JRuby

BigDecimal divided by Rational gives nil in --1.9 mode

Details

  • Type: Bug Bug
  • Status: Resolved Resolved
  • Priority: Blocker Blocker
  • Resolution: Fixed
  • Affects Version/s: JRuby 1.6.6, JRuby 1.6.7, JRuby 1.7.0.pre1
  • Fix Version/s: JRuby 1.7.0.pre1
  • Component/s: Core Classes/Modules
  • Labels:
    None
  • Environment:
  • Number of attachments :
    0

Description

Given the code:

require 'date'
e = Date.parse '2907-07-22'
require 'bigdecimal'
b = BigDecimal.new '1.0'
rd = e - Date.today
s = b / rd

s is 3.05798240436925e-06 in MRI and JRuby --1.8 mode and nil in JRuby --1.9 mode.

Is this correct or is it a bug?

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Thomas E Enebo added a comment -

I cannot say what changed between 1.8 and 1.9 in MRI, but it will return an error in 1.9.3:

ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-darwin11.3.0]
snippets/jruby-6556.rb:6:in `/': Rational can't be coerced into BigDecimal (TypeError)
	from snippets/jruby-6556.rb:6:in `<main>'

Assuming that is correct, then we should also be throwing a type error and not nil.

Show
Thomas E Enebo added a comment - I cannot say what changed between 1.8 and 1.9 in MRI, but it will return an error in 1.9.3: ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-darwin11.3.0] snippets/jruby-6556.rb:6:in `/': Rational can't be coerced into BigDecimal (TypeError) from snippets/jruby-6556.rb:6:in `<main>' Assuming that is correct, then we should also be throwing a type error and not nil.
Hide
Permalink
Thomas E Enebo added a comment -

reduced case:

require 'bigdecimal'

a = Rational(1, 2)
b = BigDecimal '1.0'
p b / a
Show
Thomas E Enebo added a comment - reduced case: require 'bigdecimal' a = Rational(1, 2) b = BigDecimal '1.0' p b / a
Hide
Permalink
Thomas E Enebo added a comment -

Fixed in commit 693b9ef on master

Show
Thomas E Enebo added a comment - Fixed in commit 693b9ef on master
Hide
Permalink
Uwe Kubosch added a comment -

Thanks!

Bummer that the difference between two dates is not a usable value for calculations. Having to convert from Rational to Integer/Float seems unnecessary.

Show
Uwe Kubosch added a comment - Thanks! Bummer that the difference between two dates is not a usable value for calculations. Having to convert from Rational to Integer/Float seems unnecessary.
Hide
Permalink
Thomas E Enebo added a comment -

The change seems somewhat intentional on MRI's part, but you might want to raise this issue on MRI's bug tracker or on ruby-core/talk to get a better answer.

Show
Thomas E Enebo added a comment - The change seems somewhat intentional on MRI's part, but you might want to raise this issue on MRI's bug tracker or on ruby-core/talk to get a better answer.

People

  • Assignee:
    Thomas E Enebo
    Reporter:
    Uwe Kubosch
Vote (0)
Watch (1)

Dates

  • Created:
    Updated:
    Resolved: