Rename module_utils/elb_utils.py to module_utils/elbv2_utils.py

[alinabuzachis]

SUMMARY

Rename module_utils/elb_utils.py to module_utils/elbv2_utils.py, since it includes utilities for the elbv2 client and we may also have utilities just for elb.

ISSUE TYPE

• Bugfix Pull Request
• Docs Pull Request
• Feature Pull Request
• New Module Pull Request

COMPONENT NAME

ADDITIONAL INFORMATION

[softwarefactory-project-zuul]

Build succeeded.
https://ansible.softwarefactory-project.io/zuul/buildset/c72ffa453f734a61838bf237dfab5ef0

✔️ ansible-galaxy-importer SUCCESS in 4m 41s
✔️ build-ansible-collection SUCCESS in 10m 34s
✔️ ansible-test-splitter SUCCESS in 4m 20s
✔️ integration-amazon.aws-1 SUCCESS in 32m 56s
✔️ integration-amazon.aws-2 SUCCESS in 10m 20s
✔️ integration-community.aws-1 SUCCESS in 17m 54s
Skipped 41 jobs

[tremble]

Not completely against the idea of moving this code about, but will ask the question if module_utils/elbv2_utils is the right name...

I'd either go for module_utils/elbv2/utils.py or possibly splitting the contents into multiple files under elbv2/ I would also ask where we want folks loading things like the error handler / exception from: maybe module_utils/elbv2? (probably needs to live in module_utils/elbv2/utils.py to avoid loops though)

[tremble]

Not completely against the idea of moving this code about, but will ask the question if module_utils/elbv2_utils is the right name...

To answer my own question here. We can't use module_utils/elbv2/ because of the "empty-init" policy:
https://docs.ansible.com/ansible-core/devel/dev_guide/testing/sanity/empty-init.html

For others (module_utils), we want the possibility of using Python namespaces which an empty init.py will allow for.

Personally I think this policy is being a little overly restrictive, but it is what it is.

[softwarefactory-project-zuul]

Build failed.
https://ansible.softwarefactory-project.io/zuul/buildset/e4ff9c1eea66497ba2b7a8c95b33b188

✔️ ansible-galaxy-importer SUCCESS in 8m 12s
✔️ build-ansible-collection SUCCESS in 10m 43s
✔️ ansible-test-splitter SUCCESS in 4m 16s
❌ integration-amazon.aws-1 FAILURE in 1h 00m 05s
✔️ integration-amazon.aws-2 SUCCESS in 13m 11s
✔️ integration-community.aws-1 SUCCESS in 16m 09s
Skipped 41 jobs

[tremble]

Suggested change

	minor_changes:
	breaking_changes:

Technically this is a breaking change, but I think it's the right change to make in this case.

Further thought:
It's a breaking change because someone can in theory import from module_utils/elb_utils/...

Do we want people directly importing things from module_utils/elbv2_utils ? The bulk of this code is just calls wrapped with helpers.

Do we in fact move this to module_utils/_elbv2 (marking the module private), and only directly import AnsibleELBv2Error which we want to publicly expose (for the other calls we import ._elbv2 _elbv2_utils and then do things like _elbv2_utils.describe_load_balancer_attributes(self.connection, self.elb["LoadBalancerArn"])