state,worker/provisioner: fix test timeouts

state,worker/provisioner: fix test timeouts

A couple of places that were relying on testing.ShortWait
to wait for something to happen.
There are a few more, but this should help matters
for the time being.

https://code.launchpad.net/~rogpeppe/juju-core/343-fix-test-timeouts/+merge/176155

(do not edit description out of merge proposal)

[rog, 2013-07-22 10:45:03]

Please take a look.

[dimitern, 2013-07-22 10:46:08]

LGTM, thank you!

[gz, 2013-07-22 10:53:39]

LGTM

https://codereview.appspot.com/11525045/diff/3001/state/testing/watcher.go
File state/testing/watcher.go (right):

https://codereview.appspot.com/11525045/diff/3001/state/testing/watcher.go#ne...
state/testing/watcher.go:170: for {
Okay, to check I understand this, previously the code would stop at:
1) timeout
2) when some changes arrived

If a list of more than one change was expected, it could recieve one, but not
the remainder, and the test would fail.

So, now it waits for:
1) timeout
2) as many or more changes as were expected

https://codereview.appspot.com/11525045/diff/3001/worker/provisioner/provisio...
File worker/provisioner/provisioner_test.go (right):

https://codereview.appspot.com/11525045/diff/3001/worker/provisioner/provisio...
worker/provisioner/provisioner_test.go:343: for time.Since(t0) <
coretesting.LongWait {
Feels like this should be factorable into a common helper.

[rog, 2013-07-22 11:00:48]

https://codereview.appspot.com/11525045/diff/3001/state/testing/watcher.go
File state/testing/watcher.go (right):

https://codereview.appspot.com/11525045/diff/3001/state/testing/watcher.go#ne...
state/testing/watcher.go:170: for {
On 2013/07/22 10:53:39, gz wrote:
> Okay, to check I understand this, previously the code would stop at:
> 1) timeout
> 2) when some changes arrived
> 
> If a list of more than one change was expected, it could recieve one, but not
> the remainder, and the test would fail.
> 
> So, now it waits for:
> 1) timeout
> 2) as many or more changes as were expected

that's right.

i'm not actually *quite* sure why the duration of ShortWait influenced the
number of changes delivered ( should probably investigate further to understand
that better), but this change seems right regardless - changes can always be
delivered piecemeal.

https://codereview.appspot.com/11525045/diff/3001/worker/provisioner/provisio...
File worker/provisioner/provisioner_test.go (right):

https://codereview.appspot.com/11525045/diff/3001/worker/provisioner/provisio...
worker/provisioner/provisioner_test.go:343: for time.Since(t0) <
coretesting.LongWait {
On 2013/07/22 10:53:39, gz wrote:
> Feels like this should be factorable into a common helper.

there is already a common helper (AttemptStrategy). it's
not so convenient to use here though - i intend to reimport
gustavo's modifications to it from ec2 at some point,
and we can revisit.