Allow container dsl to run outside of arquillian

[aahmed-se]

I am trying to run the container dsl outside the context to arquillian runner , trying to integrate integrate with a custom test framework

Container pingpong = Container.withContainerName("pingpong")
            .fromImage("jonmorehouse/ping-pong")
            .withPortBinding(8080)
            .build();

As if understand the manager can start as so

Manager manager = ManagerBuilder.from().extension(LoadableExtensionLoader.class).create();
        manager.start();

what would be the other things that I need to initialize for the container dsl to work ?

[aahmed-se]

I tried doing adding the following

manager.inject(pingpong);

It seems to have injected some fields correctly but it didn't work as expected

[dipak-pawar]

@aahmed-se Can you share your project to understand how are you invoking arquillian-cube with another framework? so maybe I can help you

[aahmed-se]

it's an in house test framework similar to testng, I need to add custom setup and teardown methods for the containers so need more fine grain control of the process.

[lordofthejars]

@aahmed-se I would do it differently. I am not an expert of TestNg but maybe you can get the idea from https://github.com/arquillian/arquillian-cube/tree/master/docker/junit-rule/src/main/java/org/arquillian/cube/docker/junit/rule

[dipak-pawar]

If you want to use Arquillian Runner then there is no need to fire event, any configuration in arquillian.xml with required dependency in pom.xml should work. Look here for example

If you don't want to use arquillian runner you have to fire some events by your own to work cube as per expectation

• Initialize docker client
• Fire required event - Have a look into this method in arquillian cube to get an idea.

[lordofthejars]

In this concrete example and if you are only planning to integrate with this part, then I suggest using the Rule approach in TestNg. If you plan to use more features of Arquillian then what @dipak-pawar suggests is the best way to proceed.

[aahmed-se]

seems this has been discussed before, is this approach wrong ?

#242