Should it be an error if RestartRead is not present in omega.yml?

[xylar]

I got an error when RestartRead was not present in omega.yml. For most Polaris tests, we don't anticipate needing a restart. It's not a problem to include a RestartRead entry in the IOStreams but I wanted to make sure that the requirement that it be there is intentional.

[philipwjones]

Didn't think about the case where the entry isn't needed and is missing so I guess I need to catch that while still making sure the state is initialized. Are you initializing with the InitialState in these cases or will there be cases where we will initialize a test case with internal analytical data without reading one of those two streams?

[xylar]

We always initialize with InitialState whenever we wouldn't be using RestartRead. I don't think we plan to initialize from an analytic initial condition in Omega itself.