> If I understand correctly, these do two different things: the former
> #cleanPackages actually ensured that all packages really were clean by
> forcing them to be dirty and then check their clean-ness. Sometimes
> I've seen packages that _looked_ clean, but only because something
> failed to notify the package that it had altered.
>
> #verifyPackagesClean doesn't seem to do that: it only checks packages
> that claim to be dirty. In theory that should be fine, except for the
> apparently-clean-only-not-really problem.

Okay, we should probably invoke Berts new function to make sure.
Calling #prepareNewBuild with dirty packages should be a user-error.
If they / it is trying to build with a dirty package I'd rather
understand why and fix that than potentially throwing away a quick fix
they forgot about.