Bug: Running terraform inventory on empty tfstate returns invalid inventory

[marcospedreiro]

Context

I define a cluster-name for some infrastructure I'm provisioning in aws. Before I run terraform apply I want to run terraform destroy on that cluster-name to make sure that any previous/existing infrastructure is cleaned up before I create new infrastructure.

Bug

If I pull tfstate for cluster-name that does not exist and then try to run an ansible playbook I get the following error:

ERROR! The file /usr/local/bin/terraform-inventory is marked as executable, but failed to execute correctly. If this is not supposed to be an executable script, correct this with `chmod -x /usr/local/bin/terraform-inventory`.
You defined a group "all" with bad data for the host list:
 {u'hosts': None, u'vars': {<vars here>}}
'utf8' codec can't decode byte 0x80 in position 24: invalid start byte
'utf8' codec can't decode byte 0x80 in position 24: invalid start byte

If I run

/usr/local/bin/terraform-inventory --list <empty-tfstate> | jq .

I get

{
  "all": {
    "hosts": null,
    "vars": {
      <vars here>
    }
  }
}

If I write this to a file and point ansible to it directly I get:

ERROR! 'NoneType' object is not iterable
./test-inventory:2: Expected key=value host variable assignment, got: {

If I change the null field to an empty json array as below:

{
  "all": {
    "hosts": [], 
    "vars": {
      <vars here>
    }
  }
}

And then run ansible with this file it runs as expected.