On Thu, Feb 11, 2010 at 10:52 AM, John Admanski <[email protected]> wrote:
> I think we could accept as a general rule that you can change a database
> migration (rather than adding a new one) so long as the only actual change
> in behavior is that the migration now works for people who it would've
> failed for before (i.e. for people who could successfully run the old
> version of the migration, the new version does not change the behavior in
> any way).

Right, as long as it failed before (and will rerun), sounds fine.
_______________________________________________
Autotest mailing list
[email protected]
http://test.kernel.org/cgi-bin/mailman/listinfo/autotest

Reply via email to