Ticket #4513 -- wrapping db:fixtures:load in a transaction

Any chance of having this patch applied?

Shouldn't break anything, I believe, and is pretty useful when your
test db has FK constraints.

Regards,
Isak

But it will break stuff. Transaction in this case should be optional, not implicit.

>
> Shouldn't break anything, I believe

But it will break stuff.

Yup, you're right. I'm sure someone out there depend on the data that
gets loaded before the task hits an error and aborts.

There's probably other ways for stuff to break as well, or?

Transaction in this case should be optional, not

implicit.

That doesn't sound like a sensible default to me, but I guess altering
behaviour is bad no matter what.

Isak