(#19845) block confines

[dalen]

Allow confine to take a block parameter, like:

confine { code... }

Builds on top of pull #408

[jeffmccune]

You could get rid of the conditional and simply do @block = block. If no block is passed then the block argument will be nil.

[dalen]

yes, but I still needed it for the others. but I can have them in a unless block_given?

[jeffmccune]

Ah yes, sorry.

So... I don't yet understand how this change set works. If we discard the fact we're confining to, how does Facter know what value to yield to the block?

[dalen]

It doesn't yield any value. Either you pass only a block to confine like for example:

confine { File.exist?('/usr/bin/lldpctl') }

Or you pass a fact and one or more values that it should match like

confine :manufacturer => /Dell.*/

The value it is matched against can however also be a Proc object in which case the fact value is yielded to it, like:

confine :ipaddress => lambda { |addr| IPAddr.new('192.168.0.0/16').include? addr }

But the new thing in this pull request compared to Pull #408 is the first variant.

[puppetcla]

CLA Signed by dalen on 2012-07-17 21:00:00 -0700

[jeffmccune]

I'm trying to use this and I've written a simple fact to try out the new behavior:

Facter.add("jeff") do
  confine :operatingsystem do |os|
    require 'pry'; binding.pry
    true
  end
  setcode do
    "Jeff #{rand(1000)}"
  end
end

It seems as though the block never gets executed. The confine method in Facter::Util::Resolution#confine which is the receiver of my confine :operatingsystem message, doesn't accept a block...

Could you give me an example of how you're using these block based confinements?

Unfortunately, I don't think we're going to be able to get this into 1.7.0, sorry... I'll look at cleaning it up in the next hour, but I'm going to have to switch back to working against a deadline task if my refactoring goes beyond that.

-Jeff

[adrienthebo]

@dalen it's been a week since this has seen activity, and it looks like this PR needs to be rebased. Do you know what the status of this PR is?

[dalen]

It is just waiting for review the last week. Jeff had some thoughts about how the syntax should be and then I've implemented that.
I can rebase, but it just needs a review now.

[jeffmccune]

summary: Merged into master as b7f75cf. This should be released in Facter 2. Thanks again for the contribution! Please note, I re-wrote the Facter::Util::Resolution#confine API documentation in commit 49e3f85

The updated documentation provides specific documentation for each overloaded form of #confine, as seen below.

-Jeff