Testing for Kubernetes installed with Docker for Mac:

- Run `make build-coexecutor-container`
- Install and start rabbitmq on the localhost with Homebrew.
- PULSAR_TEST_INFRASTRUCTURE_HOST=docker.for.mac.localhost  PULSAR_TEST_FILE_SERVER_HOST=0.0.0.0 PULSAR_RABBIT_MQ_CONNECTION='amqp://guest:guest@localhost:5672//' nosetests test/integration_test.py:ExternalQueueIntegrationTests.test_integration_kubernetes

This manager is designed to just handle one job. It should be able to stage a job, write a command line to the job directory, and wait for an exit code to appear. Meant to be used alongside a container that actually runs the resulting tool.

See https://docs.python.org/2/library/unittest.html#deprecated-aliases

plus tempdir naming for ease of deciphering tracebacks.

unless explicitly requested.

So it turns out nose.tools.timed doesn't actually kill functions on timeout. Create a new decorator that actually does using the stopit library. Add decorator to another function that appears to hang under Python 3.4 sometimes.

Finally, under Python 3.4 on Traivs print the functions running so we can see hung entries.

 - Attempt to set maximum time for tests to prevent travisci hanging.
 - In various places loosen up skip conditions, try to never run a test that will fail because of the environment.

 - Depend on galaxy-lib.
 - Move things into pulsar.util and/or pulsar.client.util as makes sense.
 - Use more Galaxy idioms.

Fixes #62.

(In case I need it in the future again.)

Go further than runtime and ensure threads join so that they cannot interfer with other tests.

 - Rearrange things so ssh_key can be alternatively defined in file_actions.yml or the other properties ssh_user, ssh_host, ssh_port can be defined in destination_params. This allows these actions to be usable as the default file action and made it much easier to test.
 - Fixes when constructing scp command-line (my version of Python didn't like a tuple in middle of subprocess.check_call).
 - Increase integration test timeout (these are slower than localhost RESTful transfer).
 - Add integration test for each of these commands.

 - Imports must come before everything else now.
 - Cannot assign lambda to vars - boo!

The LWR use to not make a distinction between cancelled jobs and complete jobs but with server side postprocessing it is more nessecary. Bugfixes including actually recording and use cancelled status in external job runners and persisting the the recorded cancel status in queued and unqueued Python job managers (in some conditions it would revert the cancelled and mark the job as complete).

Catch and suppress exceptions related to cancelling jobs and fetching that status. Failure to ping the metadata store (disk) and determine if a job has been cancelled shouldn't be interpreted as a failure to cancel.

Lots of refactoring and tests to support this.

Will let travis ci see if this works.

Poster is not the best candidate because it is kind of old and likely uncompatible with Python 3 - but is available in Galaxy. Should be targetting requests + requests_toolbelt I guess.

... with pulsar counterparts.

... with pulsar counterparts.

... also improve error log location.

Deprecating old wsgi app_factory module in lieu of lwr.web.wsgi.

Outputs weren't getting rewritten when using remote_copy and remote_transfer, this made a bunch of tests pass that should have been failing and things seem more correct than they actually were.

The exact outlines of how this works will probably change as Galaxy is plugged in, but I am trying to get client as independent of Galaxy as reasonably possible.

This also allows a full integration test using RemoteTransferActions.