Runnable parameters

[marioaag]

What is the right syntax to pass arguments to an avocado-instrumented test through a Runnable?

Runnable('avocado-instrumented', uri, ???)

[beraldoleal]

Hi @marioaag , you could pass a kind, uri, config, args and kwargs. args and kwargs are used by specific runners, And you will find the docstring for each runner on avocado/core/nrunner.py. For instance, ExecRunner uses args as arguments to be given on the command line to the binary of the uri, and kwargs as enviroment variables.

If you are interested on InstrumentedTest (avocado/core/runners/avocado_instrumented.py), args and kwargs (unless you have tags) are not used. Something like this would be enough:

Runnable('avocado-instrumented', '/path/to/your/instrumented_test.py', config=your_config_dict)

Sorry if this is not so well documented, but we are not exporting this, publicly yet, and it is used only in our core. So if you have any specific requirement, an use case, and would like to share in more details with us, we can evaluate how incorporate this on our Job API (that is meant to be more public).

[beraldoleal]

Are you using nrunner or the default runner?

I did a test with the default runner against our examples/tests/sleeptest.py and it worked as expected:

avocado run tests/sleeptest.py -p sleep_length=10

However, using the next runner (--test-runner=nrunner), it seems that we have a problem, indeed. I just created #4569 so we can track this.

[richtja]

Hi @marioaag If you are trying to run multiple variants of one test, you can use CIT or Yaml_to_mux plugins? It might be useful for you.

[marioaag]

@beraldoleal, yes I'm using nrunner, one of my principal goals is have a solution that let me run multiple test in parallel.

[marioaag]

@richtja, the Yaml_to_mux as I know is not available for nrunner or did something change?

[richtja]

Yes you are right, if you are using the nrunner, the varianter plugins are not available right now.