Fixes #36462 - check for presence of content facet before assigning CVE

[jeremylenz]

What are the changes introduced in this pull request?

Sometimes in Katello we assume that all hosts have content facets. We should stop doing that.

Specifically in this case, the host edit page will try to inherit content facet attributes from a hostgroup without first checking if the host has a content facet.

Before assigning a content view environment based on the deprecated content_view_id and lifecycle_environment_id attributes, we should check if the content facet actually exists.

Considerations taken when implementing this change?

same as #10595 and #10600 ugh I'm getting tired of saying this

What are the testing steps for this pull request?

Create a hostgroup with CV and LCE
Create a host (Hosts > Create Host) - I did it by syncing a kickstart repo
In rails console, remove the host's content facet (verify that myhost.content_facet returns nil)
Edit the host and attempt to add/remove the hostgroup (you may have to fill in other required fields, it's a pain I know sorry)

Before patch:

NoMethodError (undefined method `assign_single_environment' for nil:NilClass):
13:38:08 rails.1   |  a198b8f9 |   
13:38:08 rails.1   |  a198b8f9 | /home/vagrant/katello/app/models/katello/concerns/host_managed_extensions.rb:14:in `update'

After patch: host is edited without error

[theforeman-bot]

Issues: #36462

[ianballou]

Works like a charm. Do you think adding any tests make sense to look out for regressions?

[jeremylenz]

Do you think adding any tests make sense to look out for regressions?

I did think about adding tests. But once the multi-CV feature is finished, a lot of this logic will be quite different anyway. Since the ubiquity of assign_single_environment is (hopefully) temporary, I wasn't sure if it was worth the effort. Happy to add some if you think so, though.

[ianballou]

All good, if this is taken out then we shouldn't have regressions.

[jeremylenz]

Added a test in hosts_controller_test 👍