Ideas for skipping cleanup on error?

Just ran an apply/destroy cycle through tftest, where on destroy I ran into a timing error when waiting for a resource to change to the expected state:

```
Error: Error waiting for VPC Peering Connection (pcx-....) to be deleted: timeout while waiting for state to become 'rejected, deleted' (last state: 'active', timeout: 1m0s)
```

Ok, not great, but no big deal, just go into the test directory and re-run destroy... Except, the tfstate file is gone, so can't do that!