Fixes #12053

1. The test did "name: '{{hostnames}}.{{item}}'" inside a with_sequence
   loop, which didn't do what was intended: it expanded hostnames into
   an array, appended ".1", and set name to the resulting string. This
   can be converted to a simple with_items loop.

2. Some of the entries in hostnames contained punctuation characters,
   which I see no reason to support in inventory hostnames anyway.

3. Once the add_host failures are fixed, the playbook later fails when
   the unicode hostnames are interpolated into debug output in ssh.py
   due to an encoding error. This is only one of the many places that
   may fail when using unicode inventory hostnames; we work around it
   by providing an ansible_ssh_host setting.

Labels must start with an alphanumeric character, may contain
alphanumeric characters or hyphens, but must not end with a hyphen.
We enforce those rules, but allow underscores wherever hyphens are
accepted, and allow alphanumeric ranges anywhere.

We relax the definition of "alphanumeric" to include Unicode characters
even though such inventory hostnames cannot be used in practice unless
an ansible_ssh_host is set for each of them.

We still don't enforce length restrictions—the fact that we have to
accept ranges makes it more complex, and it doesn't seem especially
worthwhile.

This adds a parse_address(pattern) utility function that returns
(host,port), and uses it wherever where we accept IPv4 and IPv6
addresses and hostnames (or host patterns): the inventory parser
the the add_host action plugin.

It also introduces a more extensive set of unit tests that supersedes
the old add_host unit tests (which didn't actually test add_host, but
only the parsing function).

There was code to support set literals (on Python 2.7 and newer), but it
was buggy: SAFE_NODES.union() doesn't modify SAFE_NODES in place,
instead it returns a new set object that is then silently discarded.

I added a unit test and fixed the code.  I also changed the version
check to use sys.version_tuple instead of a string comparison, for
consistency with the subsequent Python 3.4 version check that I added in
the previous commit.

Fixes #12245

The full error was

======================================================================
ERROR: test_task_executor_execute (units.executor.test_task_executor.TestTaskExecutor)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/mg/src/ansible/test/units/executor/test_task_executor.py", line 252, in test_task_executor_execute
    mock_action.run.return_value = dict(ansible_facts=dict())
  File "/home/mg/src/ansible/lib/ansible/executor/task_executor.py", line 317, in _execute
    if self._task.async > 0:
TypeError: unorderable types: MagicMock() > int()

----------------------------------------------------------------------

Experiments show that Python 2 MagicMock() > 0 is true, so I'm setting
the async property on mock_task to 1.  (If I set it to 0, the test fails
anyway.)

Required some rewiring in inventory code to make sure we're using
the DataLoader class for some data file operations, which makes mocking
them much easier.

Also identified two corner cases not currently handled by the code, related
to inventory variable sources and which one "wins". Also noticed we weren't
properly merging variables from multiple group/host_var file locations
(inventory directory vs. playbook directory locations) so fixed as well.

simplify become testing and handling, we had drifted and were doulbe checking prompt, become and become_pass
fixed tests to conform to new signature and now tests both with and w/o password
now we are more explicit about self.prompt

Fixes #12198
Fixes #12241

Also fix the bug (missing from six import iteritems) I introduced in
823677b4.

Replace .iteritems() with six.iteritems() everywhere except in
module_utils (because there's no 'six' on the remote host).  And except
in lib/ansible/galaxy/data/metadata_template.j2, because I'm not sure
six is available there.

  ansible is passing unicode arond internally so we should test the same
  data.
* Add a zero length test for _count_newlines and fix the zero newlines
  test to have no newlines.

Because __cmp__ is not supported on Python 3.

* Dedupe combine_vars() code (removed from VariableManager)
* Fix merge_hash algorithm to preserve the type
* unittest combine_vars and merge_hash

Fixes #11891

`assert (condition, message)` gets parsed by Python as `assert
a_two_tuple`, and a 2-element tuple is never False.

Discovered by compileall on Python 3.4, which emits a SyntaxWarning for
this common mistake.

Note that this test was broken in devel because it was really just
duplicating the AES256 test because setting v.cipher_name to 'AES'
no longer selected AES after it was de-write-whitelisted.

Now that we've removed the VaultAES encryption code, we embed static
output from an earlier version and test that we can decrypt it.

Fixes #11981
Fixes #11995
Fixes #12039
Fixes #12077