The book says that when using a case statement "the logic of == is assumed". I'm pretty sure the === (triple equals, not double) is used because you can test the class of an object. For example, the following returns true:

case 65
when Fixnum: true
else false
end

which is only possible if === is used.

Note from the Author or Editor:
This information submitted is correct according to p. 106 of The Ruby Programming Language by Matz and David Flanagan. However, if a class inherits from Object, == is invoked instead.

From: Adri&#65533;n Mugnolo <adrian@giro54.com>
Date: Fri, 11 Jul 2008 16:10:25 -0300
Subject: Ruby Pocket Reference, examples on pages 45-46

Hi,

I originally sent this email regarding your Learning Ruby title and,
was acknowledged by the author. Now that I have a copy of Ruby Pocket
Reference in my hands I see the same code example listed on two pages.

Mathematically correct and idiomatic Ruby for this example should be
something like:

module Dice
def roll
r1 = rand(6) + 1
r2 = rand(6) + 1
total = r1 + r2
puts "You rolled #{r1} and #{r2} (#{total})."
total
end
end

class Game
include Dice
end

g = Game.new
g.roll

Hope this helps to improve any upcoming edition of both books.

Regards

--
Adri&#65533;n Mugnolo
Giro54

Begin forwarded message:

> From: Adri&#65533;n Mugnolo <adrian@giro54.com>
> Date: December 24, 2007 8:05:11 PM GMT-03:00
> To: bookquestions@oreilly.com
> Subject: Learning Ruby, examples 9-12 and 9-13
>
> Hi,
>
> Just a few lines about some issues with the "Learning Ruby" title,
> examples 9-12 and 9-13.
>
> Looks like the author/editor suspected something about the examples:
> "(...) Admittedly, it may not be the most efficient way to guarantee
> a nonzero result, but it works".
>
> According to Ruby's documentation: "(...) [Kernel#rand] returns a
> pseudorandom integer greater than or equal to zero and less than
> [the parameter] max1".
>
> The examples 9-12 and 9-13 have these lines in common:
>
> r_1 = rand(6); r_2 = rand(6)
> r1 = r_1 > 0 ? r_1 : 1; r2 = r_2 > 0 ? r_2 : 1
> total = r1 + r2
>
> IMO, they present several conceptual mistakes:
>
> 1. The dice model is "loaded" for 1. The frequency for 1 is 1/3 --
> twice the correct 1/6.
>
> 2. The dice model has face 6 missing. Because Kernel#rand will
> never return 6, total can't exceed 10.
>
> 3. Use of unneeded variable pairs with very similar names (r_1 and
> r1, r_2 and r2).
>
> 4. Contrived examples in a book meant for beginners, etc.
>
> Let me suggest a simpler replacement:
>
> r1 = rand(6) + 1 # Kernel#rand returns an integer between 0 and n - 1.
> r2 = rand(6) + 1 # Then, increment returned value by one.
> total = r1 + r2
>
> Other than that, the book is great. I am an long-time fan of
> everything O'Reilly: publications, conferences, etc. I spotted this
> problem while reviewing this title to be used as an introductory
> textbook on Ruby, OOP and general programming.
>
> Hope this helps to improve an upcoming edition.
>
> Best regards
>
> --
> Adri&#65533;n Mugnolo
> Giro54

Mathematically correct and idiomatic Ruby for this example should be
something like:

module Dice
def roll
r1 = rand(6) + 1
r2 = rand(6) + 1
total = r1 + r2
puts "You rolled #{r1} and #{r2} (#{total})."
total
end
end

class Game
include Dice
end

g = Game.new
g.roll

The method name should be each_pair, not each_key.

Note from the Author or Editor:
This is correct. The method each_key (third line from the bottom of 109) should be each_pair.

The method name should be each_value, not each_key.

Note from the Author or Editor:
This is correct.

The method name on the first line of p. 110 should be: each_value.