On Thu, 2009-07-30 at 06:23 +0000, James Henstridge wrote:
> Review: Approve
> > That sounds good, but such test would not fail if I reverted my changes,
> > so I thought of moving my original test to PostgresDisconnectionTest and
> > writing a new one as you suggest on DatabaseDisconnectionTest.
>
> It seems that that the test doesn't fail on MySQL due to MySQL issuing a DatabaseError subclass in this situation. So that indicates that this particular bug was only affecting the PostgreSQL backend.
>
> That doesn't make the test useless though: it makes sure we don't reintroduce the bug, or have it occur in future backends.
>
Absolutely; I was just arguing for a test that was specific for the
InterfaceError exception.
>
> > I've pushed these changes up. Please let me know what you think
>
> They look good. I've merged them as r319.
On Thu, 2009-07-30 at 06:23 +0000, James Henstridge wrote: ectionTest and ectionTest.
> Review: Approve
> > That sounds good, but such test would not fail if I reverted my changes,
> > so I thought of moving my original test to PostgresDisconn
> > writing a new one as you suggest on DatabaseDisconn
>
> It seems that that the test doesn't fail on MySQL due to MySQL issuing a DatabaseError subclass in this situation. So that indicates that this particular bug was only affecting the PostgreSQL backend.
>
> That doesn't make the test useless though: it makes sure we don't reintroduce the bug, or have it occur in future backends.
>
Absolutely; I was just arguing for a test that was specific for the
InterfaceError exception.
>
> > I've pushed these changes up. Please let me know what you think
>
> They look good. I've merged them as r319.
Thanks!
--
Guilherme Salgado <email address hidden>