fix: CI/CD tests sometimes fail if deployment takes longer than default Task timeout

[eob]

This PR introduces a wait_forever method on Task, and then uses it for Package & Plugin deployments within tests.

The hope is that some of the occasional deployment timeout failures can be avoided by simply extending the wait time.

(In response to a PR earlier today in which a single failure related to deployment timeouts happened.)

[douglas-reid]

@eob do you think this is an indication that we should add something like either: task.wait_forever() or task.until_succeed_or_failed() ?

[eob]

@douglas-reid I sat on it for the day to see if I had an opinion, and I think I'm completely agnostic.

It's just a few LoC change -- LMK if you think that's a better route and I'll update this PR to be that instead.

If yes, my vote between the two would be wait_forever with the implicit semantics of throwing on failure or returning on success.

[douglas-reid]

@douglas-reid I sat on it for the day to see if I had an opinion, and I think I'm completely agnostic.

It's just a few LoC change -- LMK if you think that's a better route and I'll update this PR to be that instead.

If yes, my vote between the two would be wait_forever with the implicit semantics of throwing on failure or returning on success.

No worries. I've just run across this before personally, and always end up writing something like:

   state = task.state
   while state not in [TaskState.succeeded, TaskState.failed]:
        try:
            task.wait()
            state = task.state
        except:
           ...

or whatever and have thought about adding a shorthand for that if others started stumbling too.