Let the GMail scraper save its state with pickle

[jchillerup]

The scraper should do some kind of pagination when fetching all the mail. This pagination should be kept track of so that the scraping won't have to start over if the process crashes/gets throttled/whatever. I suggest we do that with pickle, but whatever floats the implementer's boat ;)