Robot framework for distributed testing

Question

We are in the process of finalizing a test framework, and are pretty impressed with Robot framework and STAF.

Unable to decide with the optimal approach for the below:

1. Want to be able to start tests from a server by selecting clients
   • Can we display all clients in the existing network?
2. Number of clients can increase/decrease over time
3. When we select a client we want to go fetch client properties dynamically
   • Is there a way to display dynamic properties on RIDE/STAX
   • Can we use any other framework and integrate with Robot? Staf/Stax?
4. User should be able to choose tests the client supports and build a config
   • Can we use RIDE or something similar to build test configurations/per client
5. Launch all clients in parallel, monitor and report results
   • Is there a way to launch and monitor results in parallel?

Answer 1

This is only a partial answer, based on my understanding of your question and lacking full knowledge of your situation.

1. Robot Framework can be started from a .cmd or .bat file. I don't know if the results can be sent somewhere else to be saved from there, but I'm fairly certain that yes, they can.

   1.1. If you can get that list in Python or Java, then yes, you can do it in Robot Framework and pass variables and pass/fail results around the test suite. You might be able to use both at once, but I'm not sure yet.

2. Handled with the first item.

3. Robot + Python/Java can probably handle that.

   3.1. I don't know, I use PyCharm as my Robot Framework IDE. It has an integrated console and allows for quick and easy managing of Python/RobotFramework files, as well as a lot of other languages, but I'd imagine that using Robot's Log to Console keyword, you could send the results directly to the console. So, yes.

   3.2. Short answer: Not that I've ever heard of, but if you can run those with Java/Python and return the Pass/Fail results to Robot Framework, then yes.

4. Using multiple tags, the Robot Framework programmer can, at runtime, either run the test excluding particular tags or run the test including particular tags.

5. In theory: yes. Again, not something I've ever done, and honestly telling you how is beyond me, but I don't know a reason why you couldn't as long as you don't have any custom keywords that move the mouse cursor.