Improve RepoErrors and error output

[daviddavis]

Some suggestions:

• We have this RepoErrors model in the code. I like the idea of the model but we pass it to functions as a parameter and they add to it. I imagine we could instead return it from the functions and create an append method that would append errors to a RepoErrors instance.
• The RepoErrors model captures all sorts of results including non-error results so it's a bit of misnomer.
• It's strange that a dist always has to be specified when adding errors to RepoErrors
• When running the CLI and not using file output, a bunch of information just gets spewed out. A lot of the information is unnecessary such as successes