fix: Make state show command exit with code 1 if the stored state cannot be marshalled for rendering

[SarahFrench]

(Potential) Fix following changes in #33116

Prior to this change, if a user runs a state show command with an address for a resource that existed in state but didn't match the current schema of the resource then the user would see

• exit with code 0
• Error logged to stderr, for example : "Failed to marshal state to json: unsupported attribute "foobar""
• This message logged to stdout: "The state file is empty. No resources are represented"

This is misleading because the state file used for state show isn't empty, its just that the command hasn't been able to assemble a jsonformat.State correctly for output. And that suggests that we should return with code 1 before that is attempted.

Target Release

1.15.x

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.

[SarahFrench]

Could this be considered a breaking change?

I'd argue not, as the old behaviour was misleading and incorrect, though someone could have built a system that relied on the command not failing somewhere.

[mildwonkey]

On some level every fix in terraform has the potential to be a breaking change - the longer the behavior is present, the more likely people have already accounted for it in their workflows.

This is about the least-concerning breaking change I can think of since it's not a change in anything that modifies state and even if someone's automation for e.g. carefully checks for that error message if the show command succeeds (to understand that it actually failed), changing the exit code to properly return 1 probably isn't going to break anything.

I'd still call it a bug fix, but not a bad idea to get some other opinions! My process instincts are outdated

[SarahFrench]

Yeah agreed, previous projects I've worked on have been strict about breaking changes but in those cases state was always being modified. I don't think we have a blanket block on breaking changes here in Core as there's more grey area, but figured I'd flag to the reviewer anyways.

I'm happy to go ahead with this as a bug fix.

[github-actions]

I'm going to lock this pull request because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active contributions.
If you have found a problem that seems related to this change, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.