Never decrement next id in nova testservce

Never decrement next id in nova testservce

Adjusting the next id downwards is not necessary and can cause breakage
across testcases when cleanup does not exactly match creation, which is
common in juju.

For example:
* TestA creates two rules: next id is 3, rules is {1, 2}
* TestA deletes rule 1: next id is 2, rules is {2}
* TestB creates one rule: boom... only when TestA runs first

https://code.launchpad.net/~gz/goose/never_decrement_next_id/+merge/147208

(do not edit description out of merge proposal)

[gz, 2013-02-07 18:49:45]

Please take a look.

[jameinel, 2013-02-07 19:49:08]

LGTM. It isn't like we are worried about overflowing an int here. :)

I'm curious what the real behavior is, would it be hard to test creating and
releasing a bunch of Ids on canonistack and seeing how the numbers change (Maybe
not ideal, since there is noise in that system.)

Maybe audit the openstack code, but both seem like a lot of time spent for
something trivial.

[dimitern, 2013-02-07 20:19:47]

LGTM. I wasn't sure if I had to do that in the first place (the decrementing),
but it seemed cleaner then. It's a trivial change anyway and since it helps
tests, perfect!

[gz, 2013-02-08 17:15:08]

On 2013/02/07 19:49:08, jameinel wrote:
> 
> I'm curious what the real behavior is, would it be hard to test creating and
> releasing a bunch of Ids on canonistack and seeing how the numbers change
(Maybe
> not ideal, since there is noise in that system.)

Pretty sure it's just an auto-increment in mysql for the things that aren't
UUIDs still.