Errata for Agile Web Development with Rails, 1st Edition

#24522: "Sometimes these class-level methods return collections of objects.
Order.find(:all, :conditions => "name='dave'") do |order| 
puts order.amount
end"

Find returns a collection but requires an iterator to call the block on it. So I think you're missing an ".each" before the "do".--Kieron Browne

#24533: The code at the bottom of the page is unclear. What should one do with it? Type it at an 'mysql' prompt? If so, the result is an error (1046) that states that no database is selected.

#24116: It seems that the rails script now adds the application to database.yml:

database: depot_development

It automatically replaces the rails_development with depot_development.
--John Crowhurst

#24500: Repeated reinstallations of entire package on OSX 10.4.6 continually produce the following:

./script/../config/../config/environment.rb:8: warning: already initialized constant RAILS_GEM_VERSION
exists app/controllers/
exists app/helpers/
exists app/views/admin
exists test/functional/
dependency model
exists app/models/
exists test/unit/
exists test/fixtures/
identical app/models/product.rb
identical test/unit/product_test.rb
identical test/fixtures/products.yml
Unknown database 'depot_development'

Error is not documented in footnotes and I can't find anything on it an forums. Any ideas?

#24510: for,
depot> ruby script/generate scaffold Product Admin

I had to use infor from http://dev.rubyonrails.org/ticket/200 to get it to work on Linux (gentoo distro).
--Mike Nelson

#24532: When attempting to run the line "ruby script/generate scaffold Products Admin" in Windows, I receive the following error listed in the book (addressed only for Mac): Before updating scaffolding from new DB schema, try creating a table for your model (Product).

#24556: On OS X (10.4.3), when I attempt to run script/generate scaffold Product Admin, I get the following error:

You have a nil object when you didn't expect it!
You might have expected an instance of Array.
The error occured while evaluating nil.each

Not sure how to proceed. I have been on the IRC channels looking for help but to no avail. Any ideas?

#26704: ppc:~/work/depot tomcoady$ ruby script/generate scaffold Product Admin
./script/../config/boot.rb:18:in `require': No such file to load -- rubygems (LoadError)
from ./script/../config/boot.rb:18
from script/generate:2:in `require'
from script/generate:2--Tom Coady

#25428: After running "ruby script/generate scaffold Product Admin" on linux box using SuSE v10.0 and MySql Ver 14.12 Distrib 5.0.22, I get an error stating /tmp/mysql.sock cannot be found. I resolved this by searching for mysql.sock and creating soft link to it from /tmp dir. e.g. ln -s /var/lib/mysql/mysql.sock tmp/mysql.sock I was then able to get form depicted on p.58--Mark Glass

#31649: My version of the paper book is actually P5 (2006-02-16). The problem is that scaffold has changed. I am running Rails 2.02 and the following command will work.

ruby script/generate scaffold Product id:int title:varchar description:text image_url:varchar price:decimal -f

So, in answer the the question "That wasn't hard now, was it?" is actually "it was a nightmare"!!--Antony Scott

#24784: I've seen this error elsewhere but never a solution. Here goes: On page 57, I entered, at depot prompt,
ruby script/generate scaffold Product Admin
The error I get is:
"#28000Access denied for user 'root'@'localhost' (using password: NO)"

The author's statement after the description of this is: "That wasn't hard now, was it"...actually, it's turning out to be very hard.

I had to battle to get the commands going on page 54 too.

It really shouldn't be this hard...

--pat lynch

#24333: The unless clause of the validate method is wrong. To verify this, comment out 'validates_numericality_of :price' and try submitting a product without a price; you'll get a null constraint violation. Something like 'unless price && price >= 0.01' works better.--Kevin Christen

#25300: In the eBook we have this code for product.rb:
errors.add(:price, "should be positive") unless price.nil? || price >= 0.01
Where as the downloadable code has the following:
errors.add(:price, "should be positive") unless price.nil? || price > 0.0

Notice that the last line compares "price > 0.0" instead of "price > 0.01"

--Max

#24557: Cannot "destroy" items.

Both file listings 66 & 67 omit ':post' argument in the link_to 'Destroy' line. The line should read:

<%= link_to 'Destroy', { :action => 'destroy', :id => product },
:confirm => "Are you sure?",
:post => true %>
--Randy W. Sims

#24254: Relative URL's not allowed by current validates_format_of :with parameter; where the text says: "Put some images in the public/images directory and enter some product descriptions, and the resulting product listing might look something like Figure 6.7, on the next page.", the current validation functionality in the model (/depot/app/models/product.rb) will not allow the correct relative url (i.e. /images/sk_utc_small.jpg) to be entered, since it requires a full URL.
In the downloadable code for depot4/app/models/product.rb, the regular expression should be commented out or deleted, and replaced with one allowing relative url's for the images.
And this should be among the modifications mentioned in the text for Iteration A4.--Victor Kane

#25734: The list action for Destroy as it is in the book didn't work for me: Clicking Destroy would pop up the confirmation dialog, but choosing ok would just refresh the list but not delete the product. After looking at the initially generated scaffolding code, I've noticed that this included an additional parameter

:post => true

After I changed the whole line to

<%= link_to 'Destroy', { :action => 'destroy', :id => product },
:confirm => "Are you sure?", :post => true %>

the destroy action worked as expected.

PS: It might be a version issue. I've observed it with Rails 1.1.6
PPS: I have the 5th printing, February 06. This not available from the version drop down box.--Stefan Moser

#24228: The text starting: 'put some images in the public/images directory' led me to think that the correct url for the images in the production form was /public/images/xxx.jpg Actually, after some considerable head scratching, it turns out the base for static resources is public, so the correct href is /images/xxx.jpg. The text gives no inkling of this..--Michael Karliner

#26711: The suggested code does not work for me but I found the right code elsewhere in the code bundle - detail here to avoid unititiated Product: http://groups.google.com/group/rubyonrails-talk/browse_thread/thread/6b7333e92877c94d/8cd2f0bde6dd1a3f--Tom Coady

#24091: As was previously stated, " In the implementaion of find_cart, the 3rd line reads "session[:cart] ||=Cart.new" This line should read "@cart = session[:cart] ||= Cart.new" " I ran into an issue in which if I accessed /store or /store/display_cart directly in a fresh (meaning session cookie-less) browser window, I could refresh the page as many times as I want and keep the same session, however if I added an item to the cart or otherwise changed pages in the browser, it would set a new cookie everytime (even everytime I refreshed the new page). Making the change noted in the quote above fixed the issue. I'm running Ruby 1.8.2, Rails 1.0.0 on a Win XP Pro SP2 pc, and I caught the error by adding the line <%= debug session %> to both my index.rhtml and display_cart.rhtml (in store views).

#25164: As pointed out before, there is an error in
session[:cart] ||= Cart.new
I replaced the code with
session[@cart] ||= Cart.new
and it magically worked again

Need someone to explain this to me

#25862: I made all the additions to the code to the end of making the cart

I get the rails error:

"You have a nil object when you didn't expect it!
You might have expected an instance of Array.
The error occured while evaluating nil.size"

I don't know how to fix this--Gerry

#24329: IWhen I run the sql at the bottom of the page to create the new table, I get an error "ERROR 1005 at line 13: Can't create table '.\depot_development\line_items.frm' <errno: 150>"

MySQL (I'm running 4.0.16) needs an index to be declared on the foreign key can be created. Like so:
create table line_items (
id int not null auto_increment,
product_id int not null,
quantity int not null default 0,
unit_price decimal(10,2) not null,
primary key (id),
index ix_line_items (product_id),
constraint fk_items_product foreign key (product_id) references products(id)
) TYPE=InnoDB;

Also MyIsam is the default table type, so I've forced innodb for the foreign key support. Not sure if that would have affected anything though.--Nick Coyne

#25904: similar to the error already reported...

ERROR 1064 (42000) at line 12: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near '
product_id int not null,
quantity int not null default 0,
unit_price decimal' at line 2

(mysql 5.0.21)

typing the code as is on the bottom of 81, mySQL is kicking back this error. help?--dave rupert

#24395: Others have mentioned this error and I've tried all their suggestions but I can't get any to work. Upon testing the cart to see if it will return the @item.size of the cart's contents I get an error of:

NoMethodError in Store#display_cart

Showing app/views/store/display_cart.rhtml where line #2 raised:

You have a nil object when you didn't expect it!
You might have expected an instance of Array.
The error occured while evaluating nil.size

Now what? I've rebooted the WebBrick server. Killed the browser window after deleting all the cookies, session data, etc. I'm using RadRails so I closed it as well. Booted the server back up from a command line and get the same option. I've searched the internet through and tons of users have had this problem but there are no solutions.

Please, oh please, help!--Philip Joyner

#24393: The online version of the file for '...app/models/cart.rb' (http://media.pragprog.com/titles/rails/code/depot9/app/models/cart.rb) differs in the fourth last line from the class definition in the book.

Book (pp 83-4):
...
@total_price += item.unit_price
...

Downloadable file 'cart.rb':
...
@total_price += product.price
...

The dowloadable file seems to be correct.--Gus Gollings

#24422: In reference to problem #2532, "You have a nil object when you didn't expect it!", I also ran into the problem in chapter 8. I typoed a simple part of the application, tested it, fixed the typo, then got that error no matter what I tried (even deleting everything and typing it up again). The answer was indeed to delete session data: depot/tmp/sessions--April

#26229: When looking at "depot/public/store/display_cart" with my browser I was getting the error:

undefined local variable or method `find_cart' for #<StoreController:0x4080b768>

I then checked the downloadable Code snippets and noticed the find_cart method defined in File 75. I added this method to my code (as I was typing it from the book) and things worked.

It should say in the book to add the find_cart method (it doesn't).

(note: the version I'm using is:
Third Printing, September 2005
Version: 2005-9-13
(the closest option to submit this erratum was P3.0 - September 29))--Sean Lerner

#24256: Re #2532 - This also happend to me running under Webrick. The solution was to both close the browser and restart Webrick. Hope this helps the other reader.--brian

#29669: Concerning the invalid IDs of objects in the URL line: it is not certain that an exception will always be thrown. I assume that the ID field, extracted from the URL, is transformed to a number without any checks made. So if there is a product with ID=0 in the database, it is actually added in the cart! This is way worse than crackers looking at an exception report page! One solution is to make sure somehow that the database starts counting from 1, but this doesn't make me feel secure enough. I believe that there should be further error checks on the ID parsing level.--Nikos Mouchtaris

#24448: In regard to the common error
"NoMethodError in Store#display_cart
Showing app/views/store/display_cart.rhtml where line #5 raised:

You have a nil object when you didn't expect it!
You might have expected an instance of Array.
The error occured while evaluating nil.each
"
I found that I had a typo in the store_controler.rb
I had entered.
---
def display_cart
@cart = find_cart
@items = @cart_items
end
---

but correct code is
---
def display_cart
@cart = find_cart
@items = @cart.items
end
---

Oh the joys of being a newbie!!! 5 hours of head scratching for that one!--glenn

#24860: Unable to create the session table.

C:\apps\ruby\files\work\depot>rake --trace db:session:create
(in C:/apps/ruby/files/work/depot)
rake aborted!
Don't know how to build task 'db:session:create'
c:/apps/ruby/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:1449:in `[]'
c:/apps/ruby/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:455:in `[]'
c:/apps/ruby/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:1906:in `run'
c:/apps/ruby/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:1906:in `run'
c:/apps/ruby/lib/ruby/gems/1.8/gems/rake-0.7.1/bin/rake:7
c:/apps/ruby/bin/rake.bat:25
--Dave Pawson

#25138: Flash messages are not displaying correctly using IE 6 sp2.

Results of invalid add_to_cart, empty_cart, etc no flash message appears to display.

Resizing IE window with mouse, flash message appears & disappears.

Resizing IE window with F11 key yields flash menu with missing bottom red line of the flash box.

Hitting key F11 to maximize IE, flash message displays --Steven

#23580: using the fmt_dollars(item.unit_price) method, values less than 1, (values < 1) on the display_cart details show up as 0.00 but get added correctly to the total.--Jeff Bell

#26216: @order.line_items << @cart.items has to be @order.LineItems << @cart.items otherwise, ruby will report an unknown method "line_items" for the order object.--Volker

#24213: The code in checkout.rhtml that displays the error message is given as:

<%= error_messages_for(:order) %>

This works fine if there is an error present. However, when initially connecting to the page I get the error:

-------------------------------------------------
NoMethodError in Store#check-out

Showing app/views/store/check-out.rhtml where line #2 raised:

You have a nil object when you didn't expect it!
You might have expected an instance of ActiveRecord::Base.
The error occured while evaluating nil.errors

Extracted source (around line #2):

1: <% @page_title = "Checkout" -%>
2: <%= error_messages_for(:order) %>
3: <%= start_form_tag(:action => "save_order") %>
4: <table>
5: <tr>

RAILS_ROOT: c:/web/depot/public/../config/..
-------------------------------------------------
It seems that on initial access a nil object is passed to error_messages_for.

However, if I change the line to:

<%= error_messages_for(:order) if @order %>

so that the presence of an object "order" is checked before passing anything to error_messages_for, the page works on opening from new and still displays the errors correctly.--Rob Nichols

#23894: Adding the scaffold stylesheet to the store layout seems to disable any formatting from the depot stylesheet. It doesn't seem to matter which I list first--scaffold styling always overrides (even though in html source the order does change). I've tried both firefox and ie. Interestingly, I looked in the download files and neither the chapter 9 nor the chapter 10 versions of the store layout files link to the scaffold stylesheet.

#24289: There is an issue with the cascading style sheets being used here. The problem is that both the scaffold and depot stylesheets are combining to produce strange text size and the scaffold stylesheet is overriding the depot stylesheet's link and link-hover code. (it seems to be thse use of pt, px, ex, em text sizes... setting sizes then uses percentages of those newly set sizes) I believe this is an issue of IE and FireFox dealing with CSS differnetly than whatever browser/version the author wrote these examples with. We need a fix for this... almost everyone on the internet uses IE or FireFox.--Brian

#23947: The code given for the ship method should not include the check (if count > 0) if you want to be able to show the intended message "No orders marked as shipped" flash message when no items are checked off before submitting the form.
--Alan M

#24162: The code snippet given for login_controller.rb is missing an "end" statement at the end of it to close :

class LoginController < ApplicationController
--BrianWarren

#24004: It seems that the generator for unit tests no longer generates the setup method (Rails 1.0.0) so the code in the PDF/Book doesn't match watch is actually produced.--Dave Myron

#24246: the "assert_equal 29.95, @product.price" fails for me, even though product.price is 29.95. The only way to make this assert to pass it was changing the line to "assert_equal 29.95.to_s, @product.price.to_s"

#24340: Maybe it helps someone: here is the correct code for test_read_with_fixture_variable:

def test_read_with_fixture_variable
assert_kind_of Product, @product
assert_equal products(:version_control_book).id, @product.id
assert_equal products(:version_control_book).title, @product.title
assert_equal products(:version_control_book).description, @product.description
assert_equal products(:version_control_book).image_url, @product.image_url
assert_equal products(:version_control_book).price, @product.price
assert_equal products(:version_control_book).date_available, @product.date_available
end

#24295: New to RoR, so this might be the same as the erratum beginning "With the new testing rules, a good chunk ".
In the test_read_with_hash method, the line
vc_book = @products["version_control_book"]
fails because @products is nil, i.e. does not seem to have been filled in.
Similarly, in the test_read_with_fixture_variable method, the line
assert_equal @version_control_book.id, @product.id
fails because @version_control_book is nil.--David Hislop

#24005: test_helper.rb looks (mostly) nothing like the example in the PDF.

It now looks like:

ENV["RAILS_ENV"] = "test"
require File.expand_path(File.dirname(__FILE__) + "/../config/environment")
require 'test_help'

class Test::Unit::TestCase
# Transactional fixtures accelerate your tests by wrapping each test method
# in a transaction that's rolled back on completion. This ensures that the
# test database remains unchanged so your fixtures don't have to be reloaded
# between every test method. Fewer database queries means faster tests.
#
# Read Mike Clark's excellent walkthrough at
# http://clarkware.com/cgi/blosxom/2005/10/24#Rails10FastTesting
#
# Every Active Record database supports transactions except MyISAM tables
# in MySQL. Turn off transactional fixtures in this case; however, if you
# don't care one way or the other, switching from MyISAM to InnoDB tables
# is recommended.
self.use_transactional_fixtures = false

# Instantiated fixtures are slow, but give you @david where otherwise you
# would need people(:david). If you don't want to migrate your existing
# test cases which use the @david style and don't mind the speed hit (each
# instantiated fixtures translates to a database query per test method),
# then set this back to true.
self.use_instantiated_fixtures = true

# Add more helper methods to be used by all tests here...
end
--Dave Myron

#23885: test_helper.rb must be updated in order for the test_destroy method to be used if the MySQL database is using MyISAM tables (the default for CocoaMySQL). One must set "self.use_transactional_fixtures = false"--Parker McGee

#23887: Rails has recently, in 1.0, updated the way it handles fixtures. If you're having problems, see Mike Clark's weblog here: http://clarkware.com/cgi/blosxom/2005/10/24#Rails10FastTesting--Parker McGee

#24022: The @products and @version_control_book instance variables are not available unless you set the 'use_instantiated_fixtures' attribute to true. This can be set in the TestCase class Rails creates for you.--Cathal

#24561: when running test_read_with_hash omitting the before_type_cast suffix to the date assertion seems to allow the test to pass - including it causes a mismatch between the hashed fcture date and that called from test.database --laura herald

#28629: in your tests, you are testing the equality of ambiguous numbers (floats). The prices you chose for the fixtures just _happen_ to work, but other prices will cause this error:

1) Failure:
test_add_unique_products(CartTest) [test/unit/cart_test.rb:15]:
<#<BigDecimal:b74e84fc,'0.5988E2',8(16)>> expected but was
<59.88>.

use (for example)

assert_equal sprintf("%0.2f", (@version_control_book.price + @automation_book.price) ), sprintf("%0.2f", @cart.total_price)--Andrew Yates

#24080: The file examples for store_controller.rb thoughout this section contain arguments to the fixture() method that look like this.
fixtures :products, :orders.
Having both these fixtures referenced causes ActiveRecord::StatementInvalid: Mysql::Error: type errors to be raised when the test is run.
This appears to be the result of the fact that orders.yml at this point has not been modified, and contains the default generated values that Rails provides. The errors specifically mention that the orders table does not specify default values that the ActiveRecord tries to insert into this table using such assign statements as assert_equal 1, assigns(:items).size.
The addition of the necessary fixture information to oreders.yml is discussed at the end of this section, and unless I'm mistaken, has not been suggested before this.
Perhaps this will help others to trace this problem without wasting much time.--m gentzel

#24441: Book Version 2005-12-20. Running "ruby test/functional/login_controller_test.rb" produces error: "method 'before_destroy' for LoginController:Class (NoMethodError)". I comment out the "before_destroy dont_destroy_dave" line from login_controller.rb and I get the result specified in the book (the "302" error). Being new at ruby/rails (learning directly from this book) not sure how to fix this... (running 1.8.2 on XPPro)--Martin Crundall

#24197: This may be a Ruby/Rails version problem (Ruby 1.8.2, Rails 1.0.0). The session array is problematic. I see examples with it as a global variable (your book) and as a session variable (e.g., Four Days on Rails). Confusing, but not your fault. However, on my version(s), session is not a global variable in the functional test code, it is a field of @request. This makes sense to me and my code runs.--Jeffrey L. Taylor

#27028: The listing of application.rb is inaccurate. It should include the following lines just after the class declaration:

# Pick a unique cookie name to distinguish our session data from others'
session :session_key => '_depot_session_id'

--Tom Pollard

#24297: File 121 online (http://media.pragprog.com/titles/rails/code/depot_testing/test/performance/order_test.rb) doesn't quite match the code on the page, nor the code from the zip file I downloaded. I didn't diff it, but at least the elapsedSeconds test is different (8.0 in the book and zip, 3.0 in the online file, probably as it once was in the book?)--David Hislop

#24151: Running the code listed for File 121 on Page 176 of the book (12.7 Performance Testing) produced an integrity constraint violation error from MySQL caused by the Order.delete_all statement, which attempts to deletes orders which are tied to LineItems.

Modifying the fk_items_order constraint (create.sql, file 37/106, page 102/502) to include an ON DELETE CASCADE clause resolves the problem, by removing the line item record when deleting the associated order.

ALTER TABLE line_items DROP FOREIGN KEY fk_items_order;

ALTER TABLE line_items
ADD CONSTRAINT fk_items_order FOREIGN KEY (order_id) REFERENCES orders(id) ON DELETE CASCADE

I did use phpMyAdmin to create the tables and constraints, but going over it, I'm not sure why it would have worked for you guys, unless MySQL prior to 4.12 had default different behavior.--Nathan Youngman

#23936: Grouping controllers in modules: I have grouped controllers as described in this section using the command format:
myapp> ruby script/generate controller Admin::Book action1 action2 ...
FIles and directories were generated as described, however, a request formatted according to your example:
http://my.app/admin/book/edit/1234
results in an error "no action corresponds to book".--sharon

#24548: Order.find_on_page(page_num, page_size)

the above function defined should probably be more like,

def Order.find_on_page(page_num,page_size)
find(:all, :order => 'id', :limit => page_size, :offset => (page_num-1)*page_size)
end

note: (page_num-1)*page_size is the change--Mike Nelson