Proof of concept for yield_value rspec matcher

yield_value_spec.rb

RSpec::Matchers.define :yield_value do |expected_yielded_value|
  match do |block|
    yielded_value = nil
    block.call(lambda { |arg| yielded_value = arg })
    yielded_value == expected_yielded_value
  end
end

describe "yield_value matcher" do
  it 'passes a valid positive expectation' do
    expect { |b| 3.tap(&b) }.to yield_value(3)
  end

  it 'fails an invalid positive expectation' do
    expect {
      expect { |b| 3.tap(&b) }.to yield_value(2)
    }.to raise_error(RSpec::Expectations::ExpectationNotMetError)
  end
end

BTW, on the topic of the naming of the matchers, I think I like yield_with_args better than yield_args. It has a nice symmetry to yield_without_args, and it's more closely aligned with how I talk about this stuff: "3.tap yields with 3 as the argument".

I'll start taking a stab at this, and open a pull request with what I come up with.

FWIW, I've pushed what I've got so far into a branch:

https://github.com/rspec/rspec-expectations/tree/yield_matchers

There's more I plan to do before submitting a pull request, but I figured I'd push something to back it up and so people who care to can take a look.