More tests for parsers

[erikh]

A recurring theme is that the parser tests are anemic. While testing the UI and interaction with google calendar in particular is a bugbear that will take a lot more work (and separate tickets, almost surely), I think tackling better tests for the parsers will result in better release quality as well as help me ensure I'm not spinning my wheels cutting releases that don't actually fix bugs.

#72 is a good example of where tests could have kept me from not releasing a "bugfix" that really wasn't.