdentity: secondUrl,
relativePath: secondPath, requestedRef: "HEAD", requestKey: "second-request", requestKeys: ["second-request"] });
const admission = { companyId: ids.company, issueId: ids.issue, workspaceId: local.id, cwd: restored.remoteRoot,
agentId: ids.agent, runId: nextRun, responsibleUserId: "local-board", remoteTarget: remote, remoteLease: restored.lease };
const beforeAdmission = await db.select().from(workspaceOperations).where(and(
eq(workspaceOperations.executionWorkspaceId, local.id), eq(workspaceOperations.phase, "remote_workspace_snapshot")));
const prepared = await executionWorkspaceRepositoryService(db).prepareForAdmission({ companyId: ids.company,
issueId: ids.issue, workspaceId: local.id, cwd: restored.remoteRoot, agentId: ids.agent, runId: nextRun,
responsibleUserId: "local-board", remoteTarget: remote, remoteLease: restored.lease });
expect(prepared).toHaveLength(2);
expect(prepared).toEqual(expect.arrayContaining([expect.objectContaining({ id: repository.id, relativePath,
pinnedCommit: (await git(origin, "rev-parse", "HEAD")).stdout.trim() })]));
expect(await fs.readFile(path.join(restored.remoteRoot, secondPath, "code.txt"), "utf8")).toBe("acquired on next admission\n");
const afterAdmission = await db.select().from(workspaceOperations).where(and(
eq(workspaceOperations.executionWorkspaceId, local.id), eq(workspaceOperations.phase, "remote_workspace_snapshot")));
expect(afterAdmission).toHaveLength(beforeAdmission.length + 1);
const savedId = readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.snapshotId;
await executionWorkspaceRepositoryService(db).prepareForAdmission(admission);
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.snapshotId).toBe(savedId);
expect(await fs.readFile(path.join(restored.remoteRoot, relativePath, "code.txt"), "utf8")).toBe("acquired on next admission\n");
expect(await db.select().from(executionWorkspaceRepositories).where(eq(executionWorkspaceRepositories.id, repository.id)))
.toMatchObject([{ ...repository, state: "ready", pinnedCommit: prepared.find(row => row.id === repository.id)!.pinnedCommit,
branchName: prepared.find(row => row.id === repository.id)!.branchName, updatedAt: expect.any(Date) }]);
} finally { await new Promise<void>((resolve, reject) => server.close(error => error ? reject(error) : resolve())); }
}, 60_000);
it.each(["path", "file-url"] as const)("cold restores an imported secondary %s repository and acquires only a new descriptor", async sourceKind => {
const origin = path.join(root, "import-origin"), original = path.join(root, "old-checkout");
await fs.mkdir(origin); await fs.mkdir(original);
const git = (cwd: string, ...args: string[]) => exec("git", ["-C", cwd, "-c", "user.name=Import fixture", "-c", "user.email=import@example.invalid", ...args]);
await git(origin, "init", "--initial-branch=main");
await fs.writeFile(path.join(origin, "code.txt"), "saved secondary baseline\n");
await git(origin, "add", "."); await git(origin, "commit", "-m", "Import baseline");
const repoUrl = sourceKind === "path" ? origin : pathToFileURL(origin).href;
await git(original, "clone", repoUrl, "secondary");
await fs.writeFile(path.join(original, "secondary", "untracked.txt"), "imported secondary work\n");
const pinnedCommit = (await git(origin, "rev-parse", "HEAD")).stdout.trim();
const repository = { id: randomUUID(), relativePath: "secondary", repoUrl, ref: "HEAD", pinnedCommit };
const source = { kind: "empty" as const, repositories: [repository] };
await db.update(heartbeatRuns).set({ status: "succeeded", finishedAt: new Date() }).where(eq(heartbeatRuns.id, ids.run));
const [local] = await db.insert(executionWorkspaces).values({ companyId: ids.company, sourceIssueId: ids.issue,
mode: "isolated_workspace", strategyType: "project_primary", providerType: "local_fs", name: "Imported secondary",
cwd: original, providerRef: original }).returning();
await db.update(issues).set({ executionWorkspaceId: local.id }).where(eq(issues.id, ids.issue));
await migrateRemoteWorkspace({ db, storage, companyId: ids.company, workspaceId: local.id, source, apply: true });
const imported = readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!;
expect(imported.source).toEqual(source);
await fs.rm(original, { recursive: true }); await fs.rm(origin, { recursive: true });
// This descriptor was not in the committed import. It must still use normal
// acquisition rather than treating a missing directory as saved bytes.
const newOrigin = path.join(root, "new-origin"); await fs.mkdir(newOrigin);
await git(newOrigin, "init", "--initial-branch=main");
await fs.writeFile(path.join(newOrigin, "new.txt"), "newly acquired work\n");
await git(newOrigin, "add", "."); await git(newOrigin, "commit", "-m", "New source");
const added = { id: randomUUID(), relativePath: "new-repository", repoUrl: newOrigin, ref: "HEAD" };
const nextRun = randomUUID(), nextLease = randomUUID();
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue,
runtimeMode: "native", status: "running" });
await db.insert(environmentLeases).values({ id: nextLease, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: nextRun, issueId: ids.issue, provider: "daytona", providerLeaseId: `allocation-${nextLease}`, status: "active" });
const reserved = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue, runId: nextRun,
agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null,
existingWorkspaceId: local.id, source });
const next = (await db.select().from(environmentLeases).where(eq(environmentLeases.id, nextLease)))[0]! as EnvironmentLease;
const remote = await target("import-restored-sandbox");
const restored = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: nextRun, lease: next, target: remote,
source: { ...source, repositories: [repository, added] } });
const secondary = path.join(restored.remoteRoot, "secondary");
expect((await git(secondary, "rev-parse", "HEAD")).stdout.trim()).toBe(pinnedCommit);
expect((await git(secondary, "remote")).stdout.trim()).toBe("");
expect(await fs.readFile(path.join(secondary, "untracked.txt"), "utf8")).toBe("imported secondary work\n");
expect(await fs.readFile(path.join(restored.remoteRoot, added.relativePath, "new.txt"), "utf8")).toBe("newly acquired work\n");
expect(readRemoteWorkspaceAuthority(restored.workspace.metadata)).toMatchObject({ source, snapshotId: imported.snapshotId });
const resolveGitAuth = vi.fn(async () => null);
const admission = { db, storage, workspace: restored.workspace, runId: nextRun, lease: restored.lease, target: remote,
repository: { ...repository, ref: null }, resolveGitAuth };
// Registered admission encodes the stored default HEAD as null. It must
// verify the same saved descriptor without changing the source identity.
await expect(prepareRemoteWorkspaceRepository(admission)).resolves.toMatchObject({ cwd: secondary, pinnedCommit });
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.snapshotId).toBe(imported.snapshotId);
const beforeChanged = readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!;
await expect(prepareRemoteWorkspaceRepository({ ...admission,
workspace: await workspace(local.id), repository: { ...repository, ref: null, repoUrl: newOrigin } }))
.rejects.toThrow("remote_workspace_operation_failed");
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)).toEqual(beforeChanged);
expect(resolveGitAuth).not.toHaveBeenCalled();
const newRepository = { ...added, id: randomUUID(), relativePath: "wrapper-new", ref: null };
const failingStorage: StorageService = { ...storage, putFile: async () => { throw new Error("repository upload failed"); } };
await expect(prepareRemoteWorkspaceRepository({ ...admission, workspace: await workspace(local.id),
repository: newRepository, storage: failingStorage })).rejects.toThrow("repository upload failed");
await expect(prepareRemoteWorkspaceRepository({ ...admission, workspace: await workspace(local.id), repository: newRepository }))
.resolves.toMatchObject({ cwd: path.join(restored.remoteRoot, "wrapper-new") });
const durable = await workspace(local.id);
const durableId = readRemoteWorkspaceAuthority(durable.metadata)!.snapshotId;
await prepareRemoteWorkspaceRepository({ ...admission, workspace: durable,
repository: { ...newRepository, pinnedCommit: (await git(newOrigin, "rev-parse", "HEAD")).stdout.trim() } });
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.snapshotId).toBe(durableId);
await saveRemoteWorkspace({ ...admission, workspace: await workspace(local.id), boundary: "repository-admission-batch",
assertOwned: async () => undefined,
repositories: [{ ...newRepository, pinnedCommit: (await git(newOrigin, "rev-parse", "HEAD")).stdout.trim() }] });
const afterBatch = await workspace(local.id);
await expect(prepareRemoteWorkspaceRepository({ ...admission, workspace: afterBatch }))
.resolves.toMatchObject({ cwd: secondary, pinnedCommit });
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.snapshotId)
.toBe(readRemoteWorkspaceAuthority(afterBatch.metadata)!.snapshotId);
expect(resolveGitAuth).toHaveBeenCalledOnce();
await fs.rm(path.join(restored.remoteRoot, "wrapper-new"), { recursive: true });
await saveRemoteWorkspace({ ...admission, workspace: await workspace(local.id), boundary: "after-checkout-removal",
assertOwned: async () => undefined });
const removed = await workspace(local.id);
expect((removed.metadata?.remoteWorkspaceSnapshotRepositories as { repositories: Array<{ id: string }> }).repositories)
.not.toEqual(expect.arrayContaining([expect.objectContaining({ id: newRepository.id })]));
await expect(prepareRemoteWorkspaceRepository({ ...admission, workspace: removed, repository: newRepository }))
.resolves.toMatchObject({ cwd: path.join(restored.remoteRoot, "wrapper-new") });
expect(await fs.readFile(path.join(restored.remoteRoot, "wrapper-new", "new.txt"), "utf8")).toBe("newly acquired work\n");
expect(readRemoteWorkspaceAuthority((await workspace(local.id)).metadata)!.source).toEqual(source);
}, 60_000);
it.runIf(process.platform !== "win32" && process.getuid?.() !== 0)("cleans saved reads with read-only folders without changing modes or following symlinks", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("read-only-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const locked = path.join(prepared.remoteRoot, "locked"), outside = path.join(root, "outside");
await fs.mkdir(locked); await fs.writeFile(path.join(locked, "report.txt"), "saved read-only result\n", { mode: 0o444 });
await fs.mkdir(outside); await fs.writeFile(path.join(outside, "keep.txt"), "outside the extraction\n", { mode: 0o444 });
await fs.chmod(locked, 0o555); await fs.chmod(outside, 0o555);
let extraction: Awaited<ReturnType<ReturnType<typeof createSavedWorkspaceLoader>>> | undefined;
try {
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
const savedTemp = path.join(root, "saved-reads"); await fs.mkdir(savedTemp); vi.stubEnv("TMPDIR", savedTemp);
const loadSavedWorkspace = createSavedWorkspaceLoader(db, storage);
extraction = await loadSavedWorkspace({ companyId: ids.company, workspaceId: reserved.id, snapshotId: saved.snapshotId });
expect((await fs.stat(path.join(extraction.root, "locked"))).mode & 0o777).toBe(0o555);
expect((await fs.stat(path.join(extraction.root, "locked", "report.txt"))).mode & 0o777).toBe(0o444);
expect(await fs.readFile(path.join(extraction.root, "locked", "report.txt"), "utf8")).toBe("saved read-only result\n");
// Cleanup must unlink a symlink without changing its external target.
await fs.symlink(outside, path.join(extraction.root, "outside-link"));
await extraction.cleanup();
await expect(fs.access(path.dirname(extraction.root))).rejects.toMatchObject({ code: "ENOENT" });
expect((await fs.stat(outside)).mode & 0o777).toBe(0o555);
expect((await fs.stat(path.join(outside, "keep.txt"))).mode & 0o777).toBe(0o444);
expect(await fs.readFile(path.join(outside, "keep.txt"), "utf8")).toBe("outside the extraction\n");
expect((await fs.stat(locked)).mode & 0o777).toBe(0o555);
const current = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(current.metadata)!;
await db.update(executionWorkspaces).set({ metadata: { ...current.metadata, remoteWorkspace: { ...authority, location: null } } })
.where(eq(executionWorkspaces.id, reserved.id));
const files = workspaceFileResourceService(db, undefined, { remoteFiles: createRemoteWorkspaceFiles(db, { loadSavedWorkspace }) });
expect(await files.readContent(ids.issue, { workspace: "execution", path: "locked/report.txt" }))
.toMatchObject({ content: { data: "saved read-only result\n" }, resource: { source: { kind: "saved" } } });
expect(await fs.readdir(savedTemp)).toEqual([]);
expect(await files.list(ids.issue, { workspace: "execution", mode: "all" })).toMatchObject({ state: "available" });
expect(await fs.readdir(savedTemp)).toEqual([]);
await expect(files.readContent(ids.issue, { workspace: "execution", path: "missing.txt" })).rejects.toMatchObject({ status: 404 });
expect(await fs.readdir(savedTemp)).toEqual([]);
} finally {
if (extraction) {
await fs.chmod(path.join(extraction.root, "locked"), 0o700).catch(() => undefined);
await fs.rm(path.dirname(extraction.root), { recursive: true, force: true });
}
await fs.chmod(locked, 0o700); await fs.chmod(outside, 0o700);
}
}, 60_000);
it("forwards cancellation through saved file access and removes failed extraction scratch", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("cancelled-saved-read");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "report.txt"), "saved result\n");
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
const current = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(current.metadata)!;
await db.update(executionWorkspaces).set({ metadata: { ...current.metadata, remoteWorkspace: { ...authority, location: null } } })
.where(eq(executionWorkspaces.id, reserved.id));
const savedTemp = path.join(root, "cancelled-reads"); await fs.mkdir(savedTemp); vi.stubEnv("TMPDIR", savedTemp);
for (const mode of ["abort", "storage_failure"] as const) {
const controller = new AbortController(), failure = new Error("Saved object unavailable");
let observedSignal: AbortSignal | undefined;
const interruptedStorage: StorageService = { ...storage, getObject: async (companyId, objectKey, options) => {
observedSignal = options?.signal;
if (mode === "storage_failure") throw failure;
const object = await storage.getObject(companyId, objectKey, options);
controller.abort(new DOMException("Saved read cancelled", "AbortError"));
return object;
} };
const remoteFiles = createRemoteWorkspaceFiles(db, { loadSavedWorkspace: createSavedWorkspaceLoader(db, interruptedStorage) });
let readError: unknown;
const access = await remoteFiles.open({ companyId: ids.company, workspaceId: reserved.id, signal: controller.signal })
.catch(error => { readError = error; return null; });
await access?.close();
expect(observedSignal).toBe(controller.signal);
if (mode === "storage_failure") expect(readError).toBe(failure);
else expect(readError).toMatchObject({ name: "AbortError" });
expect(await fs.readdir(savedTemp)).toEqual([]);
}
const cancelled = new DOMException("Already cancelled", "AbortError");
const getObject = vi.fn(storage.getObject.bind(storage));
await expect(createSavedWorkspaceLoader(db, { ...storage, getObject })({ companyId: ids.company,
workspaceId: reserved.id, snapshotId: saved.snapshotId, signal: AbortSignal.abort(cancelled) })).rejects.toBe(cancelled);
expect(getObject).not.toHaveBeenCalled();
expect(await fs.readdir(savedTemp)).toEqual([]);
}, 60_000);
it("reconnects to the same live folder without replaying preparation or overwriting unsaved edits", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("same-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "in-flight.txt"), "still running");
const before = await workspace(reserved.id);
const reattached = await prepareRemoteWorkspace({ db, storage, workspace: before, runId: ids.run, lease: await lease(), target: first,
provisionCommand: "exit 99" });
expect(reattached.remoteRoot).toBe(prepared.remoteRoot);
expect(await fs.readFile(path.join(reattached.remoteRoot, "in-flight.txt"), "utf8")).toBe("still running");
expect((await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.companyId, ids.company))).length).toBe(1);
}, 60_000);
it("fences stale same-run preparation from replacing a dirty workspace allocation or lease", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("dirty-owner");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run,
lease: await lease(), target: first, assertOwned: async () => undefined });
await fs.writeFile(path.join(prepared.remoteRoot, "unsaved.txt"), "only the first allocation has this work\n");
const originalAuthority = readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)!;
for (const sameAllocation of [false, true]) {
const leaseId = randomUUID();
const providerLeaseId = sameAllocation ? prepared.lease.providerLeaseId! : `allocation-${leaseId}`;
await db.insert(environmentLeases).values({ id: leaseId, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: ids.run, issueId: ids.issue, provider: "daytona", providerLeaseId, status: "active" });
const replacementLease = (await db.select().from(environmentLeases).where(eq(environmentLeases.id, leaseId)))[0]! as EnvironmentLease;
const replacement = sameAllocation ? first : await target("stale-replacement");
await expect(prepareRemoteWorkspace({ db, storage, workspace: sameAllocation ? await workspace(reserved.id) : reserved, runId: ids.run,
lease: replacementLease, target: replacement, assertOwned: async () => undefined }))
.rejects.toThrow("remote_workspace_authority_changed");
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toEqual(originalAuthority);
expect(await fs.readFile(path.join(prepared.remoteRoot, "unsaved.txt"), "utf8"))
.toBe("only the first allocation has this work\n");
}
}, 60_000);
it("publishes a cold restored workspace despite lost cleanup replies and retries cleanup without overwriting later edits", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("restore-cleanup-source");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run,
lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "saved.txt"), "saved before the original allocation was lost\n");
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
await fs.rm(path.dirname(prepared.remoteRoot), { recursive: true, force: true });
const nextRun = randomUUID(), nextLease = randomUUID(), providerLeaseId = `allocation-${nextLease}`;
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent,
nativeIssueId: ids.issue, runtimeMode: "native", status: "running" });
await db.insert(environmentLeases).values({ id: nextLease, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: nextRun, issueId: ids.issue, provider: "daytona", providerLeaseId, status: "active" });
const nextWorkspace = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue,
runId: nextRun, agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null,
existingWorkspaceId: reserved.id, source: { kind: "empty" } });
const nextTarget = await target("restore-cleanup-destination");
const sandbox = nextTarget as Extract<AdapterExecutionTarget, { transport: "sandbox" }>;
const expectedRemoteRoot = path.join(sandbox.remoteCwd, ".paperclip-workspaces", path.basename(prepared.remoteRoot));
const execute = sandbox.runner!.execute.bind(sandbox.runner!);
let failCleanup = true, cleanupAttempts = 0;
sandbox.runner!.execute = async input => {
const result = await execute(input);
if (input.args?.[1]?.includes("remote_workspace_export_cleanup_failed") && result.exitCode === 0) {
cleanupAttempts++;
if (failCleanup) throw new Error("lost restore cleanup reply");
}
return result;
};
const nextLeaseRow = (await db.select().from(environmentLeases).where(eq(environmentLeases.id, nextLease)))[0]! as EnvironmentLease;
const restored = await prepareRemoteWorkspace({ db, storage, workspace: nextWorkspace, runId: nextRun,
lease: nextLeaseRow, target: nextTarget, assertOwned: async () => undefined });
expect(restored.remoteRoot).toBe(expectedRemoteRoot);
const location = { environmentId: ids.environment, leaseId: nextLease, providerLeaseId, remoteCwd: restored.remoteRoot };
expect(cleanupAttempts).toBeGreaterThan(0);
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toMatchObject({
location, snapshotId: saved.snapshotId, writerRunId: nextRun, dirty: true,
});
expect((await db.select().from(environmentLeases).where(eq(environmentLeases.id, nextLease)))[0]?.metadata?.remoteCwd).toBe(restored.remoteRoot);
expect((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, nextRun)))[0]?.runnerProfileJson?.remoteWorkspace)
.toMatchObject({ workspaceId: reserved.id, leaseId: nextLease, providerLeaseId,
remoteCwd: restored.remoteRoot, state: "prepared", snapshotId: saved.snapshotId });
expect(await fs.readFile(path.join(restored.remoteRoot, "saved.txt"), "utf8"))
.toBe("saved before the original allocation was lost\n");
const [operation] = await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.companyId, ids.company),
eq(workspaceOperations.heartbeatRunId, nextRun), eq(workspaceOperations.phase, "remote_workspace_restore")));
expect(operation).toMatchObject({ status: "succeeded", metadata: { exportCleanup: { ...location, state: "pending" } } });
await fs.writeFile(path.join(restored.remoteRoot, "saved.txt"), "later agent edits\n");
await fs.writeFile(path.join(restored.remoteRoot, "newer.txt"), "keep untracked work\n");
failCleanup = false;
const transfer = vi.spyOn(sandbox.runner!, "syncIn");
const upload = vi.spyOn(storage, "putFile");
await prepareRemoteWorkspace({ db, storage, workspace: await workspace(reserved.id), runId: nextRun,
lease: restored.lease, target: nextTarget, assertOwned: async () => undefined });
expect((await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, operation!.id)))[0])
.toMatchObject({ status: "succeeded", metadata: { exportCleanup: { ...location, state: "cleaned" } } });
expect(await fs.readFile(path.join(restored.remoteRoot, "saved.txt"), "utf8")).toBe("later agent edits\n");
expect(await fs.readFile(path.join(restored.remoteRoot, "newer.txt"), "utf8")).toBe("keep untracked work\n");
expect(transfer).not.toHaveBeenCalled();
expect(upload).not.toHaveBeenCalled();
}, 60_000);
it.each([
{ mode: "shared_workspace", inherited: false },
{ mode: "shared_workspace", inherited: true },
{ mode: "isolated_workspace", inherited: true },
{ mode: "operator_branch", inherited: true },
])("continues saved $mode files while fencing a second writer (inherited=$inherited)", async ({ mode, inherited }) => {
await instanceSettingsService(db).updateExperimental({ enableIsolatedWorkspaces: true });
const projectId = randomUUID(), sourceId = randomUUID();
await db.insert(projects).values({ id: projectId, companyId: ids.company, name: "Synthetic shared project" });
await db.insert(projectWorkspaces).values({ id: sourceId, companyId: ids.company, projectId,
name: "Shared files", sourceType: "empty", sharedWorkspaceKey: "synthetic-shared" });
await db.update(issues).set({ projectId, projectWorkspaceId: sourceId }).where(eq(issues.id, ids.issue));
const input = { db, storage, companyId: ids.company, issueId: ids.issue, runId: ids.run,
agentId: ids.agent, environmentId: ids.environment, projectId, projectWorkspaceId: sourceId,
existingWorkspaceId: null, source: { kind: "empty" as const }, mode };
const first = await target("shared-sandbox"), reserved = await reserveRemoteWorkspace(input);
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run,
lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "shared-result.txt"), "from the first task\n");
const childInput = { title: "Continue saved work", status: "in_progress" as const, assigneeAgentId: ids.agent };
const child = inherited ? await issueService(db).create(ids.company, { ...childInput, parentId: ids.issue })
: (await db.insert(issues).values({ ...childInput, companyId: ids.company, projectId, projectWorkspaceId: sourceId }).returning())[0]!;
if (inherited) expect(child.executionWorkspaceId).toBe(reserved.id);
const secondIssue = child.id, secondRun = randomUUID(), secondLease = randomUUID();
await db.insert(heartbeatRuns).values({ id: secondRun, companyId: ids.company, agentId: ids.agent,
nativeIssueId: secondIssue, runtimeMode: "native", status: "running" });
const next = { ...input, issueId: secondIssue, runId: secondRun,
existingWorkspaceId: mode === "shared_workspace" ? null : child.executionWorkspaceId };
await expect(reserveRemoteWorkspace(next)).rejects.toThrow("unfinished writer");
await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
await db.update(heartbeatRuns).set({ status: "succeeded" }).where(eq(heartbeatRuns.id, ids.run));
if (inherited) await fs.rm(path.dirname(prepared.remoteRoot), { recursive: true, force: true });
const reused = await reserveRemoteWorkspace(next);
expect(reused.id).toBe(reserved.id);
expect(readRemoteWorkspaceAuthority(reused.metadata)).toMatchObject({ epoch: 2, writerRunId: secondRun });
await db.insert(environmentLeases).values({ id: secondLease, companyId: ids.company, environmentId: ids.environment,
heartbeatRunId: secondRun, issueId: secondIssue, provider: "daytona",
providerLeaseId: `allocation-${inherited ? secondLease : ids.lease}`, status: "active" });
const newLease = (await db.select().from(environmentLeases).where(eq(environmentLeases.id, secondLease)))[0]! as EnvironmentLease;
const nextTarget = inherited ? { ...await target("child-sandbox"), leaseId: secondLease,
sandboxId: newLease.providerLeaseId! } : first;
const continued = await prepareRemoteWorkspace({ db, storage, workspace: reused, runId: secondRun, lease: newLease, target: nextTarget,
runtimeProvisionCommand: "printf 'continued effect\\n' >> continued-provision.txt" });
expect(await fs.readFile(path.join(continued.remoteRoot, "continued-provision.txt"), "utf8")).toBe("continued effect\n");
const [provisioned] = await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.heartbeatRunId, secondRun),
eq(workspaceOperations.phase, "remote_workspace_provision")));
expect(provisioned).toMatchObject({ issueId: secondIssue, status: "succeeded" });
expect(await fs.readFile(path.join(continued.remoteRoot, "shared-result.txt"), "utf8")).toBe("from the first task\n");
await fs.appendFile(path.join(continued.remoteRoot, "shared-result.txt"), "from the child task\n");
const childSaved = await saveRemoteWorkspace({ db, storage, workspace: continued.workspace, runId: secondRun,
lease: continued.lease, target: nextTarget, assertOwned: async () => undefined });
const loaded = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id,
snapshotId: childSaved.snapshotId });
try { expect(await fs.readFile(path.join(loaded.root, "shared-result.txt"), "utf8")).toBe("from the first task\nfrom the child task\n"); }
finally { await loaded.cleanup(); }
const [operation] = await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.heartbeatRunId, secondRun),
eq(workspaceOperations.phase, "remote_workspace_snapshot")));
expect(operation?.issueId).toBe(secondIssue);
}, 60_000);
it("durably saves a full sandbox and restores its work after losing the original allocation", async () => {
const reserved = await reserve({ kind: "empty" }), remote = await target("full-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: remote });
await fs.writeFile(path.join(prepared.remoteRoot, "saved.txt"), "completed work before restart\n");
const execute = remote.runner!.execute.bind(remote.runner);
let fullDiskFailures = 0;
vi.spyOn(remote.runner!, "execute").mockImplementation(input => {
if (input.args?.some(arg => arg.endsWith("/archive.cjs"))) {
fullDiskFailures++;
return Promise.resolve({ exitCode: 1, stdout: "", stderr: "remote_workspace_sandbox_disk_full: the sandbox disk is full; free space in the sandbox before saving the workspace",
timedOut: false, signal: null, pid: process.pid, startedAt: new Date().toISOString() });
}
return execute(input);
});
const syncOut = remote.runner!.syncOut!;
remote.runner!.syncOut = async (operations, options) => {
if (!operations.some(operation => operation.files.some(file => file.kind === "directory"))) return syncOut(operations, options);
for (const operation of operations) for (const mapping of operation.files) {
await fs.cp(mapping.sourcePath, mapping.targetPath, { recursive: true, verbatimSymlinks: true,
filter: source => !(mapping.exclude ?? []).some(excluded => {
const relative = path.relative(mapping.sourcePath, source).split(path.sep).join("/");
return relative === excluded || relative.startsWith(`${excluded}/`);
}) });
}
return { operations: [] };
};
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: remote, boundary: "final", assertOwned: async () => undefined });
expect(fullDiskFailures).toBe(1);
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toMatchObject({
snapshotId: saved.snapshotId, dirty: false, writerRunId: null,
});
const [snapshot] = await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, saved.snapshotId));
expect(snapshot).toMatchObject({ status: "committed" });
await fs.rm(remote.remoteCwd, { recursive: true, force: true });
const load = createSavedWorkspaceLoader(db, storage);
for (let reconnect = 0; reconnect < 2; reconnect++) {
const restored = await load({ companyId: ids.company, workspaceId: reserved.id, snapshotId: saved.snapshotId });
try { expect(await fs.readFile(path.join(restored.root, "saved.txt"), "utf8")).toBe("completed work before restart\n"); }
finally { await restored.cleanup(); }
}
const operations = await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.heartbeatRunId, ids.run),
eq(workspaceOperations.phase, "remote_workspace_snapshot")));
expect(operations.at(-1)).toMatchObject({ status: "succeeded" });
}, 60_000);
it("retains the writer and previous snapshot when object storage fails", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("outage-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "unpublished.txt"), "retain this source");
const failing: StorageService = { ...storage, putFile: async () => { throw new Error("storage offline"); } };
await expect(saveRemoteWorkspace({ db, storage: failing, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined })).rejects.toThrow("storage offline");
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toMatchObject({ dirty: true, writerRunId: ids.run,
snapshotId: readRemoteWorkspaceAuthority(prepared.workspace.metadata)!.snapshotId });
expect((await fs.readdir(path.join(prepared.remoteRoot, ".paperclip-runtime", "workspace-snapshots"))).length).toBe(1);
// Each settled save records where its time went: the failed upload keeps the
// phases it finished, and the retry of the same boundary records every phase.
const finalTiming = async () => (await db.select().from(workspaceOperations).where(and(
eq(workspaceOperations.executionWorkspaceId, reserved.id), eq(workspaceOperations.phase, "remote_workspace_snapshot"),
sql`${workspaceOperations.metadata}->>'boundary' = 'final'`))).map(save => save.metadata?.saveTiming);
const duration = expect.any(Number), archive = { byteSize: expect.any(Number), entries: expect.any(Number) };
expect(await finalTiming()).toEqual([{ schema: "paperclip.remote-workspace-save-timing.v1", outcome: "failed", totalMs: duration,
captureMs: duration, transferMs: duration, verifyMs: duration, ...archive, at: expect.any(String) }]);
const beforeSuccess = await workspace(reserved.id);
const saved = await saveRemoteWorkspace({ db, storage, workspace: beforeSuccess, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
expect(saved.generation).toBe(2);
const [succeededTiming] = await finalTiming();
// A lost-reply replay of the committed save keeps the measured record.
await expect(saveRemoteWorkspace({ db, storage, workspace: beforeSuccess, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined })).resolves.toMatchObject({ snapshotId: saved.snapshotId });
expect(await finalTiming()).toEqual([succeededTiming]);
expect(await finalTiming()).toEqual([{ schema: "paperclip.remote-workspace-save-timing.v1", outcome: "succeeded", totalMs: duration,
captureMs: duration, transferMs: duration, verifyMs: duration, uploadMs: duration, publishMs: duration, ...archive,
at: expect.any(String) }]);
expect(await fs.readdir(path.join(prepared.remoteRoot, ".paperclip-runtime", "workspace-snapshots"))).toEqual([]);
}, 60_000);
it.each(["initial", "repository:additional"])("saves current files through a different repair boundary after a failed %s upload", async boundary => {
if (boundary === "initial") await db.update(heartbeatRuns).set({ runtimeMode: "legacy", nativeIssueId: null,
issueId: null, contextSnapshot: { issueId: ids.issue } }).where(eq(heartbeatRuns.id, ids.run));
const reserved = await reserve({ kind: "empty" }), first = await target("failed-boundary");
const failing: StorageService = { ...storage, putFile: async upload => {
await storage.putFile(upload);
throw new Error("boundary storage offline");
} };
if (boundary === "initial") {
await expect(prepareRemoteWorkspace({ db, storage: failing, workspace: reserved, runId: ids.run,
lease: await lease(), target: first, assertOwned: async () => undefined,
provisionCommand: "printf 'one preparation effect\\n' > prepared.txt" })).rejects.toThrow("boundary storage offline");
} else {
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "prepared.txt"), "one preparation effect\n");
await expect(saveRemoteWorkspace({ db, storage: failing, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, boundary, assertOwned: async () => undefined })).rejects.toThrow("boundary storage offline");
}
const before = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(before.metadata)!;
const [pending] = await db.select().from(executionWorkspaceSnapshots).where(and(
eq(executionWorkspaceSnapshots.executionWorkspaceId, reserved.id), eq(executionWorkspaceSnapshots.status, "pending")));
expect(pending).toBeTruthy();
expect(authority.snapshotId).toBe(pending!.parentSnapshotId);
await fs.writeFile(path.join(authority.location!.remoteCwd, "after-failed-save.txt"), "newer live files must survive\n");
await db.update(heartbeatRuns).set({ status: "failed" }).where(eq(heartbeatRuns.id, ids.run));
const saved = await saveRemoteWorkspace({ db, storage, workspace: before, runId: ids.run,
lease: await lease(), target: first, assertOwned: async () => undefined });
expect(saved.generation).toBe(pending!.generation + 1);
const [committed] = await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, saved.snapshotId));
expect(committed).toMatchObject({ status: "committed", parentSnapshotId: authority.snapshotId });
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toMatchObject({
dirty: false, writerRunId: null, snapshotId: saved.snapshotId,
});
const [retired] = await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, pending!.sourceOperationId));
expect(retired).toMatchObject({ status: "failed", metadata: { generation: pending!.generation,
exportCleanup: { state: "cleaned" } } });
expect(await fs.readdir(path.join(authority.location!.remoteCwd, ".paperclip-runtime", "workspace-snapshots"))).toEqual([]);
const upload = vi.spyOn(storage, "putFile"), execute = vi.spyOn((first as Extract<AdapterExecutionTarget, { transport: "sandbox" }>).runner!, "execute");
try {
await expect(saveRemoteWorkspace({ db, storage, workspace: before, runId: ids.run,
lease: await lease(), target: first, boundary, assertOwned: async () => undefined })).rejects.toThrow("remote_workspace_snapshot_retired");
expect(upload).not.toHaveBeenCalled(); expect(execute).not.toHaveBeenCalled();
} finally { upload.mockRestore(); execute.mockRestore(); }
expect(await fs.readdir(path.join(authority.location!.remoteCwd, ".paperclip-runtime", "workspace-snapshots"))).toEqual([]);
const abandonedUploads = pending!.metadata!.uploads as Array<{ objectKey: string }>;
expect(abandonedUploads).toHaveLength(1);
expect((await storage.headObject(ids.company, abandonedUploads[0]!.objectKey)).exists).toBe(true);
const retention = remoteWorkspaceSnapshotRetentionService(db, { storage,
now: () => new Date(Date.now() + 8 * 86_400_000) });
await retention.tick();
expect((await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, pending!.id)))[0]?.status)
.toBe("deleted");
expect((await storage.headObject(ids.company, abandonedUploads[0]!.objectKey)).exists).toBe(false);
expect((await storage.headObject(ids.company, committed!.objectKey!)).exists).toBe(true);
const extracted = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id, snapshotId: saved.snapshotId });
try {
expect(await fs.readFile(path.join(extracted.root, "prepared.txt"), "utf8")).toBe("one preparation effect\n");
expect(await fs.readFile(path.join(extracted.root, "after-failed-save.txt"), "utf8")).toBe("newer live files must survive\n");
} finally { await extracted.cleanup(); }
}, 60_000);
it.each(["journal", "ledger"])("respects historical %s generation reservations without changing the committed parent", async highest => {
const reserved = await reserve({ kind: "empty" }), first = await target("reserved-generations");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const parent = readRemoteWorkspaceAuthority(prepared.workspace.metadata)!.snapshotId;
// A crash after journaling can leave no ledger row; retention can leave a
// deleted ledger after the source operation is gone. Both remain reserved.
await db.insert(workspaceOperations).values({ companyId: ids.company, executionWorkspaceId: reserved.id,
heartbeatRunId: ids.run, issueId: ids.issue, phase: "remote_workspace_snapshot", status: "running",
metadata: { boundary: "interrupted-before-ledger", epoch: 0, generation: highest === "journal" ? 5 : 3 } });
await db.insert(executionWorkspaceSnapshots).values({ companyId: ids.company, executionWorkspaceId: reserved.id,
sourceOperationId: randomUUID(), sourceRunId: ids.run, parentSnapshotId: null,
generation: highest === "ledger" ? 5 : 3, epoch: 0, status: "deleted" });
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined });
expect(saved.generation).toBe(6);
expect((await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, saved.snapshotId)))[0])
.toMatchObject({ status: "committed", parentSnapshotId: parent });
}, 60_000);
it("holds an already-poisoned pre-release reservation without rewriting its identity or uploading", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("old-poisoned-reservation");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const authority = readRemoteWorkspaceAuthority(prepared.workspace.metadata)!;
await fs.writeFile(path.join(prepared.remoteRoot, "held-work.txt"), "retain exact existing files\n");
const failing: StorageService = { ...storage, putFile: async () => { throw new Error("old upload failed"); } };
await expect(saveRemoteWorkspace({ db, storage: failing, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, boundary: "repository:old", assertOwned: async () => undefined }))
.rejects.toThrow("old upload failed");
const [pending] = await db.select().from(executionWorkspaceSnapshots).where(and(
eq(executionWorkspaceSnapshots.executionWorkspaceId, reserved.id), eq(executionWorkspaceSnapshots.status, "pending")));
// Old pre-release code could journal a conflicting final operation before
// beginWorkspaceSnapshot rejected it. Its identity must not be repurposed.
const [poisoned] = await db.insert(workspaceOperations).values({ companyId: ids.company, executionWorkspaceId: reserved.id,
heartbeatRunId: ids.run, issueId: ids.issue, phase: "remote_workspace_snapshot", status: "running",
metadata: { boundary: "final", epoch: authority.epoch, generation: pending!.generation,
parentSnapshotId: authority.snapshotId, exportCleanup: { ...authority.location, state: "pending" } } }).returning();
const upload = vi.spyOn(storage, "putFile");
await expect(saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run,
lease: prepared.lease, target: first, assertOwned: async () => undefined }))
.rejects.toThrow("remote_workspace_snapshot_operation_binding_changed");
expect(upload).not.toHaveBeenCalled();
const [held] = await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, poisoned!.id));
const { saveWork, saveTiming, ...originalMetadata } = held!.metadata!;
expect(originalMetadata).toEqual(poisoned!.metadata);
expect(saveWork).toMatchObject({ activeIds: [], unknown: true, quiescedAt: expect.any(String) });
// The refused invocation records only its own failed duration; it never captured or uploaded.
expect(saveTiming).toEqual({ schema: "paperclip.remote-workspace-save-timing.v1", outcome: "failed",
totalMs: expect.any(Number), at: expect.any(String) });
expect(Number.isFinite(Date.parse((saveWork as { quiescedAt: string }).quiescedAt))).toBe(true);
expect(await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.sourceOperationId, poisoned!.id)))
.toHaveLength(0);
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)).toMatchObject({
dirty: true, writerRunId: ids.run, snapshotId: authority.snapshotId,
});
expect(await fs.readFile(path.join(prepared.remoteRoot, "held-work.txt"), "utf8")).toBe("retain exact existing files\n");
}, 60_000);
it("reserves distinct generations for competing boundaries and fences the losing publication", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("competing-boundaries");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const parent = readRemoteWorkspaceAuthority(prepared.workspace.metadata)!.snapshotId;
let arrivals = 0, release!: () => void;
const gate = new Promise<void>(resolve => { release = resolve; });
const timeout = setTimeout(release, 10_000);
const competingStorage: StorageService = { ...storage, putFile: async upload => {
if (++arrivals === 2) release();
await gate;
return storage.putFile(upload);
} };
const input = { db, storage: competingStorage, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease,
target: first, assertOwned: async () => undefined };
const boundaries = ["repository:one", "repository:two"];
const results = await Promise.allSettled(boundaries.map(boundary => saveRemoteWorkspace({ ...input, boundary })))
.finally(() => { clearTimeout(timeout); release(); });
expect(arrivals).toBe(2);
expect(results.filter(result => result.status === "fulfilled")).toHaveLength(1);
const failed = results.find(result => result.status === "rejected") as PromiseRejectedResult;
expect(failed.reason).toMatchObject({ message: "remote_workspace_publication_fenced" });
const rows = (await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.executionWorkspaceId, reserved.id)))
.filter(row => row.id !== parent).sort((a, b) => a.generation - b.generation);
expect(rows.map(row => row.generation)).toEqual([2, 3]);
expect(rows.map(row => row.parentSnapshotId)).toEqual([parent, parent]);
expect(new Set(rows.map(row => row.objectKey)).size).toBe(2);
const winning = results.findIndex(result => result.status === "fulfilled");
const upload = vi.spyOn(competingStorage, "putFile");
const replay = await saveRemoteWorkspace({ ...input, boundary: boundaries[winning]! });
expect(upload).not.toHaveBeenCalled();
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.snapshotId).toBe(replay.snapshotId);
await expect(saveRemoteWorkspace({ ...input, boundary: boundaries[1 - winning]! }))
.rejects.toThrow("remote_workspace_publication_fenced");
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.snapshotId).toBe(replay.snapshotId);
}, 60_000);
it.each(["failed", "succeeded"])("protects an active duplicate after its peer %s and cleans it only after the last invocation settles", async peer => {
const reserved = await reserve({ kind: "empty" }), first = await target("active-duplicate");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
let release!: () => void, entered!: () => void, releaseOwnership!: () => void, postUpload!: () => void, uploads = 0, checks = 0;
const blocked = new Promise<void>(resolve => { release = resolve; });
const started = new Promise<void>(resolve => { entered = resolve; });
const ownershipGate = new Promise<void>(resolve => { releaseOwnership = resolve; });
const afterUpload = new Promise<void>(resolve => { postUpload = resolve; });
const delayed: StorageService = { ...storage, putFile: async upload => {
if (++uploads === 1) { entered(); await blocked; }
else if (uploads === 2 && peer === "failed") throw new Error("duplicate upload failed");
return storage.putFile(upload);
} };
const input = { db, storage: delayed, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease,
target: first, assertOwned: async () => undefined, boundary: "repository:shared" };
const late = saveRemoteWorkspace({ ...input, assertOwned: async () => {
if (++checks === 3) { postUpload(); await ownershipGate; }
} }).then(value => ({ value, error: null }), error => ({ value: null, error }));
let worker: ReturnType<typeof registerRemoteWorkspaceSnapshotRetentionWork> | undefined;
let coordinator: ReturnType<typeof createDeliveryWorkCoordinator> | undefined;
try {
await Promise.race([started, late.then(result => { throw result.error ?? new Error("Upload did not reach its barrier"); })]);
if (peer === "failed") await expect(saveRemoteWorkspace(input)).rejects.toThrow("duplicate upload failed");
else await saveRemoteWorkspace(input);
const [older] = await db.select().from(executionWorkspaceSnapshots).where(and(
eq(executionWorkspaceSnapshots.executionWorkspaceId, reserved.id), eq(executionWorkspaceSnapshots.generation, 2)));
const [olderOperation] = await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, older!.sourceOperationId));
await fs.writeFile(path.join(prepared.remoteRoot, "after-duplicate.txt"), "the final save contains newer work\n");
const final = await saveRemoteWorkspace({ ...input, workspace: await workspace(reserved.id), boundary: "final" });
expect(final.generation).toBe(3);
const staging = path.join(prepared.remoteRoot, ".paperclip-runtime", "workspace-snapshots");
expect(await fs.readdir(staging)).toContain(olderOperation!.id);
const retention = remoteWorkspaceSnapshotRetentionService(db, { storage,
now: () => new Date(Date.now() + 8 * 86_400_000) });
const tick = vi.spyOn(retention, "tick");
coordinator = createDeliveryWorkCoordinator({ owner: db, canRun: () => true, onError: error => { throw error; } });
worker = registerRemoteWorkspaceSnapshotRetentionWork(coordinator, retention, () => {});
await vi.waitFor(() => expect(tick).toHaveBeenCalled(), { timeout: 4000, interval: 20 });
await vi.waitFor(async () => expect(await retention.nextDeadlineAt()).toBeNull(), { timeout: 4000, interval: 20 });
expect((await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, older!.id)))[0]?.status)
.toBe(peer === "failed" ? "pending" : "superseded");
expect(await fs.readdir(staging)).toContain(olderOperation!.id);
const heldTicks = tick.mock.calls.length;
release();
await Promise.race([afterUpload, late.then(result => { throw result.error ?? new Error("Upload did not reach ownership check"); })]);
// Consume the upload's own notification while its invocation still pins
// the row. Only the last-token settlement can wake the quiet worker next.
await vi.waitFor(() => expect(tick.mock.calls.length).toBeGreaterThan(heldTicks), { timeout: 4000, interval: 20 });
await vi.waitFor(async () => expect(await retention.nextDeadlineAt()).toBeNull(), { timeout: 4000, interval: 20 });
const postUploadTicks = tick.mock.calls.length;
releaseOwnership();
expect((await late).error).toMatchObject({ message: "remote_workspace_publication_fenced" });
await vi.waitFor(async () => expect((await db.select().from(executionWorkspaceSnapshots)
.where(eq(executionWorkspaceSnapshots.id, older!.id)))[0]?.status).toBe("deleted"), { timeout: 4000, interval: 20 });
expect(tick.mock.calls.length).toBeGreaterThan(postUploadTicks);
const [settled] = await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, olderOperation!.id));
expect(settled!.status).toBe(peer === "failed" ? "failed" : "succeeded");
expect(settled!.metadata?.exportCleanup).toMatchObject({ state: "cleaned" });
expect(await fs.readdir(staging)).toEqual([]);
const keys = older!.metadata!.uploads as Array<{ objectKey: string }>;
for (const item of keys) expect((await storage.headObject(ids.company, item.objectKey)).exists).toBe(false);
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.snapshotId).toBe(final.snapshotId);
const saved = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id, snapshotId: final.snapshotId });
try { expect(await fs.readFile(path.join(saved.root, "after-duplicate.txt"), "utf8")).toBe("the final save contains newer work\n"); }
finally { await saved.cleanup(); }
} finally { release(); releaseOwnership(); await late; await worker?.stop(); await coordinator?.stop(); }
}, 60_000);
it.each(["untracked", "malformed-owner", "malformed-active"])("keeps adopted %s work pinned after a newer final save", async marker => {
const reserved = await reserve({ kind: "empty" }), first = await target("unknown-save-work");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const input = { db, storage, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease,
target: first, assertOwned: async () => undefined, boundary: "repository:legacy" };
const failed: StorageService = { ...storage, putFile: async () => { throw new Error("old upload failed"); } };
await expect(saveRemoteWorkspace({ ...input, storage: failed })).rejects.toThrow("old upload failed");
const [old] = await db.select().from(executionWorkspaceSnapshots).where(and(
eq(executionWorkspaceSnapshots.executionWorkspaceId, reserved.id), eq(executionWorkspaceSnapshots.generation, 2)));
// Older/crashed callers supplied no per-invocation completion evidence.
await db.update(workspaceOperations).set({ metadata: sql`${workspaceOperations.metadata} - 'saveWork'` })
.where(eq(workspaceOperations.id, old!.sourceOperationId));
await saveRemoteWorkspace(input);
if (marker === "malformed-owner") await db.update(workspaceOperations).set({ metadata: sql`jsonb_set(${workspaceOperations.metadata},
'{saveWork,unknown}', '"false"'::jsonb)` }).where(eq(workspaceOperations.id, old!.sourceOperationId));
if (marker === "malformed-active") await db.update(workspaceOperations).set({ metadata: sql`jsonb_set(jsonb_set(${workspaceOperations.metadata},
'{saveWork,unknown}', 'false'::jsonb), '{saveWork,activeIds}', '{"unfinished":"unknown"}'::jsonb)` })
.where(eq(workspaceOperations.id, old!.sourceOperationId));
const final = await saveRemoteWorkspace({ ...input, workspace: await workspace(reserved.id), boundary: "final" });
const [held] = await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, old!.sourceOperationId));
expect(held).toMatchObject({ status: "succeeded", metadata: { exportCleanup: { state: "pending" } } });
if (marker === "untracked") expect(held!.metadata?.saveWork).toMatchObject({ activeIds: [], unknown: true });
const retention = remoteWorkspaceSnapshotRetentionService(db, { storage, now: () => new Date(Date.now() + 8 * 86_400_000) });
await retention.tick();
expect((await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, old!.id)))[0]?.status).toBe("superseded");
expect(await fs.readdir(path.join(prepared.remoteRoot, ".paperclip-runtime", "workspace-snapshots"))).toContain(old!.sourceOperationId);
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.snapshotId).toBe(final.snapshotId);
}, 60_000);
it.each(["issue_column", "context_only"])("retains pre-native preparation files and the exact repair obligation when the first upload fails: %s", async identity => {
await db.update(heartbeatRuns).set({ ...legacyControllerClaim("legacy"), runtimeMode: "legacy", nativeIssueId: null,
issueId: identity === "issue_column" ? ids.issue : null, contextSnapshot: { issueId: ids.issue } })
.where(eq(heartbeatRuns.id, ids.run));
const [claimed] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run));
// The production preparer's ownership callback renews on another pool.
// Bound it so a row-lock inversion fails rather than hanging this test.
const assertOwned = async () => livenessDb(db).transaction(async tx => {
await tx.execute(sql`select set_config('statement_timeout', '500ms', true)`);
if (!await renewLegacyControllerLease(tx as unknown as typeof db, claimed!)) throw new Error("controller lease lost");
});
const reserved = await reserve({ kind: "empty" }), first = await target("first-save-outage");
const failing: StorageService = { ...storage, putFile: async () => { throw new Error("initial storage offline"); } };
await expect(prepareRemoteWorkspace({ db, storage: failing, workspace: reserved, runId: ids.run,
lease: await lease(), target: first, assertOwned,
provisionCommand: "printf 'one remote effect\\n' >> setup-result.txt" })).rejects.toThrow("initial storage offline");
const [beforeFailure] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run));
await terminalizeLegacyExecution({ db, run: beforeFailure!, status: "failed", patch: {
errorCode: "setup_failed", resultJson: { executionRecovery: { kind: "bootstrap", providerWorkStarted: false } },
} });
const [failed] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run));
expect(failed).toMatchObject({ runtimeMode: "legacy", status: "failed", runnerProfileJson: {
remoteWorkspace: { state: "prepared", snapshotId: null, workspaceId: reserved.id },
}, resultJson: { workspaceRestoreFailure: "restore_failed", workspaceRestoreRecovery: {
schema: "paperclip.workspace-restore-recovery.v1", leaseIds: [ids.lease],
} } });
expect(failed?.runnerProfileJson).not.toHaveProperty("nativeExecutionInput");
expect(await db.select().from(nativeRunFinalizations).where(eq(nativeRunFinalizations.runId, ids.run))).toHaveLength(0);
const source = await lease();
expect(await nativeWorkspaceExportPendingForLease(db, source)).toBe(true);
expect(source).toMatchObject({ status: "pending_cleanup", leasePolicy: "retain_on_failure" });
const runtime = environmentRuntimeService(db);
const [environment] = await db.select().from(environments).where(eq(environments.id, ids.environment));
expect(await runtime.destroyRunLease({ environment: environment! as Environment, lease: source })).toBeNull();
const retainedAuthority = readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)!;
expect(retainedAuthority).toMatchObject({
dirty: true, writerRunId: ids.run, snapshotId: null, location: { leaseId: ids.lease },
});
expect(await fs.readFile(path.join(retainedAuthority.location!.remoteCwd, "setup-result.txt"), "utf8")).toBe("one remote effect\n");
const actions = await db.select().from(issueRecoveryActions).where(eq(issueRecoveryActions.companyId, ids.company));
expect(actions).toHaveLength(1);
expect(actions[0]?.evidence).toMatchObject({ runId: ids.run, workspaceRestoreRecovery: { leaseIds: [ids.lease] } });
}, 60_000);
it("reconstructs a stopped pre-native file obligation only for its exact current authority", async () => {
await db.update(heartbeatRuns).set({ runtimeMode: "legacy", nativeIssueId: null, issueId: null,
contextSnapshot: { issueId: ids.issue } }).where(eq(heartbeatRuns.id, ids.run));
const reserved = await reserve({ kind: "empty" }), first = await target("orphan-preparation");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run,
lease: await lease(), target: first });
await db.update(heartbeatRuns).set({ status: "failed" }).where(eq(heartbeatRuns.id, ids.run));
const current = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(current.metadata)!;
for (const invalid of [
{ ...authority, epoch: authority.epoch + 1 },
{ ...authority, writerRunId: randomUUID() },
{ ...authority, location: { ...authority.location!, leaseId: randomUUID() } },
{ ...authority, location: { ...authority.location!, environmentId: randomUUID() } },
]) {
await db.update(executionWorkspaces).set({ metadata: { ...current.metadata, remoteWorkspace: invalid } })
.where(eq(executionWorkspaces.id, reserved.id));
expect(await nativeWorkspaceExportPendingForLease(db, prepared.lease)).toBe(false);
expect(await preserveRemoteWorkspacePreparationFailure(db, prepared.lease)).toBeNull();
}
await db.update(executionWorkspaces).set({ metadata: current.metadata }).where(eq(executionWorkspaces.id, reserved.id));
expect(await nativeWorkspaceExportPendingForLease(db, { ...prepared.lease, companyId: randomUUID() })).toBe(false);
expect(await preserveRemoteWorkspacePreparationFailure(db, { ...prepared.lease, id: randomUUID() })).toBeNull();
const otherIssueId = randomUUID();
await db.insert(issues).values({ id: otherIssueId, companyId: ids.company, title: "Unrelated task" });
await db.update(executionWorkspaces).set({ sourceIssueId: otherIssueId }).where(eq(executionWorkspaces.id, reserved.id));
expect(await preserveRemoteWorkspacePreparationFailure(db, prepared.lease)).toBeNull();
await db.update(executionWorkspaces).set({ sourceIssueId: ids.issue }).where(eq(executionWorkspaces.id, reserved.id));
await db.update(heartbeatRuns).set({ controllerBootId: randomUUID(), controllerLeaseExpiresAt: new Date(Date.now() + 60_000) })
.where(eq(heartbeatRuns.id, ids.run));
expect(await preserveOrphanedNativeWorkspace(db, prepared.lease)).toBe(false);
expect(await nativeWorkspaceExportPendingForLease(db, prepared.lease)).toBe(true);
expect(await preserveRemoteWorkspacePreparationFailure(db, prepared.lease, { assertOwned: async () => {
await db.update(heartbeatRuns).set({ controllerBootId: randomUUID() }).where(eq(heartbeatRuns.id, ids.run));
} })).toBeNull();
await db.update(heartbeatRuns).set({ controllerBootId: null, controllerLeaseExpiresAt: null }).where(eq(heartbeatRuns.id, ids.run));
expect(await db.select().from(issueRecoveryActions).where(eq(issueRecoveryActions.companyId, ids.company))).toHaveLength(0);
expect(await preserveOrphanedNativeWorkspace(db, prepared.lease)).toBe(true);
expect((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run)))[0]?.resultJson)
.toMatchObject({ workspaceRestoreRecovery: { leaseIds: [ids.lease] } });
expect(await nativeWorkspaceExportPendingForLease(db, prepared.lease)).toBe(true);
await saveRemoteWorkspace({ db, storage, workspace: await workspace(reserved.id), runId: ids.run,
lease: await lease(), target: first, assertOwned: async () => undefined });
expect(await nativeWorkspaceExportPendingForLease(db, prepared.lease)).toBe(false);
expect(await preserveRemoteWorkspacePreparationFailure(db, prepared.lease)).toBeNull();
}, 60_000);
it.each([
["a lost reply", "remote_workspace_export_cleanup_failed"],
["a timeout", "remote_workspace_operation_timed_out: cleanup (configured timeout 10000ms)"],
["an untrusted error", "remote_workspace_export_cleanup_failed"],
] as const)("records and retries export cleanup after %s without another upload", async (failure, diagnostic) => {
const reserved = await reserve({ kind: "empty" }), first = await target("cleanup-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "saved.txt"), "keep working files\n");
const sandbox = first as Extract<AdapterExecutionTarget, { transport: "sandbox" }>;
const execute = sandbox.runner!.execute.bind(sandbox.runner!);
let lost = false, cleanupAttempts = 0;
sandbox.runner!.execute = async input => {
const result = await execute(input);
if (input.args?.[1]?.includes("remote_workspace_export_cleanup_failed") && result.exitCode === 0) {
cleanupAttempts++;
if (lost) return result;
lost = true;
if (failure === "a timeout") return { ...result, timedOut: true,
stdout: "SYNTHETIC_PRIVATE_STDOUT", stderr: "SYNTHETIC_PRIVATE_STDERR" };
throw new Error(failure === "an untrusted error"
? "remote_workspace_operation_timed_out: cleanup (configured timeout 10000ms)\nSYNTHETIC_PRIVATE_STDERR"
: "lost cleanup reply");
}
return result;
};
const upload = vi.spyOn(storage, "putFile");
const input = { db, storage, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease,
target: first, assertOwned: async () => undefined };
const saved = await saveRemoteWorkspace(input);
const [before] = await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.heartbeatRunId, ids.run),
eq(workspaceOperations.phase, "remote_workspace_snapshot"), eq(workspaceOperations.status, "succeeded"),
eq(workspaceOperations.id, (await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.id, saved.snapshotId)))[0]!.sourceOperationId)));
expect(before?.metadata?.exportCleanup).toMatchObject({ state: "pending", lastError: diagnostic });
expect(cleanupAttempts).toBe(1);
expect(await saveRemoteWorkspace(input)).toMatchObject({ snapshotId: saved.snapshotId, generation: saved.generation });
expect(upload).toHaveBeenCalledOnce();
const [after] = await db.select().from(workspaceOperations).where(eq(workspaceOperations.id, before!.id));
expect(after?.metadata?.exportCleanup).toMatchObject({ state: "cleaned", lastError: null });
expect(cleanupAttempts).toBe(2);
expect(await fs.readFile(path.join(prepared.remoteRoot, "saved.txt"), "utf8")).toBe("keep working files\n");
}, 60_000);
it("replays an acknowledged preparation save after a newer boundary without recapture", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("boundary-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
const input = { db, storage, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease,
target: first, assertOwned: async () => undefined };
const older = await saveRemoteWorkspace({ ...input, boundary: "repository:one" });
await fs.writeFile(path.join(prepared.remoteRoot, "later.txt"), "later acknowledged work\n");
const newer = await saveRemoteWorkspace({ ...input, boundary: "repository:two" });
const upload = vi.spyOn(storage, "putFile");
expect(await saveRemoteWorkspace({ ...input, boundary: "repository:one" })).toMatchObject({ snapshotId: older.snapshotId });
expect(upload).not.toHaveBeenCalled();
expect(readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)?.snapshotId).toBe(newer.snapshotId);
}, 60_000);
it("retains an uncertain provisioning command and never replays its external effect", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("provision-sandbox");
const sandbox = first as Extract<AdapterExecutionTarget, { transport: "sandbox" }>;
const execute = sandbox.runner!.execute.bind(sandbox.runner!);
let commandCalls = 0;
sandbox.runner!.execute = async input => {
const result = await execute(input);
if (input.command === "bash") { commandCalls++; throw new Error("lost provisioning reply"); }
return result;
};
const input = { db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first,
provisionCommand: "printf 'one effect\\n' >> setup-result.txt" };
await expect(prepareRemoteWorkspace(input)).rejects.toThrow("lost provisioning reply");
const retained = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(retained.metadata)!;
expect(authority).toMatchObject({ dirty: true, writerRunId: ids.run, snapshotId: null });
expect(authority.location).toBeTruthy();
await expect(prepareRemoteWorkspace({ ...input, workspace: retained, lease: await lease() })).rejects.toThrow("remote_workspace_provision_unverified");
expect(commandCalls).toBe(1);
expect(await fs.readFile(path.join(authority.location!.remoteCwd, "setup-result.txt"), "utf8")).toBe("one effect\n");
const [journal] = await db.select().from(workspaceOperations).where(and(eq(workspaceOperations.companyId, ids.company), eq(workspaceOperations.phase, "remote_workspace_provision")));
expect(journal).toMatchObject({ status: "running", metadata: { uncertain: true } });
}, 60_000);
it("preserves enrolled workspaces after disabling enrollment and rejects a legacy writer", async () => {
const reserved = await reserve({ kind: "empty" });
expect(shouldUseRemoteWorkspaces({ adapterType: "paperclip_runner", environmentDriver: "sandbox", experimentalEnabled: false, existingWorkspace: reserved })).toBe(true);
expect(() => shouldUseRemoteWorkspaces({ adapterType: "codex_local", environmentDriver: "sandbox", experimentalEnabled: false, existingWorkspace: reserved })).toThrow("native sandbox");
});
it("keeps enrollment conditional while enrolled authority fails closed when persistence disappears", async () => {
const candidate = { adapterType: "paperclip_runner", environmentDriver: "sandbox", experimentalEnabled: true };
vi.stubEnv("PAPERCLIP_STORAGE_PROVIDER", "s3");
expect(shouldUseRemoteWorkspaces(candidate)).toBe(true);
expect(shouldUseRemoteWorkspaces({ ...candidate, experimentalEnabled: false })).toBe(false);
expect(shouldUseRemoteWorkspaces({ ...candidate, existingWorkspace: { metadata: {} } })).toBe(false);
const reserved = await reserve({ kind: "empty" });
const before = (await workspace(reserved.id)).metadata;
vi.stubEnv("PAPERCLIP_STORAGE_PROVIDER", "local_disk");
expect(shouldUseRemoteWorkspaces(candidate)).toBe(false);
expect(shouldUseRemoteWorkspaces({ ...candidate, existingWorkspace: reserved })).toBe(true);
const localStorage = createStorageService(createLocalDiskStorageProvider(path.join(root, "local-only")));
await expect(reserveRemoteWorkspace({ db, storage: localStorage, companyId: ids.company,
issueId: ids.issue, runId: ids.run, agentId: ids.agent, environmentId: ids.environment,
projectId: null, projectWorkspaceId: null, existingWorkspaceId: reserved.id, source: { kind: "empty" } }))
.rejects.toMatchObject({ details: { code: "remote_workspace_persistence_unavailable" } });
expect((await workspace(reserved.id)).metadata).toEqual(before);
vi.stubEnv("PAPERCLIP_STORAGE_PROVIDER", "s3");
vi.stubEnv("DATABASE_URL", "");
expect(shouldUseRemoteWorkspaces(candidate)).toBe(false);
expect(shouldUseRemoteWorkspaces({ ...candidate, existingWorkspace: reserved })).toBe(true);
vi.stubEnv("DATABASE_URL", databaseUrl!);
vi.stubEnv("PAPERCLIP_SECRETS_MASTER_KEY", "");
vi.stubEnv("PAPERCLIP_SECRETS_MASTER_KEY_FILE", path.join(root, "missing-stable-key"));
expect(shouldUseRemoteWorkspaces(candidate)).toBe(false);
expect(shouldUseRemoteWorkspaces({ ...candidate, existingWorkspace: reserved })).toBe(true);
});
it("selects Git credentials with the current run's agent, responsible user, and task", async () => {
await db.update(heartbeatRuns).set({ responsibleUserId: "synthetic-current-user" }).where(eq(heartbeatRuns.id, ids.run));
const contexts: unknown[] = [];
const create = ((_db, companyId, context) => {
expect(companyId).toBe(ids.company); contexts.push(context); return async () => null;
}) as typeof createGitRemoteAuthProvider;
await createRemoteWorkspaceGitAuth(db, ids.company, ids.run, create);
expect(contexts).toEqual([{ issueId: ids.issue, heartbeatRunId: ids.run, agentId: ids.agent,
responsibleUserId: "synthetic-current-user" }]);
await expect(createRemoteWorkspaceGitAuth(db, randomUUID(), ids.run, create)).rejects.toThrow("remote_workspace_git_run_unavailable");
await db.update(heartbeatRuns).set({ status: "cancelled" }).where(eq(heartbeatRuns.id, ids.run));
await expect(createRemoteWorkspaceGitAuth(db, ids.company, ids.run, create)).rejects.toThrow("remote_workspace_git_run_unavailable");
expect(contexts).toHaveLength(1);
});
it("reclaims an unused reservation after allocation failed without granting a second live writer", async () => {
const reserved = await reserve({ kind: "empty" });
const nextRun = randomUUID();
await db.insert(heartbeatRuns).values({ id: nextRun, companyId: ids.company, agentId: ids.agent, nativeIssueId: ids.issue,
runtimeMode: "native", status: "running" });
const input = { db, storage, companyId: ids.company, issueId: ids.issue, runId: nextRun, agentId: ids.agent,
environmentId: ids.environment, projectId: null, projectWorkspaceId: null, existingWorkspaceId: reserved.id, source: { kind: "empty" as const } };
await expect(reserveRemoteWorkspace(input)).rejects.toThrow("unfinished writer");
await db.delete(environmentLeases).where(eq(environmentLeases.id, ids.lease));
await db.update(heartbeatRuns).set({ status: "failed" }).where(eq(heartbeatRuns.id, ids.run));
const reclaimed = await reserveRemoteWorkspace(input);
expect(readRemoteWorkspaceAuthority(reclaimed.metadata)).toMatchObject({ writerRunId: nextRun, epoch: 2, dirty: false, location: null });
});
it("rejects publication after writer ownership changes, retaining the uploaded recovery obligation", async () => {
const reserved = await reserve({ kind: "empty" }), first = await target("stale-writer");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "work.txt"), "must not overwrite a newer owner");
let calls = 0;
await expect(saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease, target: first,
assertOwned: async () => {
if (++calls !== 3) return;
const row = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(row.metadata)!;
await db.update(executionWorkspaces).set({ metadata: { ...row.metadata, remoteWorkspace: { ...authority, epoch: authority.epoch + 1 } } })
.where(eq(executionWorkspaces.id, row.id));
} })).rejects.toThrow("remote_workspace_publication_fenced");
const current = readRemoteWorkspaceAuthority((await workspace(reserved.id)).metadata)!;
expect(current.snapshotId).toBe(readRemoteWorkspaceAuthority(prepared.workspace.metadata)!.snapshotId);
expect(current.dirty).toBe(true);
const snapshots = await db.select().from(executionWorkspaceSnapshots).where(eq(executionWorkspaceSnapshots.companyId, ids.company));
expect(snapshots.map(snapshot => snapshot.status).sort()).toEqual(["committed", "uploaded"]);
}, 60_000);
it("saves dirty and unpushed Git work and restores it without contacting the original source", async () => {
const origin = path.join(root, "source-repository"); await fs.mkdir(origin);
const git = (cwd: string, ...args: string[]) => exec("git", ["-C", cwd, "-c", "user.name=Synthetic test", "-c", "user.email=test@example.invalid", ...args]);
await git(origin, "init", "--initial-branch=main");
await fs.writeFile(path.join(origin, "code.txt"), "original\n"); await git(origin, "add", "."); await git(origin, "commit", "-m", "Synthetic baseline");
const reserved = await reserve({ kind: "git", repoUrl: origin, ref: "HEAD", branchName: "task-work" }), first = await target("git-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run, lease: await lease(), target: first });
await fs.writeFile(path.join(prepared.remoteRoot, "code.txt"), "unpushed commit\n");
await git(prepared.remoteRoot, "add", "."); await git(prepared.remoteRoot, "commit", "-m", "Synthetic unpushed work");
const head = (await git(prepared.remoteRoot, "rev-parse", "HEAD")).stdout.trim();
await fs.writeFile(path.join(prepared.remoteRoot, "code.txt"), "staged edit\n"); await git(prepared.remoteRoot, "add", "code.txt");
await fs.writeFile(path.join(prepared.remoteRoot, "code.txt"), "unstaged edit\n");
await fs.writeFile(path.join(prepared.remoteRoot, "untracked.txt"), "untracked work\n");
const status = (await git(prepared.remoteRoot, "status", "--porcelain")).stdout;
const saved = await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace, runId: ids.run, lease: prepared.lease,
target: first, assertOwned: async () => undefined });
await fs.rm(origin, { recursive: true, force: true }); await fs.rm(prepared.remoteRoot, { recursive: true, force: true });
const recovered = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id, snapshotId: saved.snapshotId });
try {
expect((await git(recovered.root, "rev-parse", "HEAD")).stdout.trim()).toBe(head);
expect((await git(recovered.root, "status", "--porcelain")).stdout).toBe(status);
expect((await git(recovered.root, "show", ":code.txt")).stdout).toBe("staged edit\n");
expect(await fs.readFile(path.join(recovered.root, "code.txt"), "utf8")).toBe("unstaged edit\n");
} finally { await recovered.cleanup(); }
}, 60_000);
it.each(["saved", "dirty", "unsaved", "foreign_snapshot", "delete_requested", "deletion_race", "controller_claim_race", "stale_epoch", "unconfirmed_stop", "unknown_action", "historical_listener_guard", "historical_listener_provider_event", "historical_listener_checkpoint", "historical_listener_resumed", "historical_listener_other_message", "historical_listener_cancelled", "historical_listener_prior_session", "historical_listener_unconfirmed_stop", "historical_listener_dirty", "historical_listener_stale_block", "historical_listener_unknown_event", "historical_listener_budget", "historical_listener_bad_preservation", "historical_listener_absent", "historical_listener_later_candidate", "historical_listener_late_proof", "historical_listener_lock_timeout", "historical_listener_late_interaction", "historical_listener_interaction_paused", "historical_listener_own_question"] as const)(
"checks durable remote work at the native safe-replacement decision (%s)", async kind => {
const historicalListener = kind.startsWith("historical_listener_");
const processAcceptance = kind === "historical_listener_guard" && process.env.PAPERCLIP_TEST_HISTORICAL_LISTENER === "1";
const scenario = kind === "historical_listener_late_proof" ? "dirty" : kind === "historical_listener_unconfirmed_stop" ? "unconfirmed_stop"
: kind === "historical_listener_dirty" ? "dirty" : historicalListener ? "saved" : kind;
const shouldSchedule = scenario === "saved" && (!historicalListener || ["historical_listener_guard", "historical_listener_later_candidate", "historical_listener_lock_timeout"].includes(kind));
const reserved = await reserve({ kind: "empty" }), first = await target("native-replacement-sandbox");
const prepared = await prepareRemoteWorkspace({ db, storage, workspace: reserved, runId: ids.run,
lease: await lease(), target: first, assertOwned: async () => undefined });
await fs.writeFile(path.join(prepared.remoteRoot, "saved.txt"), "work survives a failed provider session\n");
if (scenario !== "unsaved") await saveRemoteWorkspace({ db, storage, workspace: prepared.workspace,
runId: ids.run, lease: prepared.lease, target: first, assertOwned: async () => undefined,
boundary: kind === "historical_listener_late_proof" ? "initial" : "final" });
const current = await workspace(reserved.id), authority = readRemoteWorkspaceAuthority(current.metadata)!;
if (scenario === "dirty" || scenario === "stale_epoch") await db.update(executionWorkspaces).set({ metadata: {
...current.metadata, remoteWorkspace: { ...authority, ...(scenario === "dirty" ? { dirty: true, writerRunId: ids.run }
: { epoch: authority.epoch + 1 }) },
} }).where(eq(executionWorkspaces.id, reserved.id));
if (scenario === "foreign_snapshot") await db.update(executionWorkspaceSnapshots).set({ sourceRunId: randomUUID() })
.where(eq(executionWorkspaceSnapshots.id, authority.snapshotId!));
if (scenario === "delete_requested") await db.update(executionWorkspaceSnapshots).set({ status: "delete_requested" })
.where(eq(executionWorkspaceSnapshots.id, authority.snapshotId!));
if (processAcceptance) {
const { recordHistoricalListenerFailure } = await import("../__tests__/helpers/historical-listener-continuation.js");
await recordHistoricalListenerFailure({ db, connectionString: databaseUrl!, runId: ids.run, cwd: prepared.remoteRoot,
environmentId: ids.environment, leaseId: ids.lease });
}
const stopped = kind === "historical_listener_late_proof" ? prepared.lease : (await db.update(environmentLeases).set({ status: "released", releasedAt: new Date(), cleanupStatus: "success",
metadata: { ...prepared.lease.metadata, ...(scenario !== "unconfirmed_stop" ? {
remoteExecutionTermination: remoteTerminationReceipt(prepared.lease, {
providerLeaseId: prepared.lease.providerLeaseId, state: "destroyed",
}),
} : {}) },
}).where(eq(environmentLeases.id, ids.lease)).returning())[0]! as EnvironmentLease;
if (shouldSchedule) {
await db.transaction(tx => clearRemoteWorkspaceLocationForReleasedLease({ db: tx, lease: stopped }));
await fs.rm(prepared.remoteRoot, { recursive: true, force: true });
await expect(fs.stat(prepared.remoteRoot)).rejects.toThrow();
}
const [source] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run));
const sessionId = randomUUID();
await db.update(heartbeatRuns).set({ status: "failed", finishedAt: new Date(),
...(historicalListener && !processAcceptance ? { nativeSessionId: sessionId } : {}),
...(historicalListener ? { error: "local_runtime_api_listener_unavailable", errorCode: "native_session_interrupted" } : {}), runnerProfileJson: {
...source!.runnerProfileJson, recoveryEventInventoryVersion: 1,
nativeExecutionInput: processAcceptance ? source!.runnerProfileJson!.nativeExecutionInput : { provider: { kind: "codex" }, workspace: { cwd: prepared.remoteRoot },
...(historicalListener ? { binding: { runId: ids.run, companyId: ids.company, agentId: ids.agent, issueId: ids.issue },
session: { normalizedSessionId: sessionId } } : {}) },
} }).where(eq(heartbeatRuns.id, ids.run));
if (!processAcceptance) await db.insert(nativeRunFinalizations).values({ runId: ids.run, companyId: ids.company, issueId: ids.issue,
phase: "terminal_failure", attempt: 1, failureCode: historicalListener ? "native_session_interrupted" : "native_provider_terminal_failed",
failureDetail: historicalListener ? { originalFailureCode: "native_session_interrupted", message: "local_runtime_api_listener_unavailable",
recoveryMode: "ambiguous_state", providerSessionEstablished: false, providerEventsExist: false, checkpointExists: false }
: { originalFailureCode: "fixture_checkpoint_unusable" } });
if (historicalListener && !processAcceptance) {
const blocked = await issueService(db).update(ids.issue, { status: "blocked" });
await db.insert(issueRecoveryActions).values({ companyId: ids.company, sourceIssueId: ids.issue,
kind: "active_run_watchdog", cause: "native_session_interrupted", ownerType: "board", status: "active",
fingerprint: createHash("sha256").update(`${ids.run}:native_session_interrupted`).digest("hex"),
evidence: { runId: ids.run, nativeFailureBlock: { runId: ids.run, statusVersion: blocked!.statusVersion } },
nextAction: "Reconcile the original startup failure." });
}
if (kind === "historical_listener_guard" || kind === "historical_listener_bad_preservation") {
const actionId = randomUUID(), preservationId = randomUUID();
await db.insert(issueRecoveryActions).values({ id: actionId, companyId: ids.company, sourceIssueId: ids.issue,
kind: "active_run_watchdog", cause: "native_continuation_requires_reconciliation", ownerType: "system", status: "resolved", outcome: "blocked",
fingerprint: `native-workspace:${ids.run}`, evidence: { runId: ids.run,
workspaceRestoreFailure: "restore_failed", workspaceRestoreRecovery: { schema: "paperclip.workspace-restore-recovery.v1", leaseIds: [ids.lease] },
automaticRecovery: { replay: "blocked" }, continuationDelivery: "invalidated",
systemRecovery: { ...createSystemRecovery(ids.run), state: "complete", preservationOperationIds: [preservationId] },
workspaceRecovery: { schema: "paperclip.workspace-recovery.v1", runId: ids.run, status: "recovered", stage: "saving_receipts", attempt: 1, requestIds: [] },
}, nextAction: "Files were preserved without replay." });
await db.insert(workspaceOperations).values([{ id: preservationId, companyId: ids.company, heartbeatRunId: ids.run, issueId: ids.issue,
executionWorkspaceId: reserved.id, phase: "workspace_recovery", status: kind === "historical_listener_bad_preservation" ? "failed" : "succeeded", startedAt: new Date(),
metadata: { recoveryActionId: actionId, snapshotId: authority.snapshotId } },
{ companyId: ids.company, heartbeatRunId: ids.run, issueId: ids.issue, phase: "workspace_finalize", status: "succeeded", startedAt: new Date(),
metadata: { owningService: "workspace_recovery", preservationOperationIds: [preservationId] } }]);
}
if (kind === "historical_listener_absent") await db.update(heartbeatRuns).set({
runnerProfileJson: sql`${heartbeatRuns.runnerProfileJson} - 'remoteWorkspace'`,
}).where(eq(heartbeatRuns.id, ids.run));
if (kind === "historical_listener_stale_block") await issueService(db).update(ids.issue, { status: "blocked", actorUserId: "operator" });
if (kind === "historical_listener_unknown_event") await db.insert(heartbeatRunEvents).values({ companyId: ids.company, runId: ids.run,
agentId: ids.agent, seq: 100, eventType: "future.provider.activity", sourceEventId: "source-provider-event" });
if (kind === "historical_listener_budget") await db.update(nativeRunFinalizations).set({ attempt: 3 }).where(eq(nativeRunFinalizations.runId, ids.run));
if (kind === "historical_listener_checkpoint") await db.update(heartbeatRuns).set({ runnerProfileJson: {
...((await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, ids.run)))[0]!.runnerProfileJson), sessionCheckpoint: { providerSessionId: "already-started" },
} }).where(eq(heartbeatRuns.id, ids.run));
if (kind === "historical_listener_resumed") await db.update(heartbeatRuns).set({ sessionIdBefore: "prior-session" }).where(eq(heartbeatRuns.id, ids.run));
if (kind === "historical_listener_prior_session") await db.insert(heartbeatRuns).values({ id: randomUUID(), companyId: ids.company,
agentId: ids.agent, status: "succeeded", nativeSessionId: sessionId });
if (kind === "historical_listener_other_message") await db.update(nativeRunFinalizations).set({ failureDetail: {
message: "provider reported local_runtime_api_listener_unavailable", originalFailureCode: "native_session_interrupted", recoveryMode: "ambiguous_state",
providerSessionEstablished: false, providerEventsExist: false, checkpointExists: false,
} }).where(eq(nativeRunFinalizations.runId, ids.run));
if (kind === "historical_listener_cancelled") await db.update(issues).set({ status: "cancelled" }).where(eq(issues.id, ids.issue));
if (kind === "historical_listener_provider_event") await appendHeartbeatRunEvent(db, { companyId: ids.company, runId: ids.run,
agentId: ids.agent, eventType: "turn.started", stream: "system" });
const interactionWake = ["historical_listener_late_interaction", "historical_listener_interaction_paused", "historical_listener_own_question"].includes(kind);
const interactionId = randomUUID();
if (interactionWake) {
const priorRunId = randomUUID();
await db.insert(heartbeatRuns).values({ id: priorRunId, companyId: ids.company, agentId: ids.agent,
nativeIssueId: ids.issue, status: "succeeded", contextSnapshot: { issueId: ids.issue } });
await db.insert(issueThreadInteractions).values({ id: interactionId, companyId: ids.company, issueId: ids.issue,
kind: "request_confirmation", status: "pending", sourceRunId: kind === "historical_listener_own_question" ? ids.run : priorRunId,
payload: { version: 1, prompt: "An obsolete question from earlier work." } });
if (kind === "historical_listener_interaction_paused") await issueService(db).update(ids.issue, { status: "backlog" });
if (kind === "historical_listener_own_question") await appendHeartbeatRunEvent(db, { companyId: ids.company, runId: ids.run,
agentId: ids.agent, eventType: "tool.execution.started", stream: "system",
payload: { name: "request_confirmation", executionId: "source-question", transport: "process" } });
}
if (scenario === "unknown_action") await appendHeartbeatRunEvent(db, { companyId: ids.company, runId: ids.run,
agentId: ids.agent, eventType: "tool.execution.started", stream: "system",
payload: { name: "send_email", executionId: "unconfirmed-external-action", transport: "process" } });
if (kind === "historical_listener_later_candidate") {
const sourceNumber = BigInt(`0x${ids.run.replaceAll("-", "")}`);
const parked = Array.from({ length: 100 }, (_, index) => {
const hex = (sourceNumber - BigInt(index + 1)).toString(16).padStart(32, "0");
return { runId: `${hex.slice(0,8)}-${hex.slice(8,12)}-${hex.slice(12,16)}-${hex.slice(16,20)}-${hex.slice(20)}`,
issueId: randomUUID(), sessionId: randomUUID() };
});
await db.insert(issues).values(parked.map(row => ({ id: row.issueId, companyId: ids.company,
title: "Preserve unavailable source", status: "in_progress", assigneeAgentId: ids.agent })));
await db.insert(heartbeatRuns).values(parked.map(row => ({ id: row.runId, companyId: ids.company, agentId: ids.agent,
nativeIssueId: row.issueId, nativeSessionId: row.sessionId, runtimeMode: "native", status: "failed",
errorCode: "native_session_interrupted", error: "local_runtime_api_listener_unavailable", runnerProfileJson: {
nativeExecutionInput: { binding: { runId: row.runId, companyId: ids.company, agentId: ids.agent, issueId: row.issueId },
session: { normalizedSessionId: row.sessionId } },
} })));
await db.insert(nativeRunFinalizations).values(parked.map(row => ({ runId: row.runId, companyId: ids.company, issueId: row.issueId,
phase: "terminal_failure", attempt: 1, failureCode: "native_session_interrupted", failureDetail: {
message: "local_runtime_api_listener_unavailable", originalFailureCode: "native_session_interrupted", recoveryMode: "ambiguous_state",
providerSessionEstablished: false, providerEventsExist: false, checkpointExists: false,
} })));
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scanned: 100, scheduled: 0 });
expect(await executionControlSchedule(db).replacement()).not.toBeNull();
invalidateNativeReplacementDiscovery(db);
}
if (kind === "historical_listener_lock_timeout") {
let release!: () => void, announce!: () => void;
const locked = new Promise<void>(resolve => { announce = resolve; });
const released = new Promise<void>(resolve => { release = resolve; });
const holder = db.transaction(async tx => {
await tx.select({ id: issues.id }).from(issues).where(eq(issues.id, ids.issue)).for("update");
announce(); await released;
});
await locked;
try {
await expect(reconcileSafeNativeReplacements(db)).rejects.toThrow();
expect(await executionControlSchedule(db).replacement()).not.toBeNull();
} finally { release(); await holder; }
}
let replacement;
if (scenario === "deletion_race" || scenario === "controller_claim_race") {
let release!: () => void, announce!: () => void;
let blockingPid = 0;
const released = new Promise<void>(resolve => { release = resolve; });
const locked = new Promise<void>(resolve => { announce = resolve; });
const deletion = db.transaction(async tx => {
await tx.select({ id: issues.id }).from(issues).where(eq(issues.id, ids.issue)).for("update");
const [session] = await tx.execute(sql`select pg_backend_pid() as pid`);
blockingPid = Number(session?.pid);
announce();
await released;
if (scenario === "deletion_race") await tx.update(executionWorkspaceSnapshots).set({ status: "delete_requested" })
.where(eq(executionWorkspaceSnapshots.id, authority.snapshotId!));
else await tx.update(nativeRunFinalizations).set({ leaseOwner: "another-controller",
leaseExpiresAt: new Date(Date.now() + 60_000) }).where(eq(nativeRunFinalizations.runId, ids.run));
});
await locked;
const pending = reconcileSafeNativeReplacements(db);
try {
const deadline = performance.now() + 5_000;
let blocked = false;
while (!blocked && performance.now() < deadline) {
const waiting = await db.execute(sql`select exists(select 1 from pg_stat_activity
where datname = current_database() and ${blockingPid} = any(pg_blocking_pids(pid))) as blocked`);
blocked = waiting[0]?.blocked === true;
if (!blocked) await new Promise(resolve => setTimeout(resolve, 10));
}
expect(blocked).toBe(true);
} finally { release(); await deletion; replacement = await pending; }
} else replacement = await reconcileSafeNativeReplacements(db);
expect(replacement).toMatchObject({ scheduled: shouldSchedule ? 1 : 0 });
const successors = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.retryOfRunId, ids.run));
expect(successors).toHaveLength(shouldSchedule ? 1 : 0);
if (shouldSchedule) expect(successors[0]).toMatchObject({ status: "scheduled_retry",
contextSnapshot: { retryOfRunId: ids.run, forceFreshSession: true } });
if (kind === "historical_listener_later_candidate") {
for (let pass = 0; pass < 3 && await executionControlSchedule(db).replacement() !== null; pass++) {
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scheduled: 0 });
}
expect(await executionControlSchedule(db).replacement()).toBeNull();
}
if (kind === "historical_listener_guard") expect(await getExecutionBlocker(db, ids.company, ids.issue)).toBeNull();
if (["historical_listener_provider_event", "historical_listener_checkpoint", "historical_listener_resumed", "historical_listener_other_message",
"historical_listener_cancelled", "historical_listener_prior_session", "historical_listener_unknown_event", "historical_listener_budget", "historical_listener_stale_block", "historical_listener_bad_preservation", "historical_listener_dirty", "historical_listener_unconfirmed_stop", "historical_listener_absent", "historical_listener_late_proof"].includes(kind)) {
expect(await executionControlSchedule(db).replacement()).toBeNull();
}
if (interactionWake) {
expect(await executionControlSchedule(db).replacement()).toBeNull();
let wakes = 0;
const unsubscribe = subscribeDatabaseWork(db, DELIVERY_QUEUES.executionReplacement, event => {
if (event === "settled") { wakes++; invalidateNativeReplacementDiscovery(db); }
});
try {
await issueThreadInteractionService(db).skipInteraction({ id: ids.issue, companyId: ids.company,
status: kind === "historical_listener_interaction_paused" ? "backlog" : "blocked" }, interactionId, {}, { userId: "local-board" });
expect(wakes).toBeGreaterThan(0);
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scheduled: kind === "historical_listener_late_interaction" ? 1 : 0 });
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scheduled: 0 });
expect(await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.retryOfRunId, ids.run)))
.toHaveLength(kind === "historical_listener_late_interaction" ? 1 : 0);
expect(await executionControlSchedule(db).replacement()).toBeNull();
} finally { unsubscribe(); }
}
if (kind === "historical_listener_late_proof") {
let wakes = 0;
const unsubscribe = subscribeDatabaseWork(db, DELIVERY_QUEUES.executionReplacement, event => {
if (event === "settled") { wakes++; invalidateNativeReplacementDiscovery(db); }
});
try {
const saved = await saveRemoteWorkspace({ db, storage, workspace: await workspace(reserved.id), runId: ids.run,
lease: await lease(), target: first, assertOwned: async () => undefined });
expect(wakes).toBeGreaterThan(0);
expect(await executionControlSchedule(db).replacement()).toBeNull();
const priorWakes = wakes, active = await lease(), pluginId = randomUUID();
await db.update(environmentLeases).set({ metadata: { ...active.metadata, driver: "sandbox", pluginId,
provider: "daytona", sandboxProviderPlugin: true } }).where(eq(environmentLeases.id, ids.lease));
const runtime = environmentRuntimeService(db, { pluginWorkerManager: {
isRunning: () => true, getWorker: () => ({ supportedMethods: ["environmentReleaseLease"] }),
call: async (_id: string, method: string) => {
expect(method).toBe("environmentReleaseLease");
return { providerLeaseId: active.providerLeaseId, state: "destroyed" };
},
} as never });
await runtime.releaseRunLeases(ids.run, "failed");
expect(wakes).toBeGreaterThan(priorWakes);
expect(await executionControlSchedule(db).replacement()).not.toBeNull();
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scheduled: 1 });
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scheduled: 0 });
expect(await executionControlSchedule(db).replacement()).toBeNull();
const restored = await createSavedWorkspaceLoader(db, storage)({ companyId: ids.company, workspaceId: reserved.id, snapshotId: saved.snapshotId });
try { expect(await fs.readFile(path.join(restored.root, "saved.txt"), "utf8")).toBe("work survives a failed provider session\n"); }
finally { await restored.cleanup(); }
} finally { unsubscribe(); }
}
if (kind === "historical_listener_bad_preservation") {
expect(await getExecutionBlocker(db, ids.company, ids.issue)).toMatchObject({ cause: "native_continuation_requires_reconciliation" });
}
if (processAcceptance) {
expect(await reconcileSafeNativeReplacements(db)).toMatchObject({ scheduled: 0 });
const { completeHistoricalListenerSuccessor } = await import("../__tests__/helpers/historical-listener-continuation.js");
await completeHistoricalListenerSuccessor({ db, runId: successors[0]!.id, root: path.join(root, "successor-runtime"),
prepareWorkspace: async () => {
const leaseId = randomUUID(), runId = successors[0]!.id;
const [nextLease] = await db.insert(environmentLeases).values({ id: leaseId, companyId: ids.company,
environmentId: ids.environment, heartbeatRunId: runId, issueId: ids.issue, provider: "daytona",
providerLeaseId: `allocation-${leaseId}`, status: "active" }).returning();
const next = await reserveRemoteWorkspace({ db, storage, companyId: ids.company, issueId: ids.issue,
runId, agentId: ids.agent, environmentId: ids.environment, projectId: null, projectWorkspaceId: null,
existingWorkspaceId: reserved.id, source: { kind: "empty" } });
const nextTarget = { ...await target("historical-successor"), leaseId, sandboxId: nextLease.providerLeaseId! };
const nextPrepared = await prepareRemoteWorkspace({ db, storage, workspace: next, runId,
lease: nextLease as EnvironmentLease, target: nextTarget });
return { cwd: nextPrepared.remoteRoot, finalize: async () => {
const { workspaceOperationService } = await import("./workspace-operations.js");
const recorder = workspaceOperationService(db).createRecorder({ companyId: ids.company, issueId: ids.issue, heartbeatRunId: runId });
const saved = await recorder.recordOperation({ phase: "workspace_finalize", cwd: nextPrepared.remoteRoot, run: async () => {
await saveRemoteWorkspace({ db, storage, workspace: nextPrepared.workspace, runId, lease: nextPrepared.lease,
target: nextTarget, assertOwned: async () => undefined });
return { status: "succeeded", exitCode: 0 };
} });
expect(saved.status).toBe("succeeded");
} };
} });
}
if (!shouldSchedule) expect(await fs.readFile(path.join(prepared.remoteRoot, "saved.txt"), "utf8"))
.toBe("work survives a failed provider session\n");
if (scenario === "unknown_action") {
const [decision] = await db.select().from(nativeRunFinalizations).where(eq(nativeRunFinalizations.runId, ids.run));
expect(decision?.failureDetail?.replacementDenied).toBe("uncertain_provider_action");
}
}, 60_000,
);
});