Why aren't Enumerators in Ruby (2.0 +) lazy by default?

Question

Why don't Ruby Enumerator behave like Enumerator::Lazy by default?

Is there any case where someone would want use a non-lazy Enumerator?

Edited:

A comment to backward compatibility answers below, explaining why I'm not yet convinced:

Say we had added these 'breaking' changes to Ruby 2.0.0, which is a major version, you would test your code (especially if you're going to production) thoroughly before making the switch, no?

Edit#2

I suspected it has to do with efficiency (let me know if anything is wrong), so I did the following benchmark: (of course there are places where lazy is better. this is to demonstrate, potentially, why Ruby is not using lazy all the time?)

require 'fruity'
require 'prime'

compare do
  lazy { 
    g = Prime::EratosthenesGenerator.new; 1000.times { g.lazy.take(100).to_a } 
  }
  diligent { 
    g = Prime::EratosthenesGenerator.new; 1000.times { g.take(100).to_a } 
  }
end

I get the following: (maybe due to the calling of .lazy method instead of using Enumerator::Lazy from the beginning?)

diligent is faster than lazy by 19.999999999999996% ± 1.0%

Answer 1

Make laziness the default would be a breaking change for any code that expects/relies on side-effects.

Any previous usage of Enumerator would need to be examined and tested. The "safest"/"sane" approach is simply to keep the non-lazy behavior as the default - while providing a Lazy alternative.

Some fundamental/previous design decisions are not so easily discarded; Enumerators are supported as far back as Ruby 1.8.7.

Answer 2

Probably for compatibility. Lazy enumerators were introduced recently, and many code depend on enumerators being non-lazy.