Handle possibility of no orchestration stacks

[djberg96]

Fixes a potential bug in the get_stack_templates method.

Addresses #83

[djberg96]

@miq-bot add_label fine/yes

[djberg96]

@miq-bot add_label euwe/yes

[bzwei]

LTGM. Thanks.

[bronaghs]

@djberg96 - any chance of a spec?

[djberg96]

@bronaghs @bzwei I will need some help with that. :)

[blomquisg]

@djberg96 the easiest way I can see testing this is to refactor the way that deployments are collected by adding a get_deployments method that simply calls get_data_for_this_region(@tds, 'list').

Then, you can easily stub out that method to return an empty array, or nil, or whatever you need it to return to reproduce this situation. You could probably even stub that out in the VCR-based specs, but that might result in the dreaded "Unused HTTP Calls" error.

Without that wrapper method, it looks like lots and lots of digging holes to stub out weird things.

#justathought

[djberg96]

@blomquisg I was able to figure it out. :)

That said, I did split out a get_deployments method. While not strictly necessary for this PR, I included it anyway to make future testing potentially easier.

@bronaghs @bzwei Please review.

[miq-bot]

Checked commit https://github.com/djberg96/manageiq-providers-azure/commit/2d4a7fccf144b3734e355e9a1da9153aee65c7c3 with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0
2 files checked, 2 offenses detected

app/models/manageiq/providers/azure/cloud_manager/refresh_parser.rb

• ❗ - Line 81, Col 11 - Style/AccessorMethodName - Do not prefix reader method names with get_.
• ❗ - Line 86, Col 11 - Style/AccessorMethodName - Do not prefix reader method names with get_.

[simaishi]

@djberg96 Is there a BZ for this? Can you please create if it doesn't exist?

[djberg96]

@simaishi https://bugzilla.redhat.com/show_bug.cgi?id=1467405

[simaishi]

@djberg96 app/models/manageiq/providers/azure/cloud_manager/refresh_parser.rb is giving a conflict backporting to Euwe. Please resolve conflict and create an Euwe-specific PR (referencing this one) or suggest other PRs to backport.

++<<<<<<< HEAD
 +        old_status_message = get_resource_status_message(old_resource)
 +
 +        old_resource.properties['status_message'] = if old_status_message
 +                                                      "#{old_status_message}\n#{new_status_message}"
 +                                                    else
 +                                                      new_status_message
 +                                                    end
 +      end
 +
 +      def get_stack_template(stack, content)
 +        process_collection([stack], :orchestration_templates) { |the_stack| parse_stack_template(the_stack, content) }
++=======
+         # link stacks to templates, convert raw_template to template
+         Hash(@data_index[:orchestration_stacks]).each do |_stack_uid, stack|
+           raw_template = stack[:orchestration_template]
+           stack[:orchestration_template] = @data_index.fetch_path(:orchestration_templates, raw_template[:uid])
+         end
++>>>>>>> cb8acf4... Merge pull request #84 from djberg96/orchestration_stack_fix
        end

[djberg96]

@simaishi I don't know that it needs backporting to Euwe. @bzwei The logic changed between Euwe and Fine, didn't it?

[blomquisg]

@djberg96 you added the euwe/yes label :)

[djberg96]

@blomquisg dang, i suck.

[djberg96]

@miq-bot add_label euwe/no

[djberg96]

@simaishi I will submit a separate PR for Euwe. Sorry about that.

[djberg96]

@miq-bot remove_label euwe/yes

[simaishi]

Fine backport details:

$ git log -1
commit 0bea1a1b12325dfb3cbc9630158369cb72214fc8
Author: Greg Blomquist <[email protected]>
Date:   Tue Jun 27 16:27:31 2017 -0400

    Merge pull request #84 from djberg96/orchestration_stack_fix
    
    Handle possibility of no orchestration stacks
    (cherry picked from commit cb8acf4422799d9861f4e43a9c07d7038fbe7405)
    
    https://bugzilla.redhat.com/show_bug.cgi?id=1468703

[simaishi]

Backported to Euwe via ManageIQ/manageiq#15527