Limitations on automated GUI testing

The testing library allows more manual GUI tests to be rewritten as automated tests. However, automated tests cannot replace manual testing entirely, and there exists toolkit-dependent and platform-dependent limitations as to what can be achieved for programmatically imitating user interactions, e.g. the animation of a button being depressed when clicked may not be replicated in automated tests.

The delay parameter allows playing back a test for visual confirmation, but it may not look identical to how the GUI looks when it is tested manually.

If the GUI / trait states being asserted in tests are not consistent compared to manually testing, then that is likely a bug. Please report it.

If the GUI / trait states being asserted in tests are consistent with manual testing, then such visual discrepancies may have to be tolerated.