Currently when more than one playbook is provided on the commandline, ansible-playbook --list-hosts will only consider the first playbook and stop. This change will make it work for the various different playbooks.

When the inventory is empty, or the provided --limit subset returns an empty intersection, we don't want to loop over all plays but bail out at the very start.

refers to #1345

--limit takes localhost, FQDN, and 127.0.0.1.

Signed-off-by: Brian Coca <briancoca+ansible@gmail.com>

check. Lots of room to make this more interesting - but not today

Removes git checkout after playbook run.

Direct any ansible-pull specific messages to stderr.
Introduce try_playbook() and select_playbook() to remove try/except-y.

Merged commits: 3cd25b5 and ed9fc76.  Some massaging of changes applied.

If all hosts in a play fail, fail the whole playbook and don't bother printing out every remaining task.

being run against for each playbook

add the limit option, which can be used to further confine the patterns selected by "hosts:" in ansible-playbooks
to an additional pattern (a subset) specified on the command line.  For instance, a playbook could be reusable
and target "webservers" and "dbservers", but you want to test only in the stage environment, or a few boxes at a time.

whitespace + remove deprecated YAML parser (migration script lives in examples/scripts and warning was added
in 0.6 release)

(make colorized playbook runs not suck on black terminals)

Remove the -D module debug flag, which no longer is functional due to sudo pty requirements, and replace with -v/--verbose.

This flag will show playbook output from non-failing commands.  -v is also added to /usr/bin/ansible, but not  yet used.

I also gutted some internals code dealing with 'invocations' which allowed the callback to know what module invoked
it.  This is not something 0.5 does or needed, so callbacks have been simplified.

Since host variables are becoming important, it did not make sense to sustain --override-hosts, with the ability
to create hosts that didn't have inventory information, but also existed, in various groups.

This eliminates the creation of runner and playbook instances and just
invokes the processes ansible and ansible-playbook.

A first stab at a pull-based model for ansible.  This does two things:
    1. Invoke the git module via Runner to set up a git repository on the
       localhost.  It sets up Runner to use transport='local' and forces
       the inventory to just 'localhost'.
    2. Run any playbooks provided.  By default, this wants to run the
       playbook local.yml.  This also sets transport='local' and sets
       the host_list to a list: localhost, fqdn, and hostname.
The reason for setting the host_list and not using override_hosts is
because there may be plays in the playbook that are not meant for a
specific host.  That is, if the git repository is for the entire site
and not host-specific, you don't want to override hosts and apply all
plays to any given host.  This has the downside of potentially running a
play three times if the play is defined for 'hosts: all'.