Bug #5309: 0.6.to_r != "0.6".to_r

• Status: Rejected
• Assignee: -
• Target version: -
• ruby -v: -

[brixen (Brian Shirai)]

0.6.to_r != "0.6".to_r

Is this a bug?

$ ruby -v
ruby 1.9.2p290 (2011-07-09 revision 32553) [x86_64-darwin10.8.0]

$ irb
ruby-1.9.2-p290 :001 > 0.6.to_r
=> (5404319552844595/9007199254740992)
ruby-1.9.2-p290 :002 > "0.6".to_r
=> (3/5)
ruby-1.9.2-p290 :003 > a = "%.55f" % 0.6
=> "0.5999999999999999777955395074968691915273666381835937500"
ruby-1.9.2-p290 :004 > b = "%.55f" % "0.6".to_f
=> "0.5999999999999999777955395074968691915273666381835937500"
ruby-1.9.2-p290 :005 > a == b
=> true

Thanks,
Brian

[naruse (Yui NARUSE)]

• Status changed from Open to Rejected

0.6.to_r is considered as a rational converted from a floating point number 0.5999999999999999777955395074968691915273666381835937500.
"0.6".to_r is considered as Rational(6, 10).
So they are slightly different number.

[marcandre (Marc-Andre Lafortune)]

Use Float#rationalize instead of to_r, and you will get the same result as String#to_r. I've argued before that Float#to_r should be Float#rationalize, though.

On Sep 12, 2011, at 2:06 PM, Brian Ford wrote:

0.6.to_r != "0.6".to_r

Is this a bug?

Probably not because

0.6.to_r

requires two conversions:

decimal floating point to binary floating point to rational

while

"0.6".to_r

only requires one conversion from decimal floating point to rational.

In the first case, the intermediate value is not exactly 3/5 so the
two sequences don't give you the same result.

Gary Wright

[naruse (Yui NARUSE)]

• ruby -v changed from ruby 1.9.2p290 (2011-07-09 revision 32553) [x86_64-darwin10.8.0] to -

2011/9/13 Marc-Andre Lafortune [e-mail]:

Use Float#rationalize instead of to_r, and you will get the same result as String#to_r.
I've argued before that Float#to_r should be Float#rationalize, though.

It is http://redmine.ruby-lang.org/issues/1408

--
NARUSE, Yui  [e-mail]