Update test for autocompletion of top-level commands

[SarahFrench]

This is a rabbit hole I explored today 😅

I was trying to find a satisfying way to test changes in #37569 to workspace command autocompletion (ie 733618f and 7497796).

I've settled on testing workspace-related autocompletion with more of a unit test, but I think this change to the original test I was taking inspiration from could still be merged.

Previously I think the test only asserted that autocomplete that got no autocompletion results didn't result in an error. Now the test shows that no error happens both when autocompletion returns a result vs an empty result, and also asserts about the returned value.

Target Release

N/A

Rollback Plan

• If a change needs to be reverted, we will roll out an update to the code within 7 days.

Changes to Security Controls

Are there any changes to security controls (access controls, encryption, logging) in this pull request? If so, explain.

CHANGELOG entry

• This change is user-facing and I added a changelog entry.
• This change is not user-facing.

[radeksimko]

I wish we could find a way of testing this without having to override the global os.Stdout and os.Args but I suppose it's hard if this is to be kept as "E2E" and use realMain() as the testable entrypoint. 🤔

[SarahFrench]

Yeah I think the alternative to this particular test is setting up a full E2E test that actually builds and uses a Terraform binary.

Regarding overriding global variables, my personal favourite is:

terraform/internal/command/command.go

Lines 12 to 13 in f5a28cf

	// Set to true when we're testing
	var test bool = false