Github user neykov commented on the pull request:
https://github.com/apache/incubator-brooklyn/pull/994#issuecomment-154085081
Moving them to integration will address the immediate problem so +1.
At some point we will start running the integration tests on apache
infrastructure (read - **busy**, we already do, but they are failing for a
number of other reasons), so the same issue will crop up - but can deal with it
then.
---
The code could be restructured such that it doesn't use waits at all:
* tests which need non-running server be put in a separate test, which
doesn't start the server by default
* no sleep on server start, if the server doesn't come up immediately,
then always use `eventuallySucceeds`
* for the cases where there's `xxxxAfter(5); succeedsEventually`, could
be changed to `failsNow; task = succeedsEventuallyInAnotherTask; startNow();
task.get()`. What could be a problem here is that connections to closed ports
take a long time to timeout?
* The only place where a fixed wait is really needed is tests which
expect to fail and have a timeout, those go in integration.
All of the above said, I am perfectly fine with moving all tests to
integration as is.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
