Implement PlanResourceChange on Destroy Support

[bflad]

Module version

v0.8.0

Use-cases

The provider protocol may soon support enabling destroy plans that call the PlanResourceChange RPC. This will enable providers to return warning or error diagnostics when resources are destroy planned.

This framework can opt into the new capability and support it via the existing ResourceWithModifyPlan interface ModifyPlan methods. Technically this is already enabled in the framework logic, however there are no protections if providers do something unexpected, such as modifying the response planned state from being empty, which should have guardrails to prevent unexpected behaviors.

Proposal

• When implemented in the protocol, upgrade to terraform-plugin-go version which supports the feature.
• Enable the capability for the framework server (e.g. in the current proposal, enable GetProviderSchema > Capabilities > PlanDestroy).
• Ensure the framework server implements an error if the response plan has any values set (refer to TODOs in server_planresourcechange_test.go unit testing)
• Announce as a potentially breaking change, as providers using this framework can implement support via the existing ResourceWithModifyPlan interface ModifyPlan (future provider capabilities closer or after 1.0 should likely be surfaced as opt-in flags)

[bflad]

Depends on hashicorp/terraform-plugin-go#204, which will need to be reviewed, merged, and released, before this functionality can be enabled in this module.

[github-actions]

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