Reset trigger prefixes for workspace resources when removed from the config

[nicsnet]

Description

Fixes a bug for tfe_workspace resource updates. Previously when creating a tfe_workspace resource with a trigger_prefixes configuration, this option remained even when the trigger_prefixes config option was removed during subsequent updates to this resource. terraform plan would continue outputting this change, but it would not take effect.

Testing plan

 # tfe_workspace.pleng_staging_us_west_2 will be updated in-place
  ~ resource "tfe_workspace" "pleng_staging_us_west_2" {
        auto_apply            = false
        external_id           = "ws-6YmtKxQffbqJpVE6"
        file_triggers_enabled = true
        id                    = "chavo4/pleng-staging-us-west-2"
        name                  = "pleng-staging-us-west-2"
        operations            = true
        organization          = "chavo4"
        queue_all_runs        = true
        terraform_version     = "0.12.12"
      ~ trigger_prefixes      = [
          - "567694625192/us-west-2",
        ]
        working_directory     = "567694625192/us-west-3"

        vcs_repo {
            identifier         = "nicsnet/terraform-provider-nics"
            ingress_submodules = false
            oauth_token_id     = "#######"
        }
    }

Output from acceptance tests

Please run the full suite of acceptance tests locally and include the output here.

$ make testacc

...

[stasquatch]

Everything looks good on my end, I just have one question. Is it within scope to allow working_directory to be reset? It can't be cleared once it's been set via the provider, similar to the trigger_prefixes.

[nicsnet]

The working_directory fix is part of this #137

[stasquatch]

LGTM