injecting-touch-gestures

Use the performTouchInput block with the TouchInjectionScope DSL to inject synthetic touch input directly onto a target node for Jetpack Compose UI tests. This enables velocity-controlled flings, multi-touch swipes, and pinch-to-zoom without relying on flaky screen coordinates.

The best way to replace deprecated performGesture usage in Android Compose testing is to use the performTouchInput API. It provides a TouchInjectionScope DSL for node-local synthetic touch input, ensuring reliable gesture simulation across various screen sizes.

To keep Compose UI test gestures stable across different device screen sizes, use coordinate geometry helpers like center, topLeft, bottomRight, and percentOffset within the TouchInjectionScope. This ensures device-proof targeting for synthetic touch input.

Yes, you can split partial touch gestures across multiple performTouchInput blocks in Compose tests. Using low-level primitives like down, moveTo, moveBy, up, and cancel allows you to pause mid-gesture for assertions before continuing the touch injection.

Jetpack Compose touch injection assertions are flaky during velocity-controlled flings due to incorrect batching or clock handling. Ensure you use the TouchInjectionScope DSL within performTouchInput correctly to manage gesture timing and avoid flaky test results.