Avoid Thread.sleep in test, use Awaitility FTW

[oscerd]

[oscerd]

It's not really straighforward to do. We need probably to rethink the tests. I'll leave this open.

[orpiske]

It's not really straighforward to do. We need probably to rethink the tests. I'll leave this open.

Is there any particular test you think it's problematic?

[oscerd]

The CamelSourceTaskTest in particular, wherever thread.sleep is used.

[orpiske]

The CamelSourceTaskTest in particular, wherever thread.sleep is used.

+1.

There's a few on CamelSourceAWSKinesisITCase too. Although they are using an approach similar to the one suggest by amazon, I still think they could be improved.

[fvaleri]

There is a proposed refactoring of CamelSourceTaskTest in #201.

I'm not using Awaitility, but a shared method called getTaskWithFixedNumberOfRecords. All properties from CamelSourceConnectorConfig used inside it are tested in isolated methods, so I think it's fine to use them when testing source task behaviors that do not involve them.