Merge lp://staging/~renamer-developers/renamer/ditch-metalanguage into lp://staging/renamer

Everything looks good, the only suggested change was to move the filename parsing out of the TVRage class into a base TV class so that it could be shared by any other metadata sources (e.g. thetvdb.com) since it isn't specific to TVRage. That can wait until there is actually another source implemented.

> 24 from renamer._version import version
> 25 +version # Ssssh, Pyflakes.

Don't do this, use __all__ instead.

> 124 + ('dry-run', 'n', 'Perform a dry-run.'),

It might be more intuitive to make the long option name be --no-act.

> 6 +Copyright © 2007-2010 Slipgate Development cc

There isn't really such an entity as Slipgate Development cc.

> 234 + self.args = (FilePath(arg) for arg in args)

This should probably be a listcomp, not a genexp.

> 1033 +class _metaASC(type):

"ASC" stands for AxiomaticSubCommand, and thus isn't a very appropriate name here.

148 + return (get,)
149 +
150 + subCommands = property(*subCommands())

This is a bit pointless, just use @property on the inner function and be done.

Address review commentary.

2397 + if not (newcls.__name__ == typeName and
2398 + newcls.__module__ == moduleName):
2399 + directlyProvides(newcls, *provides)
2400 + return newcls

Okay, so. This should be alsoProvides() instead of directlyProvides(). You can also ditch the if check; instead, just throw in a noLongerProvides() after the definition of EridanusCommand to remove those interfaces from it. Finally, I think instead of this being a factory function that takes some params, it should just be a subclass of type that reads the list of interfaces off an attribute; then you just subclass it and define the interfaces as a class attribute.

2341 + if e.errno == errno.EXDEV:
2342 + raise errors.DifferentLogicalDevices(
2343 + 'Refusing to symlink "%s" to "%s" on another filesystem' % (
2344 + src.path, dst.path))

I don't understand this code at all; when would it be valid for symlink to fail with EXDEV?

1682 + key, value = line.strip().split(u'@', 1)
1683 + data[key] = value.split(u'^')

The page data is returned as a byte string, so splitting on unicode strings here is wrong.

Whew. I hope that's all!

> Whew. I hope that's all!

Fixed.