July 2019 Summaries

Firebase announced an upgrade that connects existing Firebase projects to the next-generation Google Analytics experience without requiring code changes or disrupting existing analytics data and Firebase console workflows. The upgrade adds cross-device customer journey measurement through User-ID, closed funnels for analyzing conversion paths and drop-off points, and an Analysis module with tools such as segment overlap, geography-based comparisons, explorations, funnels, and path analysis. Firebase users retain free, unlimited event reporting and existing dashboards while gaining access to advanced Google Analytics reports, with some user counts potentially decreasing if cross-device User-ID deduplication is enabled. Users can upgrade through the Firebase Analytics dashboard, while developers migrating from the deprecated Google Analytics Services SDK must still independently move to the Google Analytics for Firebase SDK.

Firebase Test Lab now supports Cloud Functions triggers that run automatically when a test matrix completes, eliminating the need for developers to manually check results or repeatedly poll the Test Lab API. Test Lab runs automated app tests across hosted Android and iOS devices, while Cloud Functions provides a managed backend for responding to completion events and accessing test status, outcome summaries, and infrastructure errors. Developers can use these triggers to automate workflow actions such as emailing teams, posting Slack notifications, or creating JIRA issues. The feature is available in version 3.2.0 of the firebase-functions Node module, with documentation, API references, GitHub samples, and Firebase Slack support available for implementation.

Firebase’s Unity SDK performs I/O-heavy operations such as dependency checks and authentication through background Tasks, requiring developers to return to Unity’s main thread before accessing UnityEngine-managed resources like PlayerPrefs or GameObjects. The discussion demonstrates several approaches for safely handling these asynchronous calls: using TaskScheduler.FromCurrentSynchronizationContext with ContinueWith, Firebase’s simpler ContinueWithOnMainThread extension, and C# async/await syntax, which offers more readable sequential code and generally resumes on the calling thread. It also examines Unity coroutines through a CustomYieldInstruction that waits for Tasks, noting their main-thread execution and automatic termination when their MonoBehaviour is destroyed, as well as manually managed action queues for dispatching completed background work during Update. Finally, it introduces UniRx as a reactive alternative that can explicitly schedule work on the main thread and may be useful for event streams such as realtime database updates. Across all methods, developers must account for errors, lifecycle concerns involving destroyed Unity objects, and performance trade-offs between background work, per-frame polling, and main-thread processing.