Errata for Programming Ruby

#29609: This errata is in the section titled "Objects and Attributes". On the top of p. 31, the book says:

"In this example, we named the accessor methods name, artist, and duration. The corresponding instance variables, @name, @artist, and @duration,
will be created automatically."

After asking about that statement on the ruby-forum, it seems to be the general consensus that the statement is false. Here is an example, which I think proves the statement is false:

class Song
attr_reader :name, :artist, :duration
end

song = Song.new
puts song.name
puts song.artist
puts song.duration

vars = song.instance_variables
puts "Here are the instance vars: --->#{vars}<----"

--output:--
nil
nil
nil
Here are the instance vars: ---><----

--7stud

#29299: In the code for the 'find' method of class Array, the word 'size' should be '@size', I believe, since size is presumably an instance variable of the Array class.--Michael Vanier

#29612: In the example at the the top of p. 66, the variable song_file is undefined. It looks like the first line of the example should be something like:

song_file = File.open("songdata.txt")--7stud

#28608: In the code for function show_regexp, the statements
"#{$`}<<#{$&}>>#${$'}" and "no match" should be preceded by
"puts", otherwise no output is produced.--Phillip Ngan

#29640: On p. 96, there is an example that starts with this line:

words[key] ||= [ ]
...

and the text says:

"The first line is equivalent to words[key] = words[key] || [ ]."

That seems obvious--unfortunately it's not true. Here is an example:

h = Hash.new(5)

h["a"] ||= 10
p h #{}

h["a"] = h["a"] || 10
p h #{"a"=>5}

Apparently in Ruby, the statement:

x ||= y

is equivalent to:

x = y unless x

Here is an example:

h = Hash.new(5)

h["a"] ||= 10
p h #{}

h["a"] = 10 unless h["a"]
p h #{}
--7stud

#29679: At the top of page the following code appears:

count = 0
threads = []
10.times do |i|
threads[i] = Thread.new do
sleep(rand(0.1))
Thread.current["mycount"] = count
count += 1
end
end
threads.each {|t| t.join; print t["mycount"], ", " }
puts "count = #{count}"

Since rand(0.1) is the same as rand, "rand(0.1)" should be replaced with either "rand" or "0.1 * rand", whichever was actually intended. This error was detected by someone other than me and posted on ruby-talk ML -- I just thought I should post it here because nobody else has.
--Morton Goldberg

#29348: Table 11.1 purports to demonstrate a race condition this will occur when either thread runs before the previously active thread has a chance to "store val back into @count" resulting in a lost update to count. The table erroneously shows suspensions after the update to count has occurred which will not affect @count.--bob

#29729: p. 156, example at top of page: The output shows the test failing, but there is no description of what was entered to get the test to fail. I tried specifying an empty data file on the command line, a file with two words in it, and a file that doesn't exist. Still the test succeeded. Then I omitted specifying a data file name on the command line, and instead I entered ^D for end-of-file when the program was running. Still the test succeeded.

Finally, I specified a file on the command line, and then I inserted the statement:

ARGF.read

before the line:

assert_not_nil(ARGF.read, "Read next...")

and that caused the test to fail. I also tried that modified code without specifying a file on the command line and entering ^D when the program was running, but the test succeeded, which I don't understand.

I asked about this on the ruby-forum, and no one knew how to make the test fail.
--7stud

#29962: The first time I tried to use the debugger was in 186 and I am seeing something different than in v2 of the book on page 163. Or perhaps I am doing something wrong.

In the book it appears that when you run with the debugger option, such as this

ruby.exe -r debug t.rbyou end up with the debugger looking at the source file (e.g., t.rb).

However, when I run it I end up in some Ruby module:

ruby.exe -r debug t.rb
C:/ruby/lib/ruby/site_ruby/1.8/ubygems.rb:10:require ‘rubygems’
(rdb:1) l
[5, 14] in C:/ruby/lib/ruby/site_ruby/1.8/ubygems.rb 5 # All rights reserved. 6 # See LICENSE.txt for permissions. 7 #++ 8 9
=> 10 require ‘rubygems’

The work-around seems pretty simple: set a breakpoint in your own file and type C, as below.

I thought you might want to know, in order to update the book.

(rdb:1) b t.rb:3
Set breakpoint 1 at t.rb:3
(rdb:1) c
Breakpoint 1, toplevel at t.rb:3
--Wes Rishel

#29754: p. 216, Installing Ruby Gems: the two commands at the top of the page read:

% cd rubygems-0.7.0
% ruby install.rb

However, according to the rubygems README file, which after I unpacked the files was in the directory

rubygems-0.9.4/

(0.9.4 is the current version number), the command to install rubygems is:

ruby setup.rb

Since there there was no file called install.rb in the directory rubygems-0.9.4/ and there was a file called setup.rb in the rubygems-0.9.4/ directory, I went with the command in README:

$ ruby setup.rb

However, I got an error mid way through the install:

...
...
mkdir -p /usr/bin/
install gem /usr/bin/
setup.rb:633:in `initialize': Permission denied - /usr/bin/gem (Errno::EACCES)
from setup.rb:633:in `open'
from setup.rb:633:in `install'
from setup.rb:1377:in `install_files'
from setup.rb:1376:in `each'
from setup.rb:1376:in `install_files'
from setup.rb:1346:in `install_dir_bin'
from setup.rb:1532:in `__send__'
from setup.rb:1532:in `traverse'
... 6 levels...
from setup.rb:1000:in `exec_install'
from setup.rb:814:in `invoke'
from setup.rb:773:in `invoke'
from setup.rb:1578

So, I reran the command like this:

$ sudo ruby setup.rb
password: <I entered my password>

and everything installed fine.
--7stud

#29759: p. 217, The command at the top of the page is:

% gem install -r rake

But, when entered that command, first I got this:

$ gem install -r rake
Bulk updating Gem source index for: <link here>
ERROR: While executing gem ... (Gem::GemNotFoundException)
Could not find rake (> 0) in any repository

So I tried the command again, and I got:

$ gem install -r rake
Bulk updating Gem source index for: <link here>
ERROR: While executing gem ... (Errno::EACCES)
Permission denied - /usr/lib/ruby/gems/1.8/cache/rake-0.7.3.gem

The command I needed to use was:

$ sudo gem install -r rake
Password: <I entered my password here>
Bulk updating Gem source index for: <link here>
Successfully installed rake-0.7.3
Installing ri documentation for rake-0.7.3...
Installing RDoc documentation for rake-0.7.3...

--7stud

#29776: p. 222: The text says:

"As of RubyGems 0.8.0, requiring rubygems.rb will install an overloaded version fo Ruby's require method. Having loaded the RubyGems framework, you could say:

require 'bluecloth' "

Ok, so according to the text there are two prerequisites to using that require statement:

1) Installation of RubyGems 0.8.0 or higher

2) Requiring rubygems.rb, i.e. the statement:

require 'rubygems'

So, if you have RubyGems 0.8.0 installed, then you have to write:

require 'rubygems'
require 'bluecloth'

How does that free your program from RubyGems-specific code?

After the explanation of the first example on the page, the rest of the section should be deleted since it is completely non-sensical.
--7stud

#29635: Running the test ruby script for the my_test extension fails on OSX with:

1) Error:
test_test(TestTest):
NameError: uninitialized constant TestTest::MyTest
method test_test in my_test.rb at line 7

Seems it cannot find the MyTest class even though the extension loaded with require 'my_test'

--Sean Wolfe

#30477: --with-name-lib=directory
Reads: "Add directory/lib to the link command"
Should read: "Add directory to the link command"

Similar error applies to
--with-name-include=directory--Matt Wallis

#25968: I believe there is a rule missing in the "Parallel Assignment" section. The last sentence on p. 91 states, "If an assignment has just one lvalue and multiple rvalues, the rvalues are converted to an array and assigned to the lvalue." This seems to contradict (or at least) modify rule 5 on p. 340.--Brian Adkins odjfs2001@yahoo.com

#25969: I think some clarification of "Ranges in Boolean Expressions" is in order :) The last sentence in the first paragraph doesn't agree with the last two sentences in the second paragraph. Also, in the first example, I would expect the slot for 18 to be nil since expr2 evaluates to true which should cause a transition to unset according to the text.--Brian Adkins odjfs2001 at yahoo dot com

#29738: p. 352, Aliases: At the top of the page, the first two lines say:

"When a method is aliased, the new name refers to a copy of the original method's body. If the method is subsequently redefined, the aliased name will still invoke the original implementation."

If that were true, then aliasing a name wouldn't do anything: the new name would be a copy of the original method's body and the aliased name would "still invoke the original implementation." As a result, you would have two names that invoke identical methods.

The second line should read: "If the method is subsequently redefined, the new name will invoke the original implementation." For proof, see the example under the text.--7stud

#30483: Section "to_hash -> Hash", first line: "(The only known use is the second parameter to Hash#replace.)" should be "(The only known use is the parameter to Hash#replace.)", since Hash#replace has no second parameter.--Felix Siegrist

#26809: In the ex0629.rb and ex0629.rb program listings this line should read:

if Fixnum === other && (other + @value) <= MAX_ROMAN

Otherwise, an miss-by-one (less) condition occurs not allowing the result of the +(other) method to include MAX_ROMAN.

Thanks for your great book!

P.S. couldn't find the book version in my printed copy -- sorry.
--Adrian Mugnolo

#29911: Regarding:

"At the top-level, we're executing code in the context of some
predefined object. When we define methods, we're actually creating
(private) instance methods for the class Object."

The implication is that the second sentence somehow follows from the
first. But it doesn't. In the context of the first sentence, the
behavior in the second sentence is magical.--Greg Weeks

#29692: In the library reference, in the Kernel module section, it says that Kernel module instance methods are listed in the Object class page, and instead in thid Kernel module page are listed only Kernel module methods as for example the method "puts" among others.... The fact is that "puts" (and other moethods listed here) isn't a module method, it is a private module instance method.--Skyblaze

#26538: Both to_a and to_s show the following statement:

h = { "c" => 300, "a" => 100, "d" => 400, "c" => 300 }

I believe one of the "c" => 300 key value pairs is redundant. It's not technically an error, but somewhat nonsensical.--Brian Adkins

#29913: "The Kernel module is included by class Object, so its methods are
available in every Ruby object." Here, "methods" should read "instance
methods". It is not clear why Kernel class methods can be called
without a receiver. It may be magical.

--Greg Weeks

#26313: The Math.sqrt function does not raise an "ArgError" (or an ArgumentError, if that's what you meant) if 'numeric' is less than 0. On both Linux and Windows it raises Errno::EDOM.

-Dan--Daniel Berger

#29441: You state in the documentation for the []= method 'the Regexp and String forms will silently ignore the assignment [if the value is out of range].' But they both raise an IndexError if there's no match.

Furthermore theres no '~' method for Strings.

(Either in 1.8.2 or 1.8.5)

#29820: gsub() documentation on p.613: Above the description is the syntax:

str.gsub(pattern, replacement) --string
str.gsub( pattern ){|match| block} ---> string

In the description is this passage:

"...the sequences \1, \2, and so on may be used to interpolate successive groups in the match."

I wonder who can make any sense of that statement. 'match' is the block syntax's parameter, which the description is not even referring to, so mentioning 'match' is confusing.

The sentence should say something like:

"...the sequences \1, \2, and so on may be used in <i>replacement</i> to refer to the actual matches to parenthesized groups."

Or, simply:

"...the sequences \1, \2, etc. may be used in <i>replacement</i>.

(The sequences 1\, \2, etc. are described on p. 75 for any trying to figure out what they mean.)

Also, it's worth highlighting that double quotes surrounding the replacement string prevent the sequences \1, \2, etc. from working. Instead, you have to use single quotes around the replacement string to get the sequences \1, \2, etc. to work. Since I'm told it that is a very common error, it should be in the docs that you need to use SINGLE QUOTES.

--7stud

#30502: Near the bottom of the page:
users.join(", ") should be users=users.join(", ")
groups.join(", ") should be groups=groups.join(", ")

My book's copyright page says "Eleventh Printing, October 2007"
"Version: 2007-9-13"--Wayne Kovsky

#26867: The first example of a weakref doesn't cause an exception. The text suggests that this is being demonstrated, but even the "produces" output shown has no exception. It seems that strings aren't being garbage collected at this time. Substituting Object.new for "fol de rol" in the 2nd line produces the expected exception after the gc runs.--Stonewall Ballard

#29557: The code listing for unpack_sockaddr_un on page 769 in the Second Edition is incorrect. It is a copy of the unpack_sockadd_in code listing. The explanation is OK.--Michael