[Cosmo-dev] testing data migration for cosmo 0.6

[Brian Moseley]

On 12/4/06, Mikeal Rogers <mikeal at osafoundation.org> wrote:

> Why not? We're inputing the functional test data into 0.5, migrating
> it, then testing it manually after the migration -- isn't that the
> definition of "testing specific data that was in the database at the
> time of the migration"

when you run the functional tests against the pre-migration database,
the tests insert a bunch of data. if you were to migrate the database
and then test the data those tests inserted, that would be
interesting. but you proposed to remove those users and run the
functional tests again, creating new users, which i don't see being at
all useful in determining whether or not the actual migration of real
existing data works.