Flutter, Google's open-source UI toolkit, has gained immense popularity for building natively compiled applications for mobile, web, desktop, and embedded devices from a single codebase. With its versatility, developers have embraced Flutter for creating dynamic and responsive user interfaces. However, when it comes to test automation, Flutter apps require a unique approach due to their underlying architecture.
“It’s no use! I can’t run an end to end test with Flutter’s integration tests”, exclaimed one of our customers about 9 months ago. I asked what the problem was and they explained that they were using Google Authentication for logging in and used the google_sign_in package for and it wasn’t possible use Flutter’s integration tests to interact with the login screens.
In Flutter’s early days in 2019, I developed a live object detection system for a major German company, despite the platform’s constraints. With release of Flutter 3.7 and advancements of TensorFlow have catalyzed the need to refine or overhaul this approach. This article discusses the newest techniques in live-stream object detection as showcased in the flutter-tflite GitHub repository.
Mobile developers using Javascript-based mobile application development platforms such as Cordova, Ionic and React Native have enjoyed the benefit of being able to push app updates over-the-air without resubmitting their apps to the App Store or Google Play for quite some time. As long as the updates are not compiled code, and don’t change the primary purpose of the application then both Apple and Google allow this.