David Chelimsky: predicate_matchers

predicate_matchers

David — Mon, 30 Apr 2007 20:00:00 -0400

Updated on 5/2/2007

In RSpec-0.8 if you say …

File.should_exist(path)

... the expectation passes if File.exist?(path). Here’s how that should look in RSpec-0.9, with the underscore removed:

File.should exist(path)

Supporting this for any arbitrary predicate would require more method_missing magic than we were willing to stomach, so we added a means of easily declaring methods like this yourself. We’ve supplied #exist out of the box, but you can add your own with a simple declaration.

Here’s how you do this for an individual behaviour:

describe Fish do
  predicate_matchers[:swim] = :can_swim?
  it "should swim" do
    Fish.new.should swim
  end
end

And here’s how you define them globally, so they are available in every example in your suite:

Spec::Runner.configure do |config|
  config.predicate_matchers[:swim] = :can_swim?
end

"predicate_matchers" by David

Thu, 03 May 2007 01:30:35 -0400

I went ahead and changed this. A couple of others had the same reaction you did. It’s in the trunk now and will be released as part of rspec 0.9.2 within the next day or so.

Cheers

"predicate_matchers" by David

Tue, 01 May 2007 08:50:12 -0400

No – this is correct – at least it’s how it works. Aslak and I debated which should be the key and which the value and felt this way made more sense. Perhaps we were wrong!

Any other opinions?

"predicate_matchers" by Tim

Tue, 01 May 2007 00:34:10 -0400

Is that backward or do I misunderstand? predicate_matchers[:can_swim?] = :swim and not: predicate_matchers[:swim] = [:can_swim?]

Is the hash keyed by the word used in the test (“swim”) so that you have to do a “backward” search? Or is it the other way round? Maybe it’s my ruby-ignorance showing, but it sure looks backward.