Bug #18677: BigDecimal#power (**) returns FloatDomainError when passing an infinite parameter

• Status: Assigned
• Assignee: mrkn (Kenta Murata)
• Target version: -
• ruby -v: ruby 3.1.1p18 (2022-02-18 revision 53f5fc4236) [x86_64-darwin21]
• Backport: 2.6: UNKNOWN, 2.7: UNKNOWN, 3.0: UNKNOWN, 3.1: UNKNOWN

[dorianmariefr (Dorian Marié)]

Example:

> BigDecimal(10) ** BigDecimal("Infinity")
FloatDomainError: Computation results in 'Infinity'

Maybe:

require "bigdecimal/util"

class BigDecimal < Numeric
  def **(other)
    if other.infinite? == 1
      if self > 1
        BigDecimal::INFINITY
      elsif self == 1
        self
      elsif self >= 0
        BigDecimal(0)
      else
        power(other)
      end
    else
      power(other)
    end
  end
end

def puts_and_eval(string)
  puts string
  p eval(string)
end

puts_and_eval "10 ** BigDecimal::INFINITY"
puts_and_eval "1 ** BigDecimal::INFINITY"
puts_and_eval "0.1 ** BigDecimal::INFINITY"
puts_and_eval "0 ** BigDecimal::INFINITY"
puts_and_eval "-1 ** BigDecimal::INFINITY"

Seems like ruby is doing very different things from math though

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Assigned
• Assignee set to mrkn (Kenta Murata)

[mrkn (Kenta Murata)]

I made a patch at https://github.com/ruby/bigdecimal/pull/227.

Note that the last case -1 ** BigDecimal::INFINITY should cause Math::DomainError.