kotlin-concurrency-expert

To fix GlobalScope usage in Android Kotlin coroutines, replace it with lifecycle-aware scopes like viewModelScope or lifecycleScope. This enforces structured concurrency and prevents memory leaks by automatically canceling jobs when their scopes are destroyed.

Android coroutine code swallows CancellationException when exceptions are caught and handled improperly. You must rethrow CancellationException within your error handling blocks to ensure structured concurrency and allow parent jobs to cancel child jobs correctly.

The best way to collect StateFlow safely in Android Fragments is using repeatOnLifecycle with lifecycleScope. This approach ties flow collection to specific lifecycle states, preventing unnecessary background processing and avoiding app not responding errors.

To inject CoroutineDispatcher for testability in Android, pass dispatchers through ViewModel constructors instead of hardcoding them. This allows you to replace real dispatchers with test dispatchers during unit testing, ensuring deterministic and controlled coroutine execution.

You should use withContext for IO-bound work in Kotlin when performing disk or network operations. Switching to Dispatchers.IO inside a viewModelScope prevents blocking the main thread, avoiding UI freezes while maintaining structured concurrency boundaries.

Yes, you can use callbackFlow to encapsulate asynchronous APIs in Android. It converts callback-based APIs into cold flows, allowing proper structured concurrency management, cancellation handling, and state exposure within lifecycle-aware components like ViewModel.