Test instability when launching a new process within a test.

[aaronpowell]

What happened?

As noted in a7ae343 we have a test that is not stable across OSes (it worked on Windows but not Linux).

There seems to be something problematic when it comes to how we launch a new process within a test and then capture the output of it to then test against.

We should review if this method we're using is the best way in which we can launch a process within a test, and if it's the best way to read the stdout and stderr streams.