Automate UI Testing
I'm trying to create an UI Test project for an desktop application.
I saw this topic that it can perform 'Click' action for all the buttons in a dialog and it also possible to do others action for the different wndclass.
So, for example: can I click all the button in a dialog and then add region checkpoint after each of the click action by using script?
If there are more than thousands of dialog in that application, I have to create the same amount of checkpoints for all the dialog.
It's something like I would like to create the test case automatically. Is it possible to do this by using TestComplete? From my understanding on Region Checkpoint, it seems like not possible to create it automatically but have to make sure that is there any possibility to achieve this goal 😅
the region checkpoint will rely on referencing a previously stored test item within the Stores-Regions. Whenever testcomplete encounters the region checkpoint, it will look at the object in question, compare it to the baseline image stored within the stores-region test item, and deteremine whether or not the checkpoint passes based on the pixel&color tolerance you defined, along with any other masks that you've created.
Now assuming that you have a thousand dialogues, which produces a thousand unique objects that you'd like validate, that would mean that you require a thousand unique baseline images to be stored in the Stores test item, and that just does not seem feasible to me.
Something else to consider is the premise with which you intend on using checkpoints. In many cases, checkpoints = assertions, where you want to put it in a test script location to provide validity to whatever it is you are testing. In many cases, you can skip over a step, knowing that if the GUI responded the correct way, your test would continue to move on (or produce an error at a step that could not be performed)
Furthermore, a region checkpoint may not be needed. Within the various dialogues of the application, there usually are other object properties that you can validate to fit your needs, without having to rely on regions (which will also fail based on the dpi and resolution settings of the machine that the test is executed on).
I really haven't answered your question so far, so in short "yes, something along those lines can be possible, but i think there may be a better solution that using region checkpoints". food for thought.