Check the validity of my object repository in case of an update

Hello everyone,

I am looking to automatically test my object repository for an intranet, which will be updated soon.

I would like to create a workflow that searches the repository for the described objects, then tests them, and of course reports any errors.

I have a .objects directory that contains 1 directory (my application), and inside this directory, I have as many directories as there are screens, and so on.

Do you have a solution to automate this?

have a good day

Arnaud

Hello,

Test workflows are your answer here, you can have them run automatically aswell if you have a Test licence and alert you in a selector should fail.

If you are wanting to just ‘plug’ the Object Repository into something and have it be validated, thats not how it works. You need to still build the workflows to navigate etc.

Building test workflows as you go is the ideal scenario.

@Arnaud_G

Workflow cannot scan tthrough object repo and validate each selector

you need to build your flows and run them…testing capabilities provide you with the ability to do so..Also if your object repo is names properly and hierarchy is proper..then you can leverage autopilot to build your flow. might not be 100% accurate but would get most of the flow in place

cheers

Hi @Arnaud_G

You can try to create a workflow that loops through the .objects directory, reads each screen folder, validates described objects using ui automation activities, logs results, and writes errors to a report file.

Happy Automation

How is it going to know how to navigate between each screen? This is not a good idea in my opinion.

A test workflow needs to be deliberately built. Thats what they are for, not trying to backwards engineer the object repository.

Thanks for all your answers, i’m looking if we’ve got the test licence or autopilot

it proposed validating whether the objects/selectors in the repository still exist, not building full navigation or end to-end tests.

What isn’t possible is automatically generating full test flows - and that’s where your point is completely correct.
Thanks