Feature #12039

Fixnum#infinite?/Bignum#infinite or Numeric#infinte, consistent with Float#infinite? and BigDecimal#infinite?

Added by ioquatix (Samuel Williams) almost 5 years ago. Updated over 4 years ago.

Target version:


We have an issue where it is not easy to ask if a number is +ve or -ve infinity.

Both Float and BigDecimal do respond to infinite? but Integer derivatives don't. It sort of makes sense, since we don't have an infinity value for Fixnum/Bignum. However, it makes polymorphic code hard.

[13] pry(main)> Float::INFINITY.infinite?                                       
=> 1
[14] pry(main)> -Float::INFINITY.infinite?                                      
=> -1
[15] pry(main)> BigDecimal::INFINITY.infinite?                                  
=> 1
[16] pry(main)> -BigDecimal::INFINITY.infinite?                                 
=> -1

Given a Numeric value x, it would be nice to query if it is INFINITY or not. Propose adding a default implementation to Numeric, which always returns false or nil.

Related issues

Related to Ruby master - Feature #10641: Introduce Fixnum#finite? and Bignum#finite?ClosedActions

Updated by ioquatix (Samuel Williams) almost 5 years ago

It might also make sense to consider if other "predicates" make sense, e.g. nan? They should be declared on Numeric with a default logical implementation where possible.

Updated by seantheprogrammer (Sean Griffin) almost 5 years ago

This is probably a useful method to be able to rely on universally, even if Fixnum and Bignum just return false.

Updated by shyouhei (Shyouhei Urabe) almost 5 years ago

Another real-world use-case where I found Numeric#infinite? useful is JSON validation. Infinities are not allowed in JSON so a programmer want to filter them out before converting a Numeric into JSON, but that is not straight-forward right now. With this method it becomes much OO-ish.

Updated by matz (Yukihiro Matsumoto) almost 5 years ago

Agreed. Add #finite? and #infinite?


Updated by nobu (Nobuyoshi Nakada) almost 5 years ago

  • Tracker changed from Bug to Feature

Updated by mrkn (Kenta Murata) almost 5 years ago

What behavior is desirable for a Complex?

Updated by mrkn (Kenta Murata) almost 5 years ago

I think the following definition is acceptable for a Complex:

def inifinite?

Updated by ioquatix (Samuel Williams) over 4 years ago

This looks good, do you think we can integrate this at some point? What do we need to do?

Updated by mrkn (Kenta Murata) over 4 years ago

  • Assignee set to mrkn (Kenta Murata)

Updated by mrkn (Kenta Murata) over 4 years ago

  • Status changed from Open to Closed

Applied in changeset r55702.

numeric.c, complex.c: Add finite? and infinite? consistent with Float

  • numeric.c (num_finite_p, num_infinite_p): Add Numeric#finite? and
    Numeric#infinite? [Feature #12039] [ruby-core:73618]

  • complex.c (rb_complex_finite_p): Add Complex#finite?

  • complex.c (rb_complex_infinite_p): Add Complex#infinite?

  • test/ruby/test_bignum.rb: Add test for Integer#finite? and

  • test/ruby/test_fixnum.rb: ditto.

  • test/ruby/test_rational.rb: Add test for Rational#finite? and

  • test/ruby/test_complex.rb: Add test for Complex#finite? and

Updated by ioquatix (Samuel Williams) over 4 years ago

Thank you for your hard work :)


Updated by shyouhei (Shyouhei Urabe) about 4 years ago

  • Related to Feature #10641: Introduce Fixnum#finite? and Bignum#finite? added

Also available in: Atom PDF