Parallelize Playwright accessibility tests and isolate functional suite

src/app/.gitignore

@@ -39,4 +39,7 @@ pnpm-lock.yaml
 
 # Temporary files
 *.tmp
-.temp/
+.temp/
+
+# Playwright test results
+test-results/

src/app/README.md

@@ -54,13 +54,16 @@ npm test
 
 Playwright starts an isolated, non-HMR test server on **127.0.0.1:4173** by default. This intentionally avoids reusing a manual development server on port 8080. Set `PLAYWRIGHT_BASE_URL` only when you explicitly want to test an already running server.
 
+> [!NOTE]
+> Direct execution via `npx playwright test` runs the accessibility (`a11y`) and `functional` test suites concurrently, which can cause state collisions in local storage/server state. For isolated, sequential execution of the suites, always use `npm test` or `npm run test:headed`.
+
 Runs all UI-safe Playwright tests headless via Chromium. Real-server smoke tests are opt-in so they fail fast instead of silently passing without a running server or loaded model:
 
 ```bash
 LEMONADE_REAL_SERVER=1 npm test
 ```
 
-Artifacts are saved under Playwright's per-test output folders, so screenshots from repeated runs are not overwritten.
+Artifacts are saved under distinct output folders per project (`test-results/a11y` and `test-results/functional`), preventing results from being overwritten when running the sequential test suites.
 
 ### Headed tests (visible browser)
 

src/app/package.json

@@ -15,8 +15,8 @@
     "build:tauri": "tauri build",
     "build:nobundle": "tauri build --no-bundle",
     "build:mac": "tauri build --bundles app",
-    "test": "npx playwright test",
-    "test:headed": "npx playwright test --headed",
+    "test": "node scripts/run-playwright-tests.mjs",
+    "test:headed": "node scripts/run-playwright-tests.mjs --headed",
     "test:a11y": "npx playwright test tests/a11y.spec.ts",
     "test:preset-intent": "node tests/preset-intent.runtime.cjs",
     "test:autoopt-glue": "node tests/autoopt-glue.runtime.cjs",

src/app/playwright.config.ts

@@ -42,7 +42,22 @@ export default defineConfig({
   },
   projects: [
     {
-      name: 'chromium',
+      name: 'a11y',
+      testMatch: '**/a11y.spec.ts',
+      fullyParallel: true,
+      outputDir: './test-results/a11y',
+      use: {
+        browserName: 'chromium',
+        viewport: { width: 1280, height: 800 },
+        ...(launchOptions ? { launchOptions } : {}),
+      },
+    },
+    {
+      name: 'functional',
+      testIgnore: '**/a11y.spec.ts',
+      fullyParallel: false,
+      workers: 1,
+      outputDir: './test-results/functional',
       use: {
         browserName: 'chromium',
         viewport: { width: 1280, height: 800 },

src/app/scripts/run-playwright-tests.mjs

@@ -0,0 +1,52 @@
+#!/usr/bin/env node
+
+import { spawnSync } from 'node:child_process';
+import process from 'node:process';
+
+const args = process.argv.slice(2);
+
+// If user explicitly specifies a project, we forward all args directly to a single Playwright run
+const hasProject = args.some(arg => arg.startsWith('--project') || arg === '-p');
+const hasHelp = args.some(arg => arg === '--help' || arg === '-h' || arg === 'help');
+
+const npxCommand = process.platform === 'win32' ? 'npx.cmd' : 'npx';
+
+if (hasProject || hasHelp) {
+  const result = spawnSync(npxCommand, ['playwright', 'test', ...args], {
+    stdio: 'inherit',
+    shell: false,
+  });
+  if (result.error) {
+    console.error('Failed to start Playwright tests:', result.error.message || result.error);
+    process.exit(1);
+  }
+  process.exit(result.status ?? 1);
+}
+
+console.log('Running accessibility (a11y) tests...');
+const a11yResult = spawnSync(npxCommand, ['playwright', 'test', '--project=a11y', ...args], {
+  stdio: 'inherit',
+  shell: false,
+});
+
+if (a11yResult.error || a11yResult.status !== 0) {
+  if (a11yResult.error) {
+    console.error('Failed to start accessibility tests:', a11yResult.error.message || a11yResult.error);
+  }
+  process.exit(a11yResult.status ?? 1);
+}
+
+console.log('Running functional tests...');
+const functionalResult = spawnSync(npxCommand, ['playwright', 'test', '--project=functional', ...args], {
+  stdio: 'inherit',
+  shell: false,
+});
+
+if (functionalResult.error || functionalResult.status !== 0) {
+  if (functionalResult.error) {
+    console.error('Failed to start functional tests:', functionalResult.error.message || functionalResult.error);
+  }
+  process.exit(functionalResult.status ?? 1);
+}
+
+process.exit(0);

src/app/tests/a11y.spec.ts

@@ -23,6 +23,8 @@
 import { test, expect, Page } from '@playwright/test';
 import AxeBuilder from '@axe-core/playwright';
 
+test.describe.configure({ mode: 'parallel' });
+
 // ─── Constants & helpers ──────────────────────────────────────────────────────
 
 const WCAG_TAGS = ['wcag2a', 'wcag2aa', 'wcag21a', 'wcag21aa'] as const;

src/app/tests/download-manager-order.spec.ts

@@ -54,10 +54,10 @@ test('download manager keeps creation order while concurrent progress updates ar
     },
   ];
 
-  await page.route('/api/v1/health', route =>
+  await page.route('**/api/v1/health**', route =>
     route.fulfill({ json: { status: 'ok', all_models_loaded: [] } }),
   );
-  await page.route('/api/v1/downloads**', route => route.fulfill({ json: { downloads: serverDownloads } }));
+  await page.route('**/api/v1/downloads**', route => route.fulfill({ json: { downloads: serverDownloads } }));
 
   await page.goto('/');
   await page.locator('.titlebar__download-toggle').click();
@@ -154,10 +154,10 @@ test('authoritative server snapshot removes stale paused renderer state', async
     updatedAt: timestamp,
   });
 
-  await page.route('/api/v1/health', route =>
+  await page.route('**/api/v1/health**', route =>
     route.fulfill({ json: { status: 'ok', all_models_loaded: [] } }),
   );
-  await page.route('/api/v1/downloads**', route => route.fulfill({ json: { downloads: [] } }));
+  await page.route('**/api/v1/downloads**', route => route.fulfill({ json: { downloads: [] } }));
 
   await page.goto('/');
   await page.locator('.titlebar__download-toggle').click();