provider/aws: Fix issue with ignoring the 'self' attribute of a Security Group rule

[catsby]

Fix issue where we ignored the self attribute of a security group rule.

To reproduce:

1. plan and apply the config below:

provider "aws" {
  region = "us-west-2"
}

resource "aws_security_group" "tf_test_self" {
  name = "tf_test_self"
  description = "tf_test_self"
  vpc_id = "<a vpc id>"

  ingress {
    from_port = 0
    to_port = 65535
    protocol = "tcp"
    self = true
  }

  tags {
    Name = "tf_test_self"
  }
}

2. After successfully creating, remove self = true from the ingress block
3. Run plan again

Expected:

A diff

Actual:

nope-nope-nope

Discovered while investigating #508, so credit to them for discovery 😄

[catsby]

@phinze sanity check please :)

[phinze]

Sanity: intact. 🙆‍♀️

Thinking about backcompat here - this will result in a recalculation of every security group rule hash, but I think that's fine because refresh will rewrite the state before plan runs a diff.

Maybe double check this quickly by doing an apply with the old code, then making sure the plan with the new code does not try to nuke-and-replace the world?

If that checks out, then this is g2g.

[catsby]

Checked the plan results between master and this branch, confirmed no change detected so backward compatibility should be 👍 , merging

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.