diff --git a/tests/restart-chaos/README.md b/tests/restart-chaos/README.md index b457d35b03f..f0e1731365d 100644 --- a/tests/restart-chaos/README.md +++ b/tests/restart-chaos/README.md @@ -511,7 +511,10 @@ windows are written to `storage-faults.json` and removed from the speed grade li Every Retry the grader needs is clicked in the real board UI: the harness opens the issue in a headless Chromium, waits for the visible Retry button and clicks it. A task whose Retry button is missing fails R3 as `ui_retry_button_missing (user saw: …)`; a click that sends no retry request fails as -`ui_retry_no_request`. Crash-point waves serve no UI and click Retry through the API. +`ui_retry_no_request`. Before it looks for Retry, the harness closes any open modal that covers the page +("Not now", else Escape) and records it in the attempt's `dismissedDialogs`. A visible Retry that stays +unreachable behind a modal that will not close fails as `ui_retry_blocked_by_modal`. Crash-point waves +serve no UI and click Retry through the API. ## Daytona (`--env daytona`) diff --git a/tests/restart-chaos/board-ui.test.ts b/tests/restart-chaos/board-ui.test.ts index 2a5ba83b88a..b8897d8f42f 100644 --- a/tests/restart-chaos/board-ui.test.ts +++ b/tests/restart-chaos/board-ui.test.ts @@ -4,10 +4,24 @@ import { tmpdir } from "node:os"; import path from "node:path"; import { once } from "node:events"; import { afterAll, beforeAll, describe, expect, it } from "vitest"; -import { BoardUi, UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST, isRetryRequestFor } from "./board-ui.js"; +import { BoardUi, UI_RETRY_BLOCKED_BY_MODAL, UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST, isRetryRequestFor } from "./board-ui.js"; const retry = (issue: string) => `fetch('/api/issues/${issue}/retry',{method:'POST'})`; +const modal = (buttons: string) => `
+

You completed 5 tasks!

${buttons}
`; +const coveredRetry = (issue: string) => `
+
Couldn’t finish this attempt.
`; +const closeModal = "document.querySelector('[role=dialog]').remove();document.querySelector('main').removeAttribute('aria-hidden')"; +const lateModal = `
+

You completed 5 tasks!

`; +const openLateModal = `document.querySelector('main').setAttribute('aria-hidden','true');document.body.insertAdjacentHTML('beforeend', ${JSON.stringify(lateModal).replace(/"/g, """)})`; const pages: Record = { + "/issues/CHA-8": coveredRetry("i8") + modal(``), + "/issues/CHA-9": coveredRetry("i9") + modal(`No way out`), + "/issues/CHA-11": `
Couldn’t finish this attempt. +
`, + "/issues/CHA-10": coveredRetry("i10") + modal(`Press Escape`) + + ``, "/issues/CHA-1": `
Couldn’t finish this attempt.
`, "/issues/CHA-2": `
Saving files from the previous attempt
`, "/issues/CHA-7": `
Couldn’t finish this attempt.
`, @@ -89,6 +103,38 @@ describe("BoardUi.clickRetry", () => { } }, 60_000); + it("dismisses a prompt that covers the Retry with Not now, then clicks Retry", async () => { + await withBoard(async board => { + expect(await board.clickRetry("CHA-8", "i8", 5_000, 1_000)).toMatchObject({ clicked: true, httpStatus: 202, accepted: true, error: null, + dismissedDialogs: [expect.stringContaining("You completed 5 tasks!")] }); + }); + expect(posts).toContain("/api/issues/i8/retry"); + }, 60_000); + + it("closes a prompt without a Not now button with Escape", async () => { + await withBoard(async board => { + expect(await board.clickRetry("CHA-10", "i10", 5_000, 1_000)).toMatchObject({ clicked: true, accepted: true, + dismissedDialogs: [expect.stringContaining("You completed 5 tasks!")] }); + }); + expect(posts).toContain("/api/issues/i10/retry"); + }, 60_000); + + it("closes a prompt that opens after Retry appeared, then clicks Retry", async () => { + await withBoard(async board => { + expect(await board.clickRetry("CHA-11", "i11", 5_000, 1_500)).toMatchObject({ clicked: true, httpStatus: 202, accepted: true, error: null, + dismissedDialogs: [expect.stringContaining("You completed 5 tasks!")] }); + }); + expect(posts).toContain("/api/issues/i11/retry"); + }, 60_000); + + it("names a modal that keeps a visible Retry unreachable", async () => { + await withBoard(async board => { + expect(await board.clickRetry("CHA-9", "i9", 2_000, 500)).toMatchObject({ clicked: false, error: UI_RETRY_BLOCKED_BY_MODAL, + bannerText: expect.stringContaining("Couldn’t finish this attempt") }); + }); + expect(posts).not.toContain("/api/issues/i9/retry"); + }, 60_000); + it("reports a Retry button whose click sends no retry request", async () => { await withBoard(async board => { expect(await board.clickRetry("CHA-3", "i3", 1_500, 500)).toMatchObject({ clicked: true, httpStatus: null, error: UI_RETRY_NO_REQUEST }); diff --git a/tests/restart-chaos/board-ui.ts b/tests/restart-chaos/board-ui.ts index 90390b45288..3677360e8a6 100644 --- a/tests/restart-chaos/board-ui.ts +++ b/tests/restart-chaos/board-ui.ts @@ -1,9 +1,9 @@ import { mkdir, writeFile } from "node:fs/promises"; import path from "node:path"; import { chromium, type Browser, type ElementHandle, type Page } from "@playwright/test"; -import { UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST } from "./types.js"; +import { UI_RETRY_BLOCKED_BY_MODAL, UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST } from "./types.js"; -export { UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST }; +export { UI_RETRY_BLOCKED_BY_MODAL, UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST }; export interface UiRetryClick { clicked: boolean; @@ -12,6 +12,7 @@ export interface UiRetryClick { error: string | null; bannerText: string | null; diagnostics?: { screenshot: string | null; dom: string | null } | null; + dismissedDialogs?: string[]; } export async function saveRetryMissingDiagnostics(page: Page, directory: string, label: string) { @@ -37,6 +38,28 @@ export async function saveRetryMissingDiagnostics(page: Page, directory: string, } const PER_BUTTON_REQUEST_MS = 5_000; +const DIALOG_SCAN_MS = 1_000; +const MAX_DISMISSED_DIALOGS = 3; +const BLOCKING_DIALOG = '[role="dialog"][data-state="open"], [role="alertdialog"][data-state="open"], ' + + '[aria-modal="true"]:not([data-state="closed"]), dialog[open]'; +const DISMISS_BUTTON = /^(not now|no thanks|maybe later|dismiss)$/i; + +const blockingDialog = (page: Page) => page.locator(BLOCKING_DIALOG).filter({ visible: true }).first(); + +async function dismissBlockingDialog(page: Page): Promise { + const dialog = blockingDialog(page); + if (!await dialog.isVisible().catch(() => false)) return null; + if (await dialog.getByRole("button", { name: "Retry", exact: true }).count().catch(() => 0)) return null; + const label = (await dialog.innerText().catch(() => "")).trim().split(/\r?\n/)[0]?.slice(0, 120) || "dialog"; + const handle = await dialog.elementHandle().catch(() => null); + const notNow = dialog.getByRole("button", { name: DISMISS_BUTTON }).filter({ visible: true }).first(); + const clicked = await notNow.count().catch(() => 0) > 0 && await notNow.click({ timeout: 2_000 }).then(() => true, () => false); + if (!clicked) await page.keyboard.press("Escape").catch(() => undefined); + const closed = handle + ? await handle.waitForElementState("hidden", { timeout: 2_000 }).then(() => true, () => false) + : !await blockingDialog(page).isVisible().catch(() => false); + return closed ? label : null; +} export function isRetryRequestFor(issueRefs: readonly string[], method: string, url: string): boolean { if (method !== "POST") return false; @@ -61,28 +84,55 @@ export async function clickRetryOnPage(page: Page, url: string, issueRefs: reado const texts = await page.locator("[role=status], [role=alert]").allInnerTexts().catch(() => []); return texts.map(text => text.trim()).filter(Boolean).join(" | ").slice(0, 500) || null; }; - try { - await page.getByRole("button", { name: "Retry", exact: true }).filter({ visible: true }).first() - .waitFor({ state: "visible", timeout: timeoutMs }); - } catch { + const reachable = page.getByRole("button", { name: "Retry", exact: true }).filter({ visible: true }).first(); + const dismissedDialogs: string[] = []; + const dismissed = () => (dismissedDialogs.length ? { dismissedDialogs } : {}); + const dismissDialogs = async () => { + const before = dismissedDialogs.length; + while (dismissedDialogs.length < MAX_DISMISSED_DIALOGS) { + const closed = await dismissBlockingDialog(page); + if (closed === null) break; + dismissedDialogs.push(closed); + } + return dismissedDialogs.length > before; + }; + const blockedByModal = async () => await blockingDialog(page).isVisible().catch(() => false) + && await page.getByRole("button", { name: "Retry", exact: true, includeHidden: true }).filter({ visible: true }).count().catch(() => 0) > 0; + const deadline = Date.now() + timeoutMs; + for (;;) { + await dismissDialogs(); + const remaining = deadline - Date.now(); + if (remaining <= 0) break; + if (await reachable.waitFor({ state: "visible", timeout: Math.min(remaining, DIALOG_SCAN_MS) }).then(() => true, () => false)) break; + } + if (!await reachable.isVisible().catch(() => false)) { const bannerText = await banner(); const diagnostics = diagnosticsDirectory ? await saveRetryMissingDiagnostics(page, diagnosticsDirectory, issueRefs[0] ?? "issue").catch(() => null) : undefined; - return { clicked: false, httpStatus: null, accepted: null, error: UI_RETRY_BUTTON_MISSING, bannerText, - ...(diagnostics !== undefined ? { diagnostics } : {}) }; + const blocked = await blockedByModal(); + return { clicked: false, httpStatus: null, accepted: null, error: blocked ? UI_RETRY_BLOCKED_BY_MODAL : UI_RETRY_BUTTON_MISSING, bannerText, + ...(diagnostics !== undefined ? { diagnostics } : {}), ...dismissed() }; } const bannerText = await banner(); for (const button of await retryButtonsBannerFirst(page)) { - const response = page.waitForResponse(candidate => isRetryRequestFor(issueRefs, candidate.request().method(), candidate.url()), - { timeout: perButtonMs }).catch(() => null); - await button.click().catch(() => undefined); - const answered = await response; - if (!answered) continue; - const body = await answered.json().catch(() => null) as { accepted?: unknown } | null; - return { clicked: true, httpStatus: answered.status(), accepted: answered.ok() ? body?.accepted === true : false, - error: answered.ok() ? null : `retry request failed with ${answered.status()}`, bannerText }; + for (let attempt = 0; attempt < 2; attempt++) { + await dismissDialogs(); + const response = page.waitForResponse(candidate => isRetryRequestFor(issueRefs, candidate.request().method(), candidate.url()), + { timeout: perButtonMs }).catch(() => null); + await button.click({ timeout: perButtonMs }).catch(() => undefined); + const answered = await response; + if (answered) { + const body = await answered.json().catch(() => null) as { accepted?: unknown } | null; + return { clicked: true, httpStatus: answered.status(), accepted: answered.ok() ? body?.accepted === true : false, + error: answered.ok() ? null : `retry request failed with ${answered.status()}`, bannerText, ...dismissed() }; + } + if (!await dismissDialogs()) break; + } } - return { clicked: true, httpStatus: null, accepted: null, error: UI_RETRY_NO_REQUEST, bannerText }; + if (await blockedByModal()) { + return { clicked: false, httpStatus: null, accepted: null, error: UI_RETRY_BLOCKED_BY_MODAL, bannerText, ...dismissed() }; + } + return { clicked: true, httpStatus: null, accepted: null, error: UI_RETRY_NO_REQUEST, bannerText, ...dismissed() }; } export class BoardUi { diff --git a/tests/restart-chaos/grader.test.ts b/tests/restart-chaos/grader.test.ts index cf6aed66e04..f6495d49e68 100644 --- a/tests/restart-chaos/grader.test.ts +++ b/tests/restart-chaos/grader.test.ts @@ -244,6 +244,8 @@ describe("T-oracles", () => { retry: { at: 2, observedRunId: "r1", httpStatus: null, accepted: null, error: "ui_retry_button_missing", via: "ui", bannerText: "Saving files" } } as unknown as TaskRecord; expect(gradeR3(record)).toEqual({ verdict: "fail", detail: "ui_retry_button_missing (user saw: Saving files)" }); expect(gradeR3({ ...record, retry: { ...record.retry!, clicked: true, error: "ui_retry_no_request" } } as TaskRecord).detail).toMatch(/^ui_retry_no_request/); + expect(gradeR3({ ...record, retry: { ...record.retry!, error: "ui_retry_blocked_by_modal" } } as TaskRecord).detail) + .toBe("ui_retry_blocked_by_modal (user saw: Saving files)"); }); it("reports R2 as not checked for a sandbox-only workspace instead of passing it", () => { diff --git a/tests/restart-chaos/grader.ts b/tests/restart-chaos/grader.ts index 14815988041..1a3730f456e 100644 --- a/tests/restart-chaos/grader.ts +++ b/tests/restart-chaos/grader.ts @@ -1,4 +1,4 @@ -import { UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST, type FaultReceipt, type OracleResult, type RuleId, type RuleResult, type TaskGrade, type TaskRecord, type TaskSample } from "./types.js"; +import { UI_RETRY_BLOCKED_BY_MODAL, UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST, type FaultReceipt, type OracleResult, type RuleId, type RuleResult, type TaskGrade, type TaskRecord, type TaskSample } from "./types.js"; import { g1Decision } from "./workloads.js"; import { enforcedUniversalFailures, gradeUniversal, type UniversalDbEvidence } from "./universal.js"; @@ -112,7 +112,7 @@ function gradeOwnRetry(record: TaskRecord): RuleResult { } return { verdict: "n/a", detail: "no Retry needed" }; } - if (record.retry.via === "ui" && (record.retry.error === UI_RETRY_BUTTON_MISSING || record.retry.error === UI_RETRY_NO_REQUEST)) { + if (record.retry.via === "ui" && [UI_RETRY_BUTTON_MISSING, UI_RETRY_NO_REQUEST, UI_RETRY_BLOCKED_BY_MODAL].includes(record.retry.error ?? "")) { return { verdict: "fail", detail: `${record.retry.error}${record.retry.bannerText ? ` (user saw: ${record.retry.bannerText})` : ""}` }; } if (record.retry.error || record.retry.accepted !== true) { diff --git a/tests/restart-chaos/observer.ts b/tests/restart-chaos/observer.ts index 7ba82e01fda..8618de508c1 100644 --- a/tests/restart-chaos/observer.ts +++ b/tests/restart-chaos/observer.ts @@ -718,7 +718,8 @@ export class TaskObserver { if (this.board) { const click = await this.board.clickRetry(issueRef, issueId); Object.assign(attempt, { via: "ui", httpStatus: click.httpStatus, accepted: click.accepted, error: click.error, bannerText: click.bannerText, - ...(click.diagnostics !== undefined ? { diagnostics: click.diagnostics } : {}) }); + ...(click.diagnostics !== undefined ? { diagnostics: click.diagnostics } : {}), + ...(click.dismissedDialogs?.length ? { dismissedDialogs: click.dismissedDialogs } : {}) }); if (click.accepted) attempt.acceptedAt = Math.max(now, Date.now()); return attempt; } diff --git a/tests/restart-chaos/types.ts b/tests/restart-chaos/types.ts index 0dee982800e..451dc26240e 100644 --- a/tests/restart-chaos/types.ts +++ b/tests/restart-chaos/types.ts @@ -256,10 +256,12 @@ export interface RetryAttempt { via?: "ui" | "api"; bannerText?: string | null; diagnostics?: { screenshot: string | null; dom: string | null } | null; + dismissedDialogs?: string[]; } export const UI_RETRY_BUTTON_MISSING = "ui_retry_button_missing"; export const UI_RETRY_NO_REQUEST = "ui_retry_no_request"; +export const UI_RETRY_BLOCKED_BY_MODAL = "ui_retry_blocked_by_modal"; export interface FaultReceipt { faultId: string;