Add MiqServer#server_role_names=

[agrare]

There is a #role and a #role= method that takes a comma separated
list, then a #server_role_names which takes an array but no
#server_roles_names=.

[agrare]

@jrafanie planning on using this for #19328

[jrafanie]

This makes sense. I don't know why we didn't have this before. I don't know how we can test this. Is it worth expecting assign_role and deactivate_roles get called?

[agrare]

@jrafanie yeah I did check that this path was tested by spec/models/miq_server/role_management_spec.rb by the role= context since that goes through server_role_names= now.

We could add a context similar to role= but instead of testing the comma separated list, test with arrays

[agrare]

@jrafanie added specs for server_role_names= specifically (mostly copied from the role= specs)

[miq-bot]

Checked commits agrare/manageiq@c57905f~...c686ae4 with ruby 2.4.6, rubocop 0.69.0, haml-lint 0.20.0, and yamllint 1.10.0
2 files checked, 3 offenses detected

app/models/miq_server/role_management.rb

• ⚠️ - Line 161, Col 5 - Lint/Void - Variable roles used in void context.
• ❗ - Line 167, Col 3 - Style/Alias - Use alias instead of alias_method in a module body.
• ❗ - Line 168, Col 3 - Style/Alias - Use alias instead of alias_method in a module body.

[jrafanie]

• ⚠️ - Line 161, Col 5 - Lint/Void - Variable roles used in void context.

that seems like a rubocop bug in parsing the code. Interesting. I wonder if that long line comment is messing it up.

[jrafanie]

I like this, only have a minor comment about sorting but LGTM so far.

[agrare]

that seems like a rubocop bug in parsing the code. Interesting. I wonder if that long line comment is messing it up.

Yeah, I was tempted to try return role just to see if it'd also complain about a pointless return 😆