Merge remote-tracking branch 'origin/master' into web-e2e-interactions

# Conflicts: # .agents/notes/implemented/testing/2026-07-24-web-gui-browser-e2e-lane.i18n.yaml # .agents/notes/implemented/testing/2026-07-24-web-gui-browser-e2e-lane.md # .agents/notes/implemented/testing/2026-07-24-web-gui-browser-e2e-lane.zh.md # apps/web/tests/scaffold.ts
2026-07-26 04:10:37 +08:00
parent 96c67df835 3ccdf4ac46
commit 07770f90f9
270 changed files with 10688 additions and 2222 deletions
--- a/apps/web/tests/replay-round-trip.e2e.ts
+++ b/apps/web/tests/replay-round-trip.e2e.ts
@@ -80,9 +80,16 @@ describe('web e2e: fresh round trip through the real assembly', () => {
      // legal — the chunk-event assertions below carry incrementality.
    })
    await expect.poll(() => page.getByText('DONE', { exact: true }).count(), { timeout: 15_000 }).toBeGreaterThanOrEqual(1)
-    // World state, not self-report: bash really ran and the turn closed clean.
-    const toolCalls = sessionEvents.filter(e => e.type === 'tool/call')
-    expect(toolCalls.map(e => (e as SessionEvent & { data: { name: string } }).data.name)).toContain('bash')
+    // World state, not self-report: the real bash executor returned the exact
+    // command output, and the turn closed cleanly.
+    const bashCall = sessionEvents.find(event => event.type === 'tool/call' && event.data.name === 'bash')
+    if (bashCall?.type !== 'tool/call') throw new Error('the replayed turn did not call the bash tool')
+    const bashResult = sessionEvents.find(event =>
+      event.type === 'tool/result' && event.data.callId === bashCall.data.callId)
+    if (bashResult?.type !== 'tool/result') throw new Error('the bash tool call produced no durable result')
+    expect(bashResult.data.isError).toBe(false)
+    expect(bashResult.data.content.filter(block => block.type === 'text').map(block => block.text).join(''))
+      .toBe('WEB_E2E_OK\n')
    const turnEnds = sessionEvents.filter(e => e.type === 'turn/end')
    expect(turnEnds.length).toBe(1)
    expect((turnEnds[0] as SessionEvent & { data: { reason: { kind: string } } }).data.reason.kind).toBe('completed')
--- a/apps/web/tests/scaffold.ts
+++ b/apps/web/tests/scaffold.ts
@@ -111,6 +111,15 @@ export interface LaunchOptions {
  paceMs?: number
 }

+/** Dispose the booted tree and remove both owned temp roots, reporting every independent cleanup failure. */
+async function cleanupScaffoldWorld(ctx: Context, workspaceCwd: string, persistenceRoot: string): Promise<unknown[]> {
+  const failures: unknown[] = []
+  await Promise.resolve(ctx.fiber.dispose()).catch((error: unknown) => failures.push(error))
+  await rm(workspaceCwd, { recursive: true, force: true }).catch((error: unknown) => failures.push(error))
+  await rm(persistenceRoot, { recursive: true, force: true }).catch((error: unknown) => failures.push(error))
+  return failures
+}
+
 /**
 * Boot the real web composition under the current snapshot mode.
 * @param options - replay fixture selection and pacing.
@@ -126,7 +135,15 @@ export async function launchWebScaffold(options: LaunchOptions = {}): Promise<We
    }
  }
  const workspaceCwd = await mkdtemp(join(tmpdir(), 'dsh-web-e2e-ws-'))
-  const persistenceRoot = await mkdtemp(join(tmpdir(), 'dsh-web-e2e-sessions-'))
+  let persistenceRoot: string
+  try {
+    persistenceRoot = await mkdtemp(join(tmpdir(), 'dsh-web-e2e-sessions-'))
+  } catch (error) {
+    const failures: unknown[] = [error]
+    await rm(workspaceCwd, { recursive: true, force: true }).catch((cleanupError: unknown) => failures.push(cleanupError))
+    if (failures.length > 1) throw new AggregateError(failures, 'web scaffold temp-root setup failed')
+    throw error
+  }

  // The include patch set — the same mechanism AppCLIEntry and the ACP
  // snapshot overlay use, applied over the SAME shipped tree (a patch id that
@@ -149,9 +166,11 @@ export async function launchWebScaffold(options: LaunchOptions = {}): Promise<We
  // Sessions inherit the gateway's process.cwd() default; run the boot from
  // the temp workspace so tool cwd, session cwd, and fixtures agree.
  const originalCwd = process.cwd()
-  process.chdir(workspaceCwd)
  const ctx = new Context()
+  let port = 0
+  let replayHandle: ReplayHandle | undefined
  try {
+    process.chdir(workspaceCwd)
    ctx.baseUrl = pathToFileURL(join(resolve(CONFIG_PATH), '..')).href + '/'
    await ctx.plugin(Loader)
    ctx.loader.builtins.include = Include
@@ -161,35 +180,35 @@ export async function launchWebScaffold(options: LaunchOptions = {}): Promise<We
    })
    await ctx.loader.await()
    assertEntriesLoaded(ctx, 'web e2e scaffold')
+    const boundPort = ctx.get('httpServer')?.port
+    if (boundPort === undefined) {
+      throw new Error('web e2e scaffold: httpServer service missing after settled boot')
+    }
+    port = boundPort
+
+    // Fill the open llm seam on the settled root ctx (llm-deepseek is disabled
+    // in keyless modes; a scenario with no fixture leaves the seam empty so a
+    // stray stream fails loud with NO_ADAPTER). The direct install, unlike the
+    // plugin row, returns the ReplayHandle for the teardown consumption check.
+    if (mode !== 'record' && options.replayFixture !== undefined) {
+      replayHandle = installLlmReplay(ctx, {
+        file: options.replayFixture,
+        providers: REPLAY_PROVIDERS,
+        ...(options.replayOverride === undefined ? {} : { overrideFile: options.replayOverride }),
+        ...(options.paceMs === undefined ? {} : { paceMs: options.paceMs }),
+      })
+    }
  } catch (error) {
-    process.chdir(originalCwd)
-    await ctx.fiber.dispose()
-    await rm(workspaceCwd, { recursive: true, force: true }).catch(() => undefined)
-    await rm(persistenceRoot, { recursive: true, force: true }).catch(() => undefined)
+    if (process.cwd() !== originalCwd) process.chdir(originalCwd)
+    const cleanupFailures = await cleanupScaffoldWorld(ctx, workspaceCwd, persistenceRoot)
+    if (cleanupFailures.length > 0) {
+      throw new AggregateError([error, ...cleanupFailures], 'web scaffold setup failed and cleanup was incomplete')
+    }
    throw error
  } finally {
    if (process.cwd() !== originalCwd) process.chdir(originalCwd)
  }

-  const port = ctx.get('httpServer')?.port
-  if (port === undefined) {
-    await ctx.fiber.dispose()
-    throw new Error('web e2e scaffold: httpServer service missing after settled boot')
-  }
-  // Fill the open llm seam on the settled root ctx (llm-deepseek is disabled
-  // in keyless modes; a scenario with no fixture leaves the seam empty so a
-  // stray stream fails loud with NO_ADAPTER). The direct install, unlike the
-  // plugin row, returns the ReplayHandle for the teardown consumption check.
-  let replayHandle: ReplayHandle | undefined
-  if (mode !== 'record' && options.replayFixture !== undefined) {
-    replayHandle = installLlmReplay(ctx, {
-      file: options.replayFixture,
-      providers: REPLAY_PROVIDERS,
-      ...(options.replayOverride === undefined ? {} : { overrideFile: options.replayOverride }),
-      ...(options.paceMs === undefined ? {} : { paceMs: options.paceMs }),
-    })
-  }
-
  return {
    mode,
    baseUrl: `http://127.0.0.1:${port}`,
@@ -229,9 +248,7 @@ export async function launchWebScaffold(options: LaunchOptions = {}): Promise<We
      } catch (error) {
        failures.push(error)
      }
-      await Promise.resolve(ctx.fiber.dispose()).catch((e: unknown) => failures.push(e))
-      await rm(workspaceCwd, { recursive: true, force: true }).catch((e: unknown) => failures.push(e))
-      await rm(persistenceRoot, { recursive: true, force: true }).catch((e: unknown) => failures.push(e))
+      failures.push(...await cleanupScaffoldWorld(ctx, workspaceCwd, persistenceRoot))
      if (failures.length > 0) throw new AggregateError(failures, 'web scaffold teardown failed')
    },
  }
@@ -254,9 +271,9 @@ function rawSessionLog(session: Session): string {
 * Record-mode fixture write-back: harvest the live session, scrub request
 * headers to {{system}}/{{tools}} (TODO(web-header-pin): the web lane pins no
 * header class — a deliberate deviation logged in the Agent Note's deferred
- * work), tokenize the run-local session id and cwd ({{sessionId}}/{{cwd}},
- * the committed ACP fixture convention — re-records then diff only on real
- * content), and write the committed fixture.
+ * work), tokenize the run-local session id, cwd, and browser RPC id
+ * ({{sessionId}}/{{cwd}}/{{rpcId}}, the committed fixture convention —
+ * re-records then diff only on real content), and write the fixture.
 * @param scaffold - the record-mode scaffold.
 * @param sessionId - the driven session.
 * @param fixturePath - the committed session.jsonl / seed.jsonl target.
@@ -267,6 +284,7 @@ export async function recordFixture(scaffold: WebScaffold, sessionId: SessionId,
  const tokenized = scrubRequestHeaders(rawSessionLog(agent.session))
    .split(sessionId).join('{{sessionId}}')
    .split(scaffold.workspaceCwd).join('{{cwd}}')
+    .replace(/"rpcId":"[^"]+"/g, '"rpcId":"{{rpcId}}"')
  await writeFile(fixturePath, tokenized)
 }

@@ -396,7 +414,7 @@ export async function compareOrRefreshGolden(goldenPath: string, actual: string,
 /**
 * Fixture-inventory guard (the TUI afterAll shape): the scenario directory
 * holds exactly the expected files and every committed JSONL is a scrub
- * fixed-point (no request-header bulk escaped the record write-back).
+ * fixed-point without a run-local browser RPC id.
 * @param dir - the scenario snapshot directory.
 * @param expected - the exact expected file inventory.
 */
@@ -406,6 +424,8 @@ export async function assertFixtureInventory(dir: string, expected: string[]): P
  for (const entry of entries.filter(name => name.endsWith('.jsonl'))) {
    const content = await readFile(join(dir, entry), 'utf8')
    expect(scrubRequestHeaders(content), `${dir}/${entry} carries request-header bulk`).toBe(content)
+    expect(content, `${dir}/${entry} carries a run-local rpcId`)
+      .not.toMatch(/"rpcId":"(?!\{\{rpcId\}\})[^"]+"/)
  }
 }

--- a/apps/web/tests/snapshots/fresh-round-trip/session.jsonl
+++ b/apps/web/tests/snapshots/fresh-round-trip/session.jsonl
@@ -1,6 +1,6 @@
 {"type":"session","version":0,"id":"{{sessionId}}","createdAt":1784973850091,"cwd":"{{cwd}}/workspace"}
-{"type":"turn/start","seq":0,"time":1784973850102,"data":{"turn":1,"trigger":{"kind":"message","source":{"kind":"user","rpcId":"c4e068dc-c277-46ae-9713-6a2694027fb5"}}}}
-{"type":"user/message","seq":1,"time":1784973850103,"data":{"content":[{"type":"text","text":"Use the bash tool to run exactly: echo WEB_E2E_OK. Then reply with the single word DONE and stop."}],"source":{"kind":"user","rpcId":"c4e068dc-c277-46ae-9713-6a2694027fb5"}},"surfaceOp":"append"}
+{"type":"turn/start","seq":0,"time":1784973850102,"data":{"turn":1,"trigger":{"kind":"message","source":{"kind":"user","rpcId":"{{rpcId}}"}}}}
+{"type":"user/message","seq":1,"time":1784973850103,"data":{"content":[{"type":"text","text":"Use the bash tool to run exactly: echo WEB_E2E_OK. Then reply with the single word DONE and stop."}],"source":{"kind":"user","rpcId":"{{rpcId}}"}},"surfaceOp":"append"}
 {"type":"session/title","seq":2,"time":1784973850105,"data":{"title":"Use the bash tool to","messageSeqs":[1],"source":{"kind":"fallback"}}}
 {"type":"step/start","seq":3,"time":1784973850164,"data":{"turn":1,"step":1}}
 {"type":"request/header","seq":4,"time":1784973850165,"data":{"header":{"config":{"provider":"deepseek","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}","messagePrefix":["{{messagePrefix}}"]},"reason":"initial"}}
--- a/apps/web/tests/snapshots/seeded-history/seed.jsonl
+++ b/apps/web/tests/snapshots/seeded-history/seed.jsonl
@@ -1,6 +1,6 @@
 {"type":"session","version":0,"id":"{{sessionId}}","createdAt":1784974100747,"cwd":"{{cwd}}/workspace"}
-{"type":"turn/start","seq":0,"time":1784974100758,"data":{"turn":1,"trigger":{"kind":"message","source":{"kind":"user","rpcId":"f95c6f1c-f1b4-42bf-ba40-c05ae0647a70"}}}}
-{"type":"user/message","seq":1,"time":1784974100759,"data":{"content":[{"type":"text","text":"Use the read tool twice in one assistant message: read a.txt and b.txt. Then reply with the single word DONE and stop."}],"source":{"kind":"user","rpcId":"f95c6f1c-f1b4-42bf-ba40-c05ae0647a70"}},"surfaceOp":"append"}
+{"type":"turn/start","seq":0,"time":1784974100758,"data":{"turn":1,"trigger":{"kind":"message","source":{"kind":"user","rpcId":"{{rpcId}}"}}}}
+{"type":"user/message","seq":1,"time":1784974100759,"data":{"content":[{"type":"text","text":"Use the read tool twice in one assistant message: read a.txt and b.txt. Then reply with the single word DONE and stop."}],"source":{"kind":"user","rpcId":"{{rpcId}}"}},"surfaceOp":"append"}
 {"type":"session/title","seq":2,"time":1784974100761,"data":{"title":"Use the read tool twice","messageSeqs":[1],"source":{"kind":"fallback"}}}
 {"type":"step/start","seq":3,"time":1784974100827,"data":{"turn":1,"step":1}}
 {"type":"request/header","seq":4,"time":1784974100828,"data":{"header":{"config":{"provider":"deepseek","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}","messagePrefix":["{{messagePrefix}}"]},"reason":"initial"}}