Added validator less bootstrap support for knife bootstrap windows winrm

[Vasu1105]

1. This change needs chef >= 12.2
2. --node-name option needs to be passed while bootstrap
3. Specs pending

[btm]

@Vasu1105, have you looked at the failing tests?

Fixes #222

[Vasu1105]

@btm Yes I have fixed test cases for Chef >= 12.2. The test cases failing for chef 11 and I am currently looking into to fix those.

[btm]

This change needs chef >= 12.2

Do you mean you need Chef >= 12.2 for this feature to work?

I presume you don't mean that this breaks knife-windows for anyone on Chef < 12.2?

--node-name option needs to be passed while bootstrap

Please incorporate chef/chef#3325 to require a node be set when using this option, otherwise you see confusing output like this:

Node  exists, overwrite it? (Y/N) n
You said no, so I'm done here.

Specs pending

It looks like you've fixed these.

[btm]

we need a note in here about how this is because Chef 11 doesn't have Chef::Knife::Bootstrap.client_builder.

If we're running against chef 12 (reverting 7aab00d), we can just say:

winrm_bootstrapper.client_builder = instance_double("Chef::Knife::Bootstrap::ClientBuilder", :run => nil, :client_path => nil)

I think we need to modify travis to test both chef11 + chef12. I'm looking at that.

[Vasu1105]

@btm

1. Yes it won't break the knife-windows if chef version is < 12.2.
2. I will be working on require --node-name change.
3. winrm_bootstrapper.client_builder = instance_double("Chef::Knife::Bootstrap::ClientBuilder", :run => nil, :client_path => nil) is possible by using rspec filter where we need to add filter for chef 12 and chef 11
4. We realized that validatorless bootstrap should not be allowed if --windows-authentication-protocol set to basic. So, we are thinking to add validation for windows-authentication-protocol and transport =ssl as well

[btm]

@Vasu1105 can you rebase this against master so you pull in #229?

Then we could add some rspec filters similar to what's in Chef

def chef_11?
  Chef::VERSION.split('.').first.to_i == 11
end

def chef_12?
  Chef::VERSION.split('.').first.to_i == 12
end

Then I think we need to add code that raises an error if you don't have a validation.pem and you've got Chef 11, and a corresponding test.

[Vasu1105]

@btm I am looking into all review comments and fixing them. I have rebased my branch.

[btm]

Doesn't this duplicate the same error being raised in Chef::Knife::WinrmCommandSharedFunctions.resolve_winrm_transport_options (in winrm_knife_base.rb)?

[btm]

It'd be great if you could rebase this against master instead of merging and force push back onto this branch. I'm seeing some other merge commits leak in from the merge.

[btm]

Merged via #234