LWRP focused rewrite

[mal]

I was having some issues reverting values, so I started down a path to cache the defaults as they're removed and kept tweaking and tweaking and here we are.

I completely understand if you'd prefer to leave this as a separate fork, but I thought I'd offer it up and let you decide for youselves.

Rather than duplicate everything here, the best option is to skim over the updated usage section in the readme for starters, and then then continue below where I'll outline the less visible changes.

---

• These changes address #1 by way of the conclusion reached in this Chef Bug tracker issue. The result of trying to set an invalid or unsupported key is an explicit error explain that. The best way to handle this would be using the LWRP and something similar to not_if platform_family?('freebsd') to prevent the attempt to set keys based on platform.
• They also address the caching issue mentioned above, allowing :remove to immediately return parameter values to their system default.
• Everything goes through the LWRP, so if a config file is available values will always be persisted in the course of the chef run (with the ability to force the write sooner using action :save.
• Support for flat keys i.e. 'vm.swappiness' => 20 and vm: { swappiness: 20 } render the same outcome
• Fix issue where if key was removed from node.sysctl.params in a subsequent convergence, the value would be removed from the persisted file, but not unset from the server, violating the expected idempotent behaviour of cookbooks
• Fix whyrun? support as it was claiming to support whyrun behavior and then changing kernel values in what should have been dry runs.
• Split out BSD testing into a separate suite due to current suites using keys BSD does not support.

---

I've run the full test suite against this PR and it passes with two exceptions; the first is the integration tests for fedora-20 which fails for me due to this vagrant issue (but it passes once the suite is restarted), and the second is a foodcritic issue which I believe to be a false positive.

Happy to answer any questions you may have.

[mal]

Fixed tests (was testing the for the wrong directory on BSD).

Also managed to work around the FC false positive, and solved another issue by upgrading to a post 4.0.0 commit of chefspec (issue here).

[svanzoest]

Hi @mal,

I appreciate your contributions. As you say this changes the syntax quite a bit, so we will need to look at the pros and cons of that approach a bit. I am currently focusing all my free time on the apache2 cookbook, but hopefully will look at this next week.

[mal]

Thanks, appreciate the heads up. Look forward to hearing your thoughts next week!

[mal]

Rebased and added some basic code to support testing of arch as soon as fauxhai supports it.

[mal]

@svanzoest pre-christmas personal PR triage, any hope for movement on this one?

[mal]

Closing due to lack of interest and the fact that it will now require further work to make it mergeable.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.