Use require_relative inplace of standard require if available.

[ileitch]

As per discussion on the mailing list: http://rubyforge.org/pipermail/rspec-users/2011-November/020760.html

Possibly the simplest approach I could think of. Provided this is OK, it could probably be moved into its own file.

[dchelimsky]

@ileitch - this seems good to me with defined? rather than Kernel.respond_to? (feels cleaner, and turns out to shave off a millionth of a second to boot!).

Once concern I have is that we're doing this in the global namespace. What do you think about something like:

module RSpec
  if defined?(require_relative)
    def _require(path)
      require_relative path
    end
  else
    def _require(path)
      require "rspec/#{path}"
    end
  end
  module_function :_require
end

RSpec._require 'core/filter_manager'
# etc

I'd want the _ prefix or similar to identify it for readers as internal to rspec, and avoid collisions with Kernel methods.

As we add this to the other libs, we'd do something more like:

module RSpec
  unless defined?(_require)
    if defined?(require_relative)
      def _require(path)
        require_relative path
      end
    else
      def _require(path)
        require "rspec/#{path}"
      end
    end
  end
  module_function :_require
end

That way we wouldn't be redefining it as each lib got loaded.

Thoughts?

[ileitch]

namespacing and defined? both look sound to me.

[dchelimsky]

Let's do it. So we need to make this change in rspec-core and then add the whole thing to the other rspec libs. Would you like to do any of that?