metadata doesn't serialize naming convention

[Mike Bayer]

from sqlalchemy import Table, Column, MetaData, Integer
import pickle

m = MetaData(naming_convention={"pk": "%(table_name)s_pk"})
t = Table('t', m, Column('id', Integer, primary_key=True))

m2 = pickle.loads(pickle.dumps(m))

t2a = Table('t2', m, Column('id', Integer, primary_key=True))
t2b = Table('t2', m2, Column('id', Integer, primary_key=True))

assert t2a.primary_key.name == t2b.primary_key.name

[Mike Bayer]

• Fixed issue where a :class:.MetaData object that used a naming convention would not properly work with pickle. The attribute was skipped leading to inconsistencies and failures if the unpickled :class:.MetaData object were used to base additional tables from. fixes #3362

→ <<cset 55c26710a1f6>>

[Mike Bayer]

• Fixed issue where a :class:.MetaData object that used a naming convention would not properly work with pickle. The attribute was skipped leading to inconsistencies and failures if the unpickled :class:.MetaData object were used to base additional tables from. fixes #3362

(cherry picked from commit 55c26710a1f6d6c6f87a9752035bc160f93b38d0)

→ <<cset e7cf56828486>>