Fixed phantom cluster blocks in databricks_pipeline (#1508)

Current problem reported in #1401 is caused by the behaviour of the complex data types
inside the blocks that use the `slice_set` as a type - when we make changes the phantom
blocks are generated in addition to required changes.  Adding a custom hash function
didn't help.

This PR attempts to fix that issue other way - by changing the type of `cluster` blocks
from the `TypeSet` to the `TypeList`.  This approach, besides fixing the original issue
will also help with decreasing the amount of plan changes when making changes in the
`cluster` block.

The side effect of this change is that when upgrading to a new version, Terraform will
show changes in the DLT pipeline resource even if no actual changes were made in the code -
this is a result of changing the order of the `cluster` blocks

Author: alexott

pipelines/acceptance/pipeline_test.go

@@ -95,7 +95,7 @@ func TestAccPipelineResource_CreatePipeline(t *testing.T) {
 
 func TestAccAwsPipelineResource_CreatePipeline(t *testing.T) {
 	if cloud, ok := os.LookupEnv("CLOUD_ENV"); !ok || cloud != "aws" {
-		t.Skip("Test is only for CLOUD_ENV=AWS")
+		t.Skip("Test is only for CLOUD_ENV=aws")
 	}
 	acceptance.Test(t, []acceptance.Step{
 		{

pipelines/resource_pipeline.go

@@ -65,7 +65,7 @@ type PipelineSpec struct {
 	Name                string            `json:"name,omitempty"`
 	Storage             string            `json:"storage,omitempty" tf:"force_new"`
 	Configuration       map[string]string `json:"configuration,omitempty"`
-	Clusters            []pipelineCluster `json:"clusters,omitempty" tf:"slice_set,alias:cluster"`
+	Clusters            []pipelineCluster `json:"clusters,omitempty" tf:"alias:cluster"`
 	Libraries           []PipelineLibrary `json:"libraries,omitempty" tf:"slice_set,alias:library"`
 	Filters             *filters          `json:"filters,omitempty"`
 	Continuous          bool              `json:"continuous,omitempty"`