amalucelli / terraform-provider-nextdns Goto Github PK

Re:

Why's that? Create will create it, so removing that resource should then remove it again. This would be the equivalent of create only importing an existing profile, but there's already terraform import for that; likewise if 'delete from state only' is what someone wants, there's terraform state rm for that.

When writing an example like this:

resource "nextdns_rewrite" "test" {
  profile_id = nextdns_profile.i.id

  rewrite {
    domain  = "test.example.com"
    address = "1.1.1.1"
  }
}

resource "nextdns_rewrite" "test2" {
  profile_id = nextdns_profile.i.id

  rewrite {
    domain  = "test2.example.com"
    address = "8.8.8.8"
  }
}

The first plan / apply works, on a subsequent plan / apply both try and overwrite each other:

  # module.nextdns.nextdns_rewrite.test will be updated in-place
  ~ resource "nextdns_rewrite" "test" {
        id         = "<snipped>"
        # (1 unchanged attribute hidden)

      - rewrite {
          - address = "8.8.8.8" -> null
          - domain  = "test2.example.com" -> null
        }

        # (1 unchanged block hidden)
    }

  # module.nextdns.nextdns_rewrite.test2 will be updated in-place
  ~ resource "nextdns_rewrite" "test2" {
        id         = "<snipped>"
        # (1 unchanged attribute hidden)

      - rewrite {
          - address = "1.1.1.1" -> null
          - domain  = "test.example.com" -> null
        }

        # (1 unchanged block hidden)
    }

I imagine this might be difficult if the API just returns a list of rewrites instead of each rewrite being its own object, is this something that can be fixed/mitigated in the provider or should we use a lifecycle block to ignore the extra values?

Today is required for the user to build and install the provider locally before consuming the resources it exposes. It's possible to improve this flow by pushing the module to the Terraform Registry by following this guide.