chore(eve): workflow - upgrade to beta.44

.changeset/upgrade-workflow-beta-44.md

@@ -0,0 +1,5 @@
+---
+"eve": patch
+---
+
+Upgrade the bundled Workflow DevKit packages to the latest 5.0 beta releases, including sealed-log event ordering and long-polling run completion.

e2e/fixtures/agent-workflow-stress/evals/sequential-workflow-turns.eval.ts

@@ -6,6 +6,7 @@ const TURN_COUNT = 100;
 export default defineEval({
   description: "Workflow stress: one durable session completes 100 sequential turns.",
   tags: ["stress", "workflow", "sequential"],
+  timeoutMs: 900_000,
 
   async test(t) {
     let sessionId: string | undefined;

e2e/fixtures/fixture-tasks/agent/subagents/busy-worker/tools/hold.ts

@@ -1,11 +1,10 @@
 import { defineTool } from "eve/tools";
+import { once } from "eve/tools/approval";
 import { z } from "zod";
 
 export default defineTool({
-  description: "Keep an admitted continuation nonterminal long enough for a later-turn check.",
+  description: "Keep an admitted continuation nonterminal until its parent releases it.",
   inputSchema: z.object({ marker: z.literal("HOLD") }),
-  execute: async ({ marker }) => {
-    await new Promise((resolve) => setTimeout(resolve, 5_000));
-    return { marker, released: true };
-  },
+  approval: once(),
+  execute: async ({ marker }) => ({ marker, released: true }),
 });

e2e/fixtures/fixture-tasks/evals/task.agent.continue.rejected-agent-busy.eval.ts

@@ -4,6 +4,7 @@ import {
   parseToolErrorOutput,
   sendAndFollowQueuedTurn,
   waitForCompletedTask,
+  waitForTaskInput,
 } from "./shared.js";
 
 /** A persistent child with a nonterminal task rejects every competing continuation. */
@@ -14,6 +15,8 @@ export default defineTaskEval({
     primary: "task.agent.continue.rejected-agent-busy",
     setup: [
       "task.dispatch.start.accepted-acknowledged",
+      "task.input.require.accepted-valid-batch",
+      "task.input.answer.accepted-complete",
       "task.lifecycle.complete.accepted-nonterminal",
       "task.agent.continue.accepted-terminal-available",
     ],
@@ -69,7 +72,17 @@ export default defineTaskEval({
       status: "failed",
     });
 
-    await waitForCompletedTask(t, later.session, "CHILD-TASK-EXCLUSIVITY-VERIFY", admittedTaskId);
+    const blocked = await waitForTaskInput(t, later.session, "hold");
+    const released = await blocked.session.respond([
+      {
+        optionId: "approve",
+        requestId: blocked.request.requestId,
+      },
+    ]);
+    released.expectOk();
+    released.noFailedActions();
+
+    await waitForCompletedTask(t, blocked.session, "CHILD-TASK-EXCLUSIVITY-VERIFY", admittedTaskId);
   },
 });
 

e2e/fixtures/fixture-tasks/evals/task.parent.wake.emitted-ready.conditional-delivery.eval.ts

@@ -1,8 +1,12 @@
-import type { EveEvalTurn } from "eve/evals";
+import type { EveEvalContext, EveEvalTurn } from "eve/evals";
 import { satisfies } from "eve/evals/expect";
 
 import { defineTaskEval } from "./task-transition.js";
-import { requireSessionStreamIndex } from "./shared.js";
+import {
+  requireSessionStreamIndex,
+  type TaskEvalSessionDriver,
+  waitForTaskInput,
+} from "./shared.js";
 
 const REVIEW_FINDING = "blocker: task admission can discard deferred user input.";
 
@@ -14,6 +18,8 @@ export default defineTaskEval({
     primary: "task.parent.wake.emitted-ready",
     setup: [
       "task.dispatch.start.accepted-acknowledged",
+      "task.input.require.accepted-valid-batch",
+      "task.input.answer.accepted-complete",
       "task.lifecycle.complete.accepted-nonterminal",
     ],
     dimensions: { transport: "local", parentPhase: "parked" },
@@ -36,11 +42,14 @@ export default defineTaskEval({
       ),
     );
 
-    const sessionId = started.sessionId;
-    const firstLive = t.target.watchTurn(sessionId, {
-      startIndex: requireSessionStreamIndex(t, "First reviewer wake"),
-    });
-    const firstWake = await firstLive.result();
+    const blocked = await waitForTaskInput(t, t, "hold");
+    const firstObserved = await waitForTaskNotification(
+      t,
+      blocked.session,
+      blocked.observedTurns,
+      "First reviewer wake",
+    );
+    const firstWake = firstObserved.turn;
 
     firstWake.expectOk();
     const firstNotification = requireTaskNotification(firstWake);
@@ -65,10 +74,22 @@ export default defineTaskEval({
       satisfies((message) => message === undefined, "the pending cohort wake is silent"),
     );
 
-    const secondLive = t.target.watchTurn(sessionId, {
-      startIndex: requireSessionStreamIndex(firstLive.session, "Late reviewer wake"),
-    });
-    const wake = await secondLive.result();
+    const released = await firstObserved.session.respond([
+      {
+        optionId: "approve",
+        requestId: blocked.request.requestId,
+      },
+    ]);
+    released.expectOk();
+    released.noFailedActions();
+
+    const lateObserved = await waitForTaskNotification(
+      t,
+      firstObserved.session,
+      [released],
+      "Late reviewer wake",
+    );
+    const wake = lateObserved.turn;
 
     wake.expectOk();
     const lateNotification = requireTaskNotification(wake);
@@ -98,11 +119,43 @@ interface TaskNotification {
   readonly taskId: string;
 }
 
+interface ObservedTaskNotification {
+  readonly session: TaskEvalSessionDriver;
+  readonly turn: EveEvalTurn;
+}
+
+async function waitForTaskNotification(
+  t: EveEvalContext,
+  initialSession: TaskEvalSessionDriver,
+  observedTurns: readonly EveEvalTurn[],
+  operation: string,
+): Promise<ObservedTaskNotification> {
+  const observed = observedTurns.find((turn) => taskNotification(turn) !== undefined);
+  if (observed !== undefined) return { session: initialSession, turn: observed };
+
+  const sessionId = initialSession.sessionId;
+  if (sessionId === undefined) throw new Error(`${operation} has no parent session id.`);
+  const live = t.target.watchTurn(sessionId, {
+    startIndex: requireSessionStreamIndex(initialSession, operation),
+  });
+  const turn = await live.result();
+  if (taskNotification(turn) === undefined) {
+    throw new Error(`${operation} has no completed-task notification.`);
+  }
+  return { session: live.session, turn };
+}
+
 function requireTaskNotification(turn: EveEvalTurn): TaskNotification {
+  const notification = taskNotification(turn);
+  if (notification !== undefined) return notification;
+  throw new Error("Reviewer wake has no completed-task notification.");
+}
+
+function taskNotification(turn: EveEvalTurn): TaskNotification | undefined {
   for (const event of turn.events) {
     if (event.type !== "message.received" || typeof event.data.message !== "string") continue;
     const taskId = /Background task (task_[a-z0-9]+)/iu.exec(event.data.message)?.[1];
     if (taskId !== undefined) return { message: event.data.message, taskId };
   }
-  throw new Error("Reviewer wake has no completed-task notification.");
+  return undefined;
 }

packages/eve/package.json

@@ -420,13 +420,13 @@
     "@vercel/sandbox": "catalog:",
     "@vercel/sandbox-drives": "catalog:",
     "@vercel/sdk": "1.28.8",
-    "@workflow/core": "5.0.0-beta.43",
-    "@workflow/errors": "5.0.0-beta.17",
+    "@workflow/core": "5.0.0-beta.44",
+    "@workflow/errors": "5.0.0-beta.18",
     "@workflow/serde": "5.0.0-beta.2",
-    "@workflow/utils": "5.0.0-beta.8",
-    "@workflow/world": "5.0.0-beta.28",
-    "@workflow/world-local": "5.0.0-beta.37",
-    "@workflow/world-vercel": "5.0.0-beta.39",
+    "@workflow/utils": "5.0.0-beta.9",
+    "@workflow/world": "5.0.0-beta.29",
+    "@workflow/world-local": "5.0.0-beta.38",
+    "@workflow/world-vercel": "5.0.0-beta.40",
     "ai": "catalog:",
     "autoevals": "0.0.132",
     "chat": "4.34.0",

packages/eve/src/internal/workflow/development-world-client.ts

@@ -89,7 +89,7 @@ async function call<T>(
 export function createDevelopmentWorkflowWorld(): World {
   const forwarded = buildForwardedOperations();
   const world = {
-    specVersion: 6 as SpecVersion,
+    specVersion: 7 as SpecVersion,
     async getDeploymentId() {
       // Inside a pinned delivery, steps and child runs must record the
       // delivery's generation — not whatever is active — so replay after a

packages/eve/test/scenarios/compiled-vendor-assets.scenario.test.ts

@@ -270,7 +270,7 @@ describe("compiled vendor assets", () => {
         readFile(join(COMPILED_VENDOR_ROOT, "@workflow/core/runtime/run.d.ts"), "utf8"),
       ]);
 
-    expect(indexDts).toContain("Just the core utilities");
+    expect(indexDts).toContain("Core utilities intended for import by user");
     expect(indexDts).toContain("from '#compiled/@workflow/errors/index.js'");
     expect(createHookDts).toContain("Creates a {@link Hook}");
     expect(workflowDts).toBe(`export * from "./workflow/index.js";\n`);