Restore launcher readiness check after boot; require dual logcat markers for launch validation

[Copilot]

macOS CI (no KVM, --gpu guest) fails at launch validation because removing the launcher focus check in PR #81 left boot_completed=1 + CPU threshold as the only readiness signal — insufficient on slow emulators. The system reports "ready" while still initializing, so the test app never fully starts.

Evidence from run 22268361795: boot 394s ✅, CPU 7.24 (threshold 30) ✅, install ✅, monkey launch ✅, but E2E_APP_STARTED_SUCCESSFULLY never appeared in logcat across 25 attempts / 10 min. Same commit on Ubuntu (KVM): marker found on attempt 2.

Emulator.cs — Restore launcher readiness check

After boot_completed=1, poll dumpsys window displays for launcher focus with:

• 120s hard timeout independent of --timeout — never blocks indefinitely (the hang that motivated PR Fix macOS emulator boot hang: remove launcher focus check, unify --timeout budget #81's removal)
• ANR dialog dismissal via KEYCODE_BACK — handles the macOS-specific ANR focus-steal that caused the original hang
• Non-blocking — proceeds after timeout since boot IS complete; the extra wait just gives the system time to stabilize

// Phase 3: best-effort launcher wait (won't fail, won't hang)
while (launcherSw.Elapsed < launcherTimeout && !token.IsCancellationRequested)
{
    var output = adb.Shell("dumpsys window displays", Serial);
    if (output.Any(l => l.Contains("mCurrentFocus", ...) && l.Contains("launcher", ...)))
        break;
    // Dismiss ANR dialogs blocking launcher focus
    if (output.Any(l => l.Contains("Application Not Responding", ...)))
        adb.Shell("input keyevent KEYCODE_BACK", Serial);
}

run.yml — Dual-marker launch validation

Require both markers, log each individually to reveal ordering:

• E2E_APP_STARTED_SUCCESSFULLY — app-emitted marker
• Displayed $TEST_APP_PACKAGE — system ActivityManager marker

This tells us definitively whether the activity displayed (system-level) vs whether the app's managed code ran to completion (app-level).

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[github-actions]

Test Results

  4 files    4 suites   22m 21s ⏱️
103 tests 100 ✅ 3 💤 0 ❌
206 runs  200 ✅ 6 💤 0 ❌

Results for commit 42c0919.

♻️ This comment has been updated with latest results.