Allow per-manager job directories.

Update: Rebased on to the latest master and documentation for this feature in docs/job_managers.rst and linked to this general manager configuration information from app.yml.sample (per excellent request of @mvdbeek).

Ensure destination directories exist when copying files

Fail jobs that error during job submission

Previously jobs that failed to submit (bad resource requests, malformed
scripts, no qsub available, staging problems) would remain in waiting status
until pulsar was restarted.

for the queued_cli manager.

Fix PBS status reporting

Without this commit only very short jobs completed successfully, because the
status was always set to complete. This was because the status returned by
qstat contains leading spaces before the `job_state = Q` part, so I'm adding a
`.strip()`.

More Progress on an Embedded Pulsar Runner

Pulsar has to be a little smarter than I assumed in 09f7ec3e because it writes out job metrics on the server.

Working directory is below metadata directory in Galaxy 16.04 - so need to handle that.

 - When client uploads files, log them.

[WIP] Metadata directory

 - Docstrings for AMQP tests.
 - Skip curl tests if pycurl is unavailable.
 - Time the AMQP test in case this is what is causing Travis to hang.

 - Fixup docstrings.
 - Log native specification and source if conflicting.

xref #94

... for metadata about ... data and jobs. So formal Galaxy metadata stuff, job instrument files, etc....

Handle transfer of metadata and working directory outputs correctly in Galaxy 16.04.

It was removed from Galaxy.

Use a global DRMAA session