Bug #5713: Fixnum#** returns Infinity for 0 ** negative Bignum

• Status: Closed
• Assignee: marcandre (Marc-Andre Lafortune)
• Target version: 2.0.0
• ruby -v: ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-darwin10.8.0]

[john_firebaugh (John Firebaugh)]

=begin
Instead it should raise ZeroDivisionError, the same as negative Fixnums.

wordsize = 8 * 1.size
fixnum_min = -2 ** (wordsize - 2)

def zero_power(exp)
0 ** exp
rescue ZeroDivisionError
"ZeroDivisionError"
end

[-1, fixnum_min, (fixnum_min-1)].each {|i| puts zero_power(i)}
=end

[marcandre (Marc-Andre Lafortune)]

• Assignee set to marcandre (Marc-Andre Lafortune)

[john_firebaugh (John Firebaugh)]

The execution path of 0 ** -Bignum goes to Rational(0) ** -Bignum, so I think the issue is there. I.e. Rational(0) ** -Bignum should raise ZeroDivisionError, the same as Rational(0) ** -Fixnum.

[marcandre (Marc-Andre Lafortune)]

• Category set to core
• Target version set to 2.0.0

John Firebaugh wrote:

The execution path of 0 ** -Bignum goes to Rational(0) ** -Bignum, so I think the issue is there. I.e. Rational(0) ** -Bignum should raise ZeroDivisionError, the same as Rational(0) ** -Fixnum.

Yes, unless there is objection, Rational#** should treat 0 and 1 as special cases before resorting to conversion to float, i.e.

Rational(0) ** (-2**100) # => Infinity, should raise an Error
Rational(0) ** (2**100) # => 0.0, should be Rational(0)
Rational(1) ** (2**100) # => 1.0, should be Rational(1)

[john_firebaugh (John Firebaugh)]

FYI, I've been doing RubySpec work on this in Rubinius: https://github.com/rubinius/rubinius/commits/master/spec/ruby/shared/rational/exponent.rb

[shyouhei (Shyouhei Urabe)]

• Status changed from Open to Assigned

[marcandre (Marc-Andre Lafortune)]

• Status changed from Assigned to Closed
• % Done changed from 0 to 100

This issue was solved with changeset r39063.
John, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.

---

• rational.c (nurat_expt): Deal with special cases for rationals 0, ±1
  [bug #5713] [bug #5715]