Also added some error handling to the fetch_url() call in the
apt_repository module, so that failures to look up the PPA info
are properly handled.

Fixes #7322

Fixes #7448

Fixes #7416

If the module directory is not writable/executable to the current user
(most likely because of a sudo to a non-root user), the ssh_wrapper
will be created in the default location for mkstemp() calls. To facilitate
the deletion of these new files, a new mechanism for cleaning up files
created by the module was also added.

Fixes #7375

Fixes #7385

ENV variable value could contain '=' (for example mysql://host/db?pool=10)

Fixes #7387

`decoded_name` was created twice, each from `rset.name`

So, the second call to `.replace(r'\100', '@')` overwrites decoded_name, discarding the result of the call to `.replace(r'\052', '*')`

I had a problem with wildcard domains that was fixed by this patch.

Since some of the vars contained in there may be used in the templating
of other variables. This also reverts e83a494e which originally fixed
issue #6979 but broke some other variable related things.

Fixes #6979
Fixes #7321
Fixes #7332

Fixes #7405

This is a corner case for remote file systems that don't support
chown() and where the source and destination for the atomic_move
are on that remote file system.

Fixes #7372

Fixes #7345

Fixes #7384

Fixes #7364

Fixes #7363

This fixed the just committed change handling non existent Ghost state.

This doesn't exist anymore and causes ansible to throw a KeyError

modify mode.

If `get_all_instances` returns multiple reservations, the old wait loop only
dealt with the first reservation. Thus, the wait loop may end before all
instances get to be running/stopped.

Also clean up the code a little.

Upon a second run, the default egress rule will be removed when a
vpc is specified but no other egress rules were set. This patch
corrects that behavior by removing the default egress rule from the
list of unmatched outbound rules.

Fixes #7309

Fixes #7308

Add integration tests for group_by module. Fixed bug introduced by ansible/ansible#7273 while also fixing issue described in ansible/ansible#6953

Fixes #7307

Fixes #7295

ignore remnants from doing a setup.py install

* the current state of the ELB was not reflected properly when checking
  the status after a change was made.
* invalid zones caused a traceback when enabling/disabling zones

Fixes #7270

Docker inventory bugfix for potential traceback