Issue LIBCLOUD-464: Add security group delete/destroy into the EC2 drive...

[cderamus]

...r

[Kami]

For consistency with other drivers and to make it easier to promote security group management methods to a top-level API in the near future, I would do the following:

1. Add two new methods - ex_destroy_security_group_by_id(group_id) and ex_destroy_security_group_by_name(group_name)
2. Add a wrapper method - ex_destroy_security_group(name) which calls ex_destroy_security_group_by_name underneath

As noted above, those methods will hopefully be promoted to a top-level API in the near-future and guaranteeing consistency and compatibility between drivers will be a bit easier then.

Edit: Fixed a typo.

[cderamus]

Let me know if this is more in the vein of what you'd like buddy.

[Kami]

Ugh, sorry, looks like I've missed one thing in my previous review.

For consistency with other drivers, can you please also rename the methods from "destroy" to "delete" (ex_destroy_security_group -> ex_delete_security_group, ...) ?

Besides that, the patch looks good.

[cderamus]

Ah, nice catch on that. I try to stay consistent, but I clearly missed that one. I've updated for your review!

[Kami]

Thanks!

While reviewing and merging the patch, I've noticed that you forgot to update ex_delete_security_group method and the tests didn't detect that because they had a typo in the name so they didn't actually run (text instead of test).

I've squashed your commits, fixed those two issues in 4df8dd2 & adba6de and merged your patch into trunk.

When I find some more time, I'll also have a look at how we can automatically detect and prevent issues like this from happening in the feature. It seems that for a start, we could do a simple integration with coverage.py and make sure that all the newly added test cases are actually executed.

[cderamus]

Ah, that's what I get for updating code with egg nog in hand. Thanks for catching that!