module vs command module for restarting.  May be some bugs and requires the service to implement 'status' -- and probably some better error handling (i.e. return JSON with "failed" element if failed).

Improvements welcome.

override the host file in a playbook to load different groups
of hosts.

Conflicts:
	TODO.md

onchange events like Puppet's file providers do.

Cleanup is due in runner.py

on the system for use in a later template module.   These values
could also be used for module-specific config.

use logger to track all events run via ansible on the remote host.

- add pattern parameter to self._matches() and match_hosts() so you can query
  for matches against the host list
- these are useful when writing scripts and you want to tell the user
  what hosts you will be running on before  actually executing anything.

* upstream/master: (2 commits)
  Added idea about how to do async + timeout in modules
  ...

/usr/bin/logger should probably be a constant or defined in some sort of config as well as the log priority

the API is important.

Because everything should have a man page

as I figure out how logging/output might look.  A couple of major TODO features
also listed in the file.