Prior to this commit the spec tests had been updated to a wrong
state. This fixes them for PE again

Prior to this commit I merged a PR while drunk, that broke all my test

Prior to this commit we failed lint.
This might be the best of both worlds

Prior to this commit we did not have a webhook for github enterprise
This commit adds the new class which is a first crack at this. This
currently requires PE.

Prior to this we did not install sinatra
PE only right now

Prior to this commit request.host is not that useful

Changed to post

Prior to this commit I was using the fac this was undef to make some
magic happen but many people found it confusing. This commit explicitly
sets this to undef to the defaults still working correctly in the config
class

This patch allows me to have a symlink for r10k.yaml that points to
where the config file is located. This is useful if you want to put the
config file in a non-standard location while not having to tell r10k to
look in a different place for the config file.

This patch allows you to parameterize the inclusion of
r10k::prerun_command and keep you from having to include a separate
class. The code optionally does an include on the class in order to
maintain backward compatibility. Recommend that in a future release the
code from the prerun_command class is moved into the r10k class where it
is currently being included.

Updating gem can have various unforseen side effects. On RHEL 6.x
the update changes the default gem path which renders all gems
previously installed unavailable without special configuration.

Issue #37

Allow a postrun_command to optionally be set.  This is useful for
running r10k after a catalog application, so that in the event of an
r10k failure, a catalog is still applied.

Add an 'ensure' parameter to both the prerun_command and
postrun_command, defaulting to 'present'.

Update the spec tests for the ensuring 'absent' and to test
postrun_command.

These could be in the same class, but this commit tries to preserve
backwards compatability.

Add ability to specify a symlink for r10k.yaml

Add class for postrun_command

never update gem with rubygems-update

ensure webhook initscript is ordered correctly

This patch allows me to have a symlink for r10k.yaml that points to
where the config file is located. This is useful if you want to put the
config file in a non-standard location while not having to tell r10k to
look in a different place for the config file.

This patch allows you to parameterize the inclusion of
r10k::prerun_command and keep you from having to include a separate
class. The code optionally does an include on the class in order to
maintain backward compatibility. Recommend that in a future release the
code from the prerun_command class is moved into the r10k class where it
is currently being included.

Updating gem can have various unforseen side effects. On RHEL 6.x
the update changes the default gem path which renders all gems
previously installed unavailable without special configuration.

Issue #37

Allow a postrun_command to optionally be set.  This is useful for
running r10k after a catalog application, so that in the event of an
r10k failure, a catalog is still applied.

Add an 'ensure' parameter to both the prerun_command and
postrun_command, defaulting to 'present'.

Update the spec tests for the ensuring 'absent' and to test
postrun_command.

These could be in the same class, but this commit tries to preserve
backwards compatability.

```
The r10k application failed to run, use -v for full error details: /opt/puppet/libexec/mcollective/mcollective/agent/r10k.ddl:64: syntax error, unexpected tASSOC, expecting keyword_end
             :display_as => "Specific environment"
                           ^

/opt/puppet/libexec/mcollective/mcollective/agent/r10k.ddl:64: syntax error, unexpected tASSOC, expecting keyword_end
             :display_as => "Specific environment"
                           ^ (SyntaxError)
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl/base.rb:80:in `instance_eval'  <----
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl/base.rb:80:in `loadddlfile'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl/base.rb:31:in `initialize'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl/agentddl.rb:43:in `initialize'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl.rb:85:in `new'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl.rb:85:in `rescue in load_and_cache'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl.rb:76:in `load_and_cache'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/ddl.rb:66:in `new'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/rpc/client.rb:90:in `initialize'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/rpc.rb:73:in `new'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/rpc.rb:73:in `rpcclient'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/application.rb:354:in `rpcclient'
    from /opt/puppet/libexec/mcollective/mcollective/application/r10k.rb:19:in `main'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/application.rb:285:in `run'
    from /opt/puppet/lib/ruby/site_ruby/1.9.1/mcollective/applications.rb:23:in `run'
    from /opt/puppet/bin/mco:20:in `<main>'
```

Syntax error on line 63. Must have committed an earlier copy.  Have to go back and review this code and submit new PR.

Fix a bug calling the wrong method.

Conflicts:
	files/agent/r10k.ddl

Prior to this commit we did not bikeshed enough on the naming for this
action, this should hopefully make it a little more intutuive

Prior to this commit we had some unimplemented code from a PR
this commit moves anything not currently be used into the land of wind
and ghosts and renames a few vars. This also removes the complexity of
how the private methods were deployed.

Support Puppet v3.6.0

Prior to this commit we were not consistant how
we passed values in given our simple loops
+ whitespace

Prior to this commit we had moved code aroudn without changing the enum
vars , also tried to make heads or tails as to why validate is not
working