ecsv files fail to write with units in metadata

This code fails because of the units in the metadata:

from astropy import table, units as u
t = table.Table()
t.meta = {"foo": 5*u.km, "foo2": u.s}
t["bar"] = [7]*u.km
t.write("test.dat", format="ascii.ecsv")

Note that either a quantity or a unit will cause the failure.

This is the end of the error messages I’m getting:

~/anaconda3/envs/py2/lib/python2.7/site-packages/yaml/representer.pyc in represent_object(self, data)
    408             reduce = copy_reg.dispatch_table[cls](data)
    409         elif hasattr(data, '__reduce_ex__'):
--> 410             reduce = data.__reduce_ex__(2)
    411         elif hasattr(data, '__reduce__'):
    412             reduce = data.__reduce__()

~/anaconda3/envs/py2/lib/python2.7/copy_reg.pyc in _reduce_ex(self, proto)
     68     else:
     69         if base is self.__class__:
---> 70             raise TypeError, "can't pickle %s objects" % base.__name__
     71         state = base(self)
     72     args = (self.__class__, base, state)

TypeError: can't pickle int objects

[taldcroft]

OK, I snuck away long enough to make #5605.

[taldcroft]

I likely won’t have time today (niece’s graduation, full day with family), but it would be great if someone could drive this home for 1.3 as a bug fix (otherwise 1.3.1). Basically just need a test writing and reading back a single ECSV table that has a unit, quantity, skycoord, and time object in the meta. I believe this is sufficient due to solid coverage for the base machinery of ECSV and YAML serialization.