Re : How to handle migrations with multiple RDBMS?

Hi again,

Is it possible to replace hash computation in some way through extension mechanism ? I would think it is possible to override changelog serializers hence we could imagine generating a string that would be uniform across databases thus generating consistent hash but there might be unknown side-effects.

If this is not possible, would you accept a patch that would compute hashing on the raw text of changesets ?

Thanks

Arnaud