Ordered output for Terraform resources with 'count'

[SamiHiltunen]

Hey,

First of, thanks for the tool! It makes using Ansible with Terraform very convenient.

This PR sorts the output groups that correlate to Terraform resources created with count.

As an example, check this old generated inventory. The kafka list is not ordered a in the same manner as it is in Terraform.

...
"kafka":  ["kafka1", "kafka3", "kafka2"],
"kafka.0": "xxx.xxx.xxx.xxx",
"kafka.1": "yyy.yyy.yyy.yyy",
"kafka.2": "zzz.zzz.zzz.zzz",
...

After this PR, the output of kafka list will be in proper order:

...
"kafka":  ["kafka1", "kafka2", "kafka3"],
"kafka.0": "xxx.xxx.xxx.xxx",
"kafka.1": "yyy.yyy.yyy.yyy",
"kafka.2": "zzz.zzz.zzz.zzz",
...

As an aside, it will print to standard error when a key is overwritten during resource gathering. The tool has an issue, where resources of different types will overwrite others with the same name. Also, having tags with same value as the name of a resource will overwrite the old value. As it is not solvable in a backwards compatible manner, I decided to just print this out so the users will at least see the errors.

[adammck]

Thanks, this looks like a very sensible change. I'll try to get it reviewed in the next couple of days. Could you rebase against master, to ensure that it works with #72?

[SamiHiltunen]

Hey, I've rebased on master and the tests are passing.

[Dwaynekj]

This would also be useful for me

[SamiHiltunen]

@adammck Polite ping! Would it be possible to have this merged?

[adammck]

Thanks very much for this! Sorry about the delay on reviewing it.

[Dwaynekj]

Awesome! @adammck When do you think another official release will be packaged?