change how alembic runs an ADD COLUMN on postgresql

gistfile1.txt

# in env.py....

from alembic.ddl.postgresql import PostgresqlImpl
class MyPostgresqlImpl(PostgresqlImpl):
    __dialect__ = 'postgresql'

    def add_column(self, table_name, column, schema=None):
        if column.server_default is not None:
            server_default = column.server_default
            column.server_default = None
            nullable = column.nullable
            column.nullable = True
            super(MyPostgresqlImpl, self).add_column(table_name, column, schema=schema)
            self.alter_column(table_name, column.name,
                        server_default=server_default, schema=schema)
            if not nullable:
                self.alter_column(table_name, column.name, nullable=False, schema=schema)

        else:
            super(MyPostgresqlImpl, self).add_column(table_name, column, schema=schema)


# profit!

# output of:

def upgrade():
    op.add_column("foo", sa.Column('data', sa.String(50), nullable=False, server_default='value'))

INFO  [alembic.migration] Running upgrade 20cab39367e3 -> 500e690e37d0, add col w default
INFO  [sqlalchemy.engine.base.Engine] ALTER TABLE foo ADD COLUMN data VARCHAR(50)
INFO  [sqlalchemy.engine.base.Engine] {}
INFO  [sqlalchemy.engine.base.Engine] ALTER TABLE foo ALTER COLUMN data SET DEFAULT 'value'
INFO  [sqlalchemy.engine.base.Engine] {}
INFO  [sqlalchemy.engine.base.Engine] ALTER TABLE foo ALTER COLUMN data SET NOT NULL
INFO  [sqlalchemy.engine.base.Engine] {}
INFO  [sqlalchemy.engine.base.Engine] UPDATE alembic_version SET version_num='500e690e37d0'
INFO  [sqlalchemy.engine.base.Engine] {}
INFO  [sqlalchemy.engine.base.Engine] COMMIT

Is it just me, or that the alter_column of lines 14-15 will not actually affect current rows?
If you would like to set not nullable at line 17, you should actually set all the rows to the requested default value or the operation fails.