Historical errata for Agile Web Development with Rails

Under “Create the Maintenance Application”, it says we’ve created our development, test, and production databases, but at this point we’ve only created the development database (on Page 59).

Caption for Figure 6.4: “rpice” should be “price”

“tablein” should be “table in” in sentence “It would be nice if we could load the data into our tablein a more controlled way.”

Bullet item “We used Ruby’s sprintf() method …” - in the preceeding rhtml fragment sprintf() is never called.

Bullet item “We used Ruby’s sprintf() method …” - “divied” should be “divided”

“is” should be “in” in “… we jointly decided to see what happened if we displayed the products is alphabetical order …”

Last paragraph on page, missing “we” in “This time, we used the h(string) method …”

“ona” should be “on a” in “Pragmatic Web Designer is off getting inspiration ona beach somewhere …”

Last paragraph, “that” should be “than” in “It’s starting to get a bit long, so rather that include it inline …”

Second to last paragraph, “renefered” should be “referred” (or possible “referred to”) in “The methods we define in a controller-specific helper are available to views reneferred by that controller.”

The format string for sprintf() could be “$%d.%02d” instead of “$%d.%02.2d” since cents will always be in 0..99 (marginally easier to parse by the reader).

Second paragraph, “one” should be “want” in “And a POST request is just the ticket when we one to do something like add an item to a cart.”

Top of page
——————

wil create and drop…

should be,

will create and drop

The word “change” is misspelled as “cdhange” in the sentence that reads, “For programmers used to other web frameworks, where a simple cdhange to the schema could involve them in half-a-dozen or more code changes, this was a revelation.”

Add a space between the words ‘table’ and ‘in’ (third paragraph, next to last sentence)

“It would be nice if we could load the data into our tablein a more controlled way.”

Should be,

“It would be nice if we could load the data into our table in a more controlled way.”

(Dave says: I see a space there)

“it you now look at your database” should be…

“if you now look at your database”

Under the “Change Log” the first sentence reads: There are probably a thousand changes that have been applied to Rails since the first version of this book wen to press.
I think it should read: “…since the first version of this book went to press.”

Running “rake db:sessions:create” does not create a new table called “sessions”, it creates a new migration for adding the “sessions” table.

You must run “rake db:migrate”, as well.

Omit the word ‘and’. (Paragraph before ‘Creating a Cart’ section.
—-

"Be careful as you add additional methods to this controller as we work further on the cart

Last line on page.
—
“The cart is basically a wrapper for an array of atems.”

should be,

“The cart is basically a wrapper for an array of items.”

Hi David, nice 2nd edition! I’m Marc, one of the developers from the RadRails team. Thanks for the plug in the book, that’s awesome! One correct though, we didn’t win a Jolt award, it was the Best Open Source Developer Tool based on Eclipse award. Thanks again!

Change ‘tart’ to ‘start’

"Let

Change “you’ve” to “you’re” (Last paragraph)

"Your browser will probably warn you that you

Change “out” to “our”

"Rather than out new cart, we

Change “out” to “our”

"…we

It may be somewhat damning to suggest that there are “There are no fully-fledged IDEs for Ruby or Rails.” especially given the existence of RadRails.

That said, Textmate owns and I’m glad to see proper respect paid here :-)

Just wouldn’t want the book to appear too dated even before it goes to print.

“WEBRick or Lightthpd?” Lighttpd is misspelt with an extra h in the header of the callout

I would be nice to see an example of prefilling tables from
migration code. For example if you want to want a table to allready contain an admin or something like this.
cheers

(Dave says: this wil be in the migrations chapter)

5th paragraph - `Just look at the
cart in the sdiebar.’

“managerl” should be “manager”

“It has one slight disadvantage: you delegate control of your installation layout to the package managerl.”

Could we get a pointer to using Subversion with Rails here? Maybe a condensed version of http://wiki.rubyonrails.org/rails/pages/HowtoUseRailsWithSubversion or http://maniacalrage.net/past/2006/4/12/heres_how_i_create_a/ or http://blog.teksol.info/articles/2006/03/09/subversion-primer-for-rails-projects

(Dave says: I think that’s probably out of scope, and I’m leary about pointing to blog entries from the book)

In the first paragraph, second sentence, the word “a” is missing between “being” and “worldwide”. It should be, “… Rails went from being an unknown toy to being a worldwide phenomenon.”

In the fifth paragraph, second sentence, you refer to “OX X” instead of “OS X”.

In the footnote at the bottom of the page, (where else do footnotes live?), in the second sentence, the word “to” is missing between “you” and “type”. It should read, “It allows you to type the …”

In the footnote at the bottom of the page, second sentence, you have the word “you’re” instead of the word “you’ve”. It should read, “That’s before you’ve written a thing….”

Change ‘out’ to ‘our’. (First paragraph on page)
—-
"To do that, let

At this point, the suggested change to the store_controller.rb should be

redirect_to :action => :index

instead of,

redirect_to_index

The redirect_to_index method requires a “msg” for the flash

(Dave says: I shows the updated version of redirect_to_index instead)

“So, open up store.rhtml in the app/views/layouts directory, and replace the button_to( ) call with something like this:”

Should be,

“So, open up index.rhtml in the app/views/store directory, and replace the button_to( ) call with something like this:”

In the “Ho! Ho! Ho!” example, an extra line break is inserted before “Merry Christmas!”, but the code would not cause that line break.

(Dave says: The output shows the generated HTML, not the appearance on the screen. It will have a blank line)

The first word on the page should be “Arachno” instead of “Aracho”.

You mention that components have fallen out of favor in the Rails community. However, nowhere does it seem to be explained why components are seeing less use, nor are we ever told what is being used to supplant them. For those of us who’re already using components in our applications, it would be nice to see what problems the Rails community has had with them, as well as have suggestions of what to migrate them to—especially if there’s any intention to remove components from Rails in subsequent releases.

The third paragraph in the section named “The Desktop”, third sentence, the word “tree” at the end of the sentence should be plural.

Although the error and subsequnt technique for fixing it is discussed on page 103, the code listing for app/models/cart.rb on page 101 already includes the line:

include Reloadable

I would assume you meant to leave it out until later…

In the first paragraph, the last sentence doesn’t make sense. It should be rewritten to be something like, “The Rails database adapters hide these differences so that a Rails application doesn’t need to know what kind of database it is running on.”

Inside the “Selecting a Different Database” box, last sentence: “… apppropriate …”

Second paragraph: “Edit the file so that in looks like the following.” s/in/it/

Footnote: “Your scheme will be transported back in time…” s/scheme/schema/

Footnote: “so Rails store …” s/store/stores/

Change “scoffolding” to “scaffolding”.

Syntax highlighting for many .rhtml code examples in the book are inconsistent and not colored properly. For instance, almost every occurance of a closing html tag has the greater-than symbol the wrong color. Minor detail, but noticable…

Third paragraph after the app/controllers/store_controller.rb code example, it says:

“Or maybe the user has tried to enter an order, but thier data hase failed validation.”

Where it should be the word “has”:

“Or maybe the user has tried to enter an order, but their data has failed validation.”

Last paragraph states:

“These helpers interactive with the controllers…”

Where it should be “interact”…

“shouw” should be “show”

Next to last paragraph: "… let

Next to last paragraph: “… as we saw, startign with a conventional application mkes it easier …” s/startign/starting/, s/mkes/makes/

Footnote 5: “… being fed values fromthe forms it creates.” s/fromthe/from the/. Also, “If there were to set an unknown payment type, …” s/there/they/?

Although mentioned several times, it appears the requisite code to implement validation on the Order model has been skipped. Consequently the suggestion to place an order without filling in any of the form fields succeeds in placing an order without any details.

example rxml code shows “xml.name” but Figure 12.2 on the next page shows it as being “title”

Footnote 3 refers to “global Window’s environment” which should probably be “global Windows environment” instead.

shouldnt YIELD be used instead of content for layout?

(I’m fighting hard not to have @content_for_layout deprecated. In fact, I just talked with DHH, and content_for_layout isn’t going away)

In ‘The Desktop’ the 2nd sentence of the 2nd paragraph ends in ’ and a browser window permentently open’ where it should be ’ and a browser window permanently open’.

In Editors paragraph:

"I

The logo.png provided at http://media.pragprog.com/titles/rails2/code/depot_e/public/images/logo.png has a different background color than the depot.css sets the top bar to.

A reminder to run ‘rake db:migrate’ might be worthwhile.

The ‘authorize’ method should probably not be public; as written it creates a new action in every controller.

In delete_user, render(:action => :list) should probably reference :list_users instead. Better yet, it should redirect.

“far deeper that I’d expected” -> “far deeper than I’d expected”

The reader is told to run the command:

depot> rake db:session:create

It should be

depot> rake db:sessions:create

In the discussion of the :action parameter of the link_to method, the code in the rhtml file has double quotes around the action “goodbye” but the description has the action in single quotes. For someone totally unfamiliar, you should either use single quotes in both locations or double quotes in both locations in order not to confuse them.

After the code listing: "Here we

End of third paragraph: “… convenient way to refenece test data …” s/refenece/reference/

Since the depot schema is now storing product prices as an integer, the fixtures should also reflect this.

When editing add_to_cart.rhtml, item.title is not correctly changed to item.product.title, causing an error upon page reload.

(Dave says: CartItem has a title method)

assert_nil/assert_not_equal should be assert_nil/assert_not_nil

In the sentence immediately before the “Configure the Application” title, you have a reference to another section that was not correctly expanded. sec.test.db.config on page ?? should reference the real location.

Under assert_tag :content, "

Change ‘iterm’ to ‘item’ (First paragraph)
—-
“Click to add an iterm to the cart,”

Should be,

“Click to add an item to the cart,”

Replace ‘and’ with ‘a’ (Forth paragraph)

"…the whole sidebar redraws on the transition between and cart that

“ClassItem” should be “CartItem”

I really think you should encourage people to not connect to mysql as the root user with no password. It is relatively easy to create a user to use with the depot databases that has all the permisions it needs. For instance, after the line that creates the development database you could have the following lines:

depot> mysql -u root -p mysql
mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP
-> ON depot_development.*
-> TO ‘depot_user’@‘localhost’
-> IDENTIFIED BY ‘agile_shopper’;
mysql> quit

This creates the user that only has access to this database when connecting from the localhost. Tell the users to update database.yml with the user and password. Also warn them that the database.yml file that is generated does not have a space after the : for the password field for the development and test databases. This space is required. The password field for the production database does have a space already.

By recommending this, people will not get into the habit of using the root user without a password. We do not want a repeat of the type of break-in that occured to the apache.org web site which was accomplished by attacking the site through its connection to mysql using the root user and no password.

(Dave says: it’s a tradeoff: do that, and folks tie themselves into knots as we proceed. Instead, I’ve added an addiitional warning)

It the paragraph describing 002_add_price.rb, the word “featutes” should be “features”.

The sentence: She notices a minor things … is obviously not right.

The sentence: Rather that use …
should be: Rather than use …

The sentence: Again, we’ll rely of CSS to do …
should read: Again, we’ll rely on CSS to do …

Refers to the _add_to_cart template. Since this is an .rjs and no longer a partial, the leading underscore should be removed.

Replace “browsers” with “browser’s”.

Fill in “Chapter xx of Pragmatic Ajax [?]” with the eventual values.

would be nice to see some real world examples like in the first book. i think you’ve got a lot to tell with all the 37 signals apps. ;-)

(Dave says: I’m not privy to 37s architectures… Ezra is writing a whole book on deployment, though)

2nd last paragraph, “disovered” -> “discovered”.

The database migration file for create_orders is numbered 004. However, if following the tutorial, it will be 005.

The create_line_items migration suffers from this same problem—it should be 006.

Third bullet:
We use JavaScript effects, Ajax,and RJS

should be “Ajax, and”

"which wasn

The locomotive URL runs off the side of the page (first line)

In the code listing for checkout.rhtml, the first div is given an id=“depot-form”. However, in the CSS for this page, depot-form is referred to as a class rather than an id (as a “.” vs. “#”). This should be changed to either a class, or the depot.css file should be changed to refer to depot-form as an id.

The code listing is missing a final ‘end’ to close the definition of regular_user.

Next to last paragraph: “… (you can get to it in the instance variable integration_session is you really need to).” s/is/if/.

Code listing, top of page: “open_sesion do |sess|” s/sesion/session/.

The example code for ‘host’ is confusing and/or erroneous.

sess.host = fred.blog_per_user.com

I imagine there should be quote marks around the assigned value. But also, underscores are not technically valid in domain labels corresponding to host names. And finally, (as a general suggestion throughout the book) it is preferable to use the reserved top-level or second-level domains for all documentation examples (e.g. example.com) rather than a live (or potentially live) domain. See RFC 2606.

Next to last paragraph: "Let

It would make more sense if “starting out with the difficult stuff” read “starting out with the easy stuff”.

“Because Rails it just so easy” -> “Because Rails is so easy”

(Dave says: I’m happy with the idiom)

In the last paragraph, second sentence, it starts with 2 “the”s.

The first ‘is’ in ‘Whenever an item is the cart is updated’ is a typo.

Under the description for @sent_on: “ATime object …” s/ATime/A Time/.

Line 1 : “a indication”> “an indication”

On page 316, there is a “Joe asks” section saying how after_find and after_initialize would slow things down, so the Rails team agreed on increased performance.

However, this section does not make sense at this stage because we were not introduced to any differences in the after_find and after_initialize methods until a page later. In other words, I was momentarily left wondering, “…alright, they agreed on increased performance how?”.

It may be more beneficial to include this “Joe Asks” section after the text “Because of a performance optimization, the only way to define callbacks for the after_find and after_initialize events is to define them as methods.”

There’s something wrong with the PDF. On this page most of the page makes the cursor really large and doesn’t allow me to select single lines of text with Adobe Reader 6.0.

(Dave says: I don;’t see this here wth either Acrobar or Preview. Anyone else see this?

The last sentence that goes to the next page about there being more information in database.yml about PostgreSQL and SQLite isn’t true. My database.yml only contains information for mysql by default.

"or if you

"And once you

"what

Spacing in footnote is messed up.

Both chapters 9 and 10 are labeled as task D

Second line, first paragraph, find is mispelled. “Whenever you fnd …” should read “Whenever you find …”

In the example loop 3.times do … the correct output should be:
Ho! without the html tag

:only_integer should be false to allow float values

()Dave says: the price is an integer)

There is a ‘return’ immediately into the ‘def verify_premium_user’ method, which seems like a typo.

Line 2: “comamnds”> “commands”

I don’t quite know how to categorize this one. The footnote says for Windows “editing your system properties and setting to 9.” Which doesn’t mean anything. I guess you mean to say “changing your \\HKLM\\Software\\Microsoft\\Command Processor\\CompletionChar registry value to 9”, and maybe mentioning XP enables it by default.

(Dave says: yes - there was a missing chunk in there)

At the end of the Editors section, you say “learn a new feature each day.” I feel it should read “every day,” while the distinction in usage is fairly subtle and somewhat subjective.

While the footnote is already lengthy, you may want to suggest using OS X’s Console application where appropriate as a free alternative with a convenient clearing/updating

This is somewhat snarky and more annoying than poignant:

Rather that use

(debatable) “100Mb of IDE” -> Be it HDD or RAM, you’ll talk in megabytes, not megabits, I guess. The proper abbreviation then is “MB”.

In the second to last paragraph, last sentence, you have “if” instead of “in”. It should read, “And, in fact, …”

Not all environments specify the correct ruby binary. The command should start with ruby: depot> ruby script/generate model user

In the third paragraph, fourth sentence, missing the word “the”. It should read, “… can interact behind the scenes with the application …”.

In the last paragraph, third sentence, you are missing the word “use”. It should read, “… the main template arranges to use a variable …”.

In the paragraph right after the code for add_to_cart.rjs, last sentence needs to be re-worded to something like, "… to use that HTML to replace the content of the

After making all the updates and reloading the store/index page and verifying that the correct HTML was generated and the javascript files loaded, the AJAX cart still didn’t work and the developer.log file had the following error:

ArgumentError (undefined class/module Cart):

Stopping WEBrick and restarting it caused this error to go away. You might want to mention this in the troubleshooting discussion as a way to make sure all the current code is loaded.

In the first paragraph of the section titled “The Customer’s Never Satisfied”, the fourth sentence, the work “now” is used instead of “know”. It should read, “… eager for the praise we know will come.”

In the second paragraph after the code fragment, second sentence, you have “reveals” instead of “reveal”. It should be, “… which will smoothly reveal the cart, …”.

Bottom of the page in the Troubleshooting section. First sentence “it can’t make (if) foolproof” (if) should read (it).

In the fourth bullet point, the word “change” needs to be plural. It should read, “To help the user see changes to the cart, …”

the line item cannot exist unless the correspond order and product rows exist.
should be
the line item cannot exist unless the corresponding order and product rows exist.

In the last paragraph, first sentence, it states that we are currently at version 3 of the database when we are actually at version 4. As mentioned in previous errata, the new migrations are numbered 005 and 006.

dealing with an object names “order”
should be
dealing with an object named “order”

In the third paragraph, third sentence, you have the word “the” instead of “then”. It should read, “… into the form fields, and then extracting …”

In the fourth paragraph, last sentence, the word “automatically” is mispelled. The sentence should read, “… it is automatically associated …”

In the fifth paragraph, second sentence, you have the word “need” instead of “needs”. It should read, “It’s important to remember that Rails needs to know …”

“…support for…mock objects” - unless it’s newly arrived in Rails 1.1, Rails supports stubs rather than genuine mock objects. You will remember this comment (from a number of people) from the first edition.

Most developers will probably recognize l33t speak, but some managers and other readers may not. You may want to add parenthetical definition.

The first page and a half do not read very well (not meaning to sound offensive). They seem like fragments gathered from various casual emails. The Intro read much more smoothly in the first edition.

Many paragraphs are tied together so there is no clear logical break between them.

“Rails went from being an unknown toy to being worldwide
phenomenon.” would sound more “professional” as “Rails went from being a private experiment to being worldwide
phenomenon.” The use of “toy” is not confidence inspiring.

The use of “strong” language can be used to good effect occasionally, but you do risk offending some.

As someone looking at various frameworks for a upcoming project, it would be nice to see more specific information comparing features of other frameworks (without bashing). Catalyst, Jifty, Maypole, PHP, Struts, etc. It would be considered flame-bait on mailing-lists, but I’ve seen it done well in books. It doesn’t have to bash, just compare features.

The example code may not be a helpful illustration to beginners. A short paragraph explaining that the code describes the relationships of various data and sets up some constraints would help.

I’m not sure the ChangeLog for the depot application is helpful to most readers-it is probably more confusing to first time readers than helpful.

The depot.css for .depot-form legend should specify font-family instead of font-style. The definition should be:

.depot-form legend {
color: #dfd;
background: #141;
font-family: sans-serif;
padding: 0.2em 1em;
}

Small typo at first line of page: application writen as appplication.

In the add_to_cart method the product variable could be local as it’s not used in the view.

product = Product.find(params[:id])
@cart.add_product(product)

Instead of:
product = Product.find(params[:id]) cart.add_product(@product)

Near top of page, “And, once youve chosen an editor” - missing apostrophe in “you’ve”.

On this page you have body text saying what script/server does, a footnote explaining what WEBrick is, and a box at the top of the page saying that you might not get WEBrick after all. It’s a bit jumpy. Could you put the description of WEBrick in line, and the ‘exception’ regarding lighttpd in a footnote?

Keeping up to date also involves updating Rails files in your application.

Gratification could be substantially more instant if you generated the Say controller with the hello action (script/generate controller Say hello). Browsing http://localhost:3000/say/hello then shows a view that tells you where it lives (“Find me in app/views/say/hello.rhtml”).
You would still show hand-addition of action and view when going on to “goodbye”. (For “goodbye” you could also demonstrate that the action method isn’t required - that Rails will go straight to the view with the same name if there’s no action method.)

(Dave says: I thought hard about tis originally, but decided this flowed better, as it let me talk about URLs a bit)

Seeing dynamic content is great, but the reader may be immediately concerned about the displayed date/time format. This is a perfect opportunity to introduce a helper method.

“Why did we go to the extra trouble of setting the time to be displayed in the controller and then using it in the view?” - Fundamentally it’s because the Time instance is a model (although it’s not persistent, and we didn’t have to write the class for it), and it’s the controller’s job to provide the model for the view to display.

(Bottom of page) "If your file uses tabs, not spaces, to indent the lines following depot_devlopment: you

“automatically” is spelled wrong.

"First, the cart and cart item classes weren

“Parameters” is misspelled.

“This takes a hash of values that are the same as the trailing par meters we passed to button_to( ).” should be “This takes a hash of values that are the same as the trailing parameters we passed to button_to( ).”

is can be hard to check
should read
it can be hard to check

You mention that you want to create a private class method self.encrypted_password. Since the readers are probably new Ruby users, you should probably include the private statement in the source snippet. Also, you should mention that create_new_salt is a private method that should occur after the private statement.

When reviewing the code for user.rb, there are two items in the source that were not mentioned in the discussion. The first is the need to require digest/sha1 for the encription done in encrypt_password. It should probably be mentioned in the discussion for this method. The second is the definition of safe_delete. Once again, if it is going to be in the source, it should probably be discussed somewhere.

While I realize inject is likely to be covered in an appendix on basic ruby techniques, as with the first edition, a breakdown of why this was chosen here and its particular mechanics would be a welcome addition.

(Dave says: tough call. I think I’d rather not break the flow here)

2nd line from the bottom should be read “open up index.rhtml” instead of “open up store.rhtml”.

In app/views/layouts/admin.rhtml, the closing paragraph tag for the link_to “Products” link should be

instead of

At the very top of the page:

work> cd depot
work> ls -p

After changing to depot, the prompt should change to depot> (as it does in the next prompt later on the page)

(Dave says: good catch!)

In the command line output on this page, you have typed “rake db:migrate” twice:

depot> rake db:migrate
dave[code/depot_p 14:41:34] rake db:migrate

It should be either one or the other. Judging by the examples in the rest of the book it should be the first.

(bottom of page) “…to indent the lines following depot_devlopment: …”
There’s a typo in the name of the database: “depot_devlopment” instead of “depot_development”

You say:

In reality, database connections work like a charm 99.9% of the time.

I know you mean to be encouraging, but it could actually be discouraging to a developer who sometimes has database issues… and let’s face it, that is not 1 developer out of every 1,000.

If things really worked well 999 times out of a thousand, would you have included a whole page with 10 possible errors to check? Each of those errors would only come up an average of 1 out of every 10,000 times! So maybe once among your readership. At that rate, you could include a warning just to me about my sticky up arrow key.

Awesome book. I am having a blast.

(Dave says: very good point)

(Middle of the page) " and with the file extension (..rb, because it’s a ruby program)"
There’s is two dots in the extension name: “..rb”

You give an example implementation of delete_user that has some problems. You talk about the problems, which is good. However, after discussing the problems, you do not give a revised implementation of delete_user. You can look at the login_controller.rb listing on pages 537-538 to get the final implementation but it would probably be better to include it here.

At the very bottom of the page you say that the code will output the time zone as “CST” When I run this code locally it prints out the full time zone name, as in “Central Standard Time”

(Dave says: I double checked, and I see the abbreviation)

You say “This uses a new command, rake.” but you did already use it earlier to test DB connectivity

The current text reads “is the exception is thrown” and I believe it should read “if the exception is thrown”

(If instead of is…)

To do that, let

(Middle of the page) “the code generated by the scaffold is indentical to”
Change “indentical” to “identical”

(Middle of the page) “(dropping the .css extension.”
The closing bracket is missing or the opening bracket needs to be replaced by a comma

The definitions for empty_cart and redirect_to_index should match the code listing on page 539. In particular, the redirect_to_index method is used throughout the book with an optional message which does not work with the definition on page 113.

(Dave says: it doesn’t need it at this point. That’s added later)

“let’s flash it’s background” should read “let’s flash its background”

In creating a Rest Interface, you should preface the “script/generate controller info” command with “ruby” since you cannot count on the #! line in script/generate to be right.

(First line) “when we created the development databases for the Depot application”
“databases” should be singular

(Middle of the page) “We can use the model’s valid?() method to check to see if it validates”

IMHO we have to choose between “to check” or “to see” :)

(First line) “and we verify the error message associated with the price attribute is what we expect”
should be “and we verify that the error message”

(Using Fixture data: second paragraph) “However, Rails makes it easier that that”
replace the first “that” by “than”

(Paragraph just before 13.3) the phrase “but these tend be be used infrequently used” should be replaced by “but these tend to be infrequently used”

(Middle of the page) “Rails is populating a response object behind the scenes.”
“scenes” should be singular

(Dave says: this one I’ll leave to the copt editor…)

in the listing for list.rhtml, the line to destroy an entry is written like this:
<%= link_to ‘Destroy’, { :action => ‘destroy’, :id => product },
:confirm => “Are you sure?” %>

This doesn’t work. I think it should be (going by the default scaffolding created by rails) like this:

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

(Dave says: This is a change to scaffolding in 1.1.1. Fixed)

The following occurs twice in a row on the page; I think either it is a duplicate line that needs to be removed, or the second line is supposed to be different, but was not editted to reflect this:

assert_tag(conditions)
assert_tag(conditions)

It’s not immediately clear what the full URL for the images or for the depot.css file mentioned in the footnote is. In fact, I haven’t been able to find the images yet.

First paragraph of the page, last sentence is:

"We

The script/generate command should be preceeded with “ruby” since you can’t count on the #! line to be correct

For ISP’s you could suggest people look at the following page as a starting point

http://wiki.rubyonrails.com/rails/pages/RailsWebHosts

In the keeping up to date section, after the gem update command. It might be a good idea to suggest using “gem cleanup” or whatever that command is. It took me a while to learn about this and I had many gem versions sitting on my computer. This made the first html page of the gem documentation cluttered with all sorts of versions.

(oops, posted this under page 41 as well which was wrong.)

Let

The brief discussion of what symbols are does not help me understand them. I have read the discriptions of symbols in Pickaxe and AWDWR and still do not have a real idea of what they are. I’m sure many programmers like me are coming from languages without a symbols analogy. I would love it if you spent a little more time on what symbols are and how they relate to strings. I know this is not a Ruby book but Rails uses symbols all the time and I feel a little whimpy without an intuative understanding of symbols.

“scaffold :product” You refer to this as a scaffold declaration. The word delaration is not really appropriate. It makes me think “int a;”. This is a call to the scaffold method, just a plain old Ruby method. It could be written as “scaffold(:product)”. This same type of thing confused me for a while with “has many :employees” or “validates_presence_of :name”. I didn’t know these were just plain old Ruby methods. I thought they were some sort of Rails voodoo. Perhaps you could write something like “This is just a call to the scaffold method provided by the Rails framework. It could be written as scaffold(:product)”

(Dave says: I’m thinking that’s too much detail at this point)

In the second paragraph, third sentence, you have the word “valud” instead of “value”. It should read, “… if we pass 0 as the id value to MySQL…”.

“you can define database migrations..”. Then phrase ends with two dots.

The info:
Rake looks for all the migrations not yet applied to the
database and applies them.
How do you know which migration has not yet been applied? I’d find this useful to know at this point.

then replace it

the code in the auto-generated product_test.rb should read “fixtures :products” instead of “fixture :products”

Whenever an item is the cart is updated (either when it is added or when we change the quantity) let

the text:
to validate you typed what you thought
you typed?
There appears to be a typo here.

“instructs the browser to immediate request the URL …” ‘immediate’ should be ‘immediately’.

“In then fades back” should be “It then fades back” on the first line of text.

_cart.rhtml on this page uses the older button_to instead of form_remote_tag code we replaced button_to with earlier in the book.

(Dave Says: I don’t see a prior _cart.rhtml using form_remote_tag. What am I missing?)

It seems that implementing the AJAX “Add to Cart” does not deal with the “Your cart is currently empty” flash if you had previously emptied your cart. Since there isn’t a form post (due to using AJAX) the flash stays on the screen even when you clearly have items in your cart. Being new to Rails I don’t know the proper solution but I “fixed” it by adding page[:notice].hide to my add_to_cart.rjs. I recognize that as a quick fix and probably not the most appropriate but I’m still learning Ruby and Rails.

Change “you access the flash as it it were a hash.” to “you access the flash as if it were a hash.”

The anchor to figure 2.1 redirects to the text below 2.1. It would be nice to be redirected to the top of the page (top of figure 2.1) there, to have the figure displayed when clicking on the reference to it. (instead of being thrown to the subtitle of the figure and having to scroll up)

(Dave says: Sorry, not much I can do about this—the link is to the caption)

“A flash is a bucket (actually closer to a Hash), Into which you can store stuff as you pro-cess a request.”

Probably better as:

“A flash is a bucket (actually closer to a Hash), in which you can store stuff as you pro-cess a request.”

In second to last paragraph, “track done what changed” should be “track down what changed”.

"an

At source code file, “depot_final” folder doesn’t have all migrations files.

(Dave says: depot_final isn’t code referenced from the book: I use it as a placeholder. depot_r is the latest ‘real’ version’)

"If we were rolling the Depot application into production, we

2nd to last line on page “I have JavaScript enabled in oe…” should read “I have JavaScript enabled in one…”

in the store index.html view you include the following code:

<% for product in @products -%>

There is no indication into why you end the code with a minus sign. Is this a special erb thing, or is it part of your code? It’s not obvious :-)

It might be an interesting idea to write the excerpt as ruby code, with comments to where you want the output to go e.g. something like (I know this isn’t that correct..):

for product in @products
#choose css class
#output image
#output product title
#output product description
#output product price with css class price
end

which would allow the reader to get a good feel for the structure of the code before all the clutter of the html is added.

I don’t understand your explanation, Dave. You say:

“I shows the updated version of redirect_to_index instead”

Where in the book is this updated? Here’s the code on this page:

def add_to_cart
begin
product = Product.find(params[:id]) rescue logger.error("Attempt to access invalid product #{params[:id]}") redirect_to_index("Invalid product") else cart = find_cart
cart.add_product(product)
redirect_to_index
end
end

One version of redirect_to_index is taking an argument, one isn’t. Throws an error for me.

In last paragraph, it mentions adding the layout directive to the login controller. We’ve already done that on p. 153, but the text makes it sound like it’s the first time we’re learning about this. I like the idea of reiterating the point with additional background info, but perhaps an acknowledgement that the layout directive should already be in the code would make it clearer.

Should the link_to for delete_user be a button_to instead?

“You might…when developing.” The end of paragraph 6.2 has already stressed this point, so this seems superfluous.

Just a consistency point.

In add_to_cart.rhtml you use the variable `cart_item` as a placeholder for all the cart_items in the cart:

<% for cart_item in @cart.items %>

yet on page 99 you use `item` as the variable name:

<% for item in @cart.items %>

Is there any real reason for this? If not it might be nice to keep them the same - either one is nice, although I personally prefer the more verbose `cart_item`. :-)

It might be useful to clarify what %{} does.

“Figure 7.1 .” There’s a leading space before the dot.

“on page ??.)” The closing paranthesis was never opened.

“Figure 7.2 .” There is a leading space before the dot.

<%= item.title %> should probably be <%= h(item.title) %> as in the index view. Same goes for later listings of the add_to_cart view.

The title should not be task D: Checkout but task E: Checkout…

2nd paragraph:

Just what to these various declarations do? -> Just what do these various declarations do?

“Remember how the render with the collection option inside the add_to_cart template set the variable cart_item inside the partial. It turns out we can do the same when we invoke a partial.”

This is rather a complicated couple of sentences. Would it be possible to break them down into more digestible chunks? Something like the following would be nice:

Remember the ‘render with collection option’ that we used inside the add_to_cart template? That render option set the variable cart_item inside the partial for us automatically. It turns out we can also do this explicitly when we invoke a partial.

I like your use of an arrow to indicate the new part of the code. However it only seems to be pointing to the line:

<%= render :partial => “cart”, :object => @cart %>

when in fact this whole section is new:

Maybe there should be an arrow for each line,

->

or the arrow should somehow indicate a range of lines:

(let’s pray this formatting works….)

—>

The object name is incorrect in the last paragraph, sentence 5. "Let

Another code inconsistency.

Your code in the _cart_item.rhtml partial references what looks like an older version of the `format_price` helper. The example given on page 126 is as follows:

<%= cart_item.quantity %>×

<%= cart_item.title %>

<%= cart_item.quantity %>×

<%= cart_item.title %>

In your example add_to_cart.rjs you wrote:

page[:cart].replace_html :partial => ‘cart’, :object => @cart

-> page[:cart].visual_effect :blind_down if @cart.total_items == 1

page[:current_item].visual_effect :highlight,
:startcolor => “#88ff88”,
:endcolor => “#114411”

placing the arrow on the centre line indicating that it was new. You haven’t indicated why you place this blind_down effect between the other two statements. Is order important here? Otherwise why not put it at the end?

Looks like redirect_to_index(message) maybe should be rewritten as redirect_to_index(message=nil). Otherwise if you call redirect_to_index without any argument as you do on this page, you get a missing argument error.

The text after the index.rhtml code suggests using <h …> regularly, but the code itself uses the h(…) construct. It’s easy to figure out what’s going on, but the discrepancy made me go back and re-read the text and try both methods to confirm that one or the other wasn’t a typo.

Maybe I missed it, but I think the “dont_destroy_dave” method got removed from the Depot application: I worked through the whole thing and never added that code. The text says that we’ve already seen this callback in action, which I think got removed.

(Dave yes: yes—you’re reading a version 1 chapter here)

“To do this, we use the belongs_to( ) declaration twice the line_item.rb file.” at the bottom of the page doesn’t seem to make sense.

Looks like there’s a missing ‘in’:

"To do this, we use the belongs_to( ) declaration twice IN the line_item.rb file.

I struggled to understand some of the concepts presented on this page:

“As this form will have to link back into a Rails model object,”

I don’t quite understand what this means, it assumes that I know why the form links to a model object. I guess it means that the data captured by the form will be manipulated within Rails by a model object, but I’m not sure.

“However, consider the general case. Maybe we want to edit an existing model. Or maybe the user has tried to enter an order, but their data hase failed validation. In these cases, we want any existing data in the model shown to the user when the form is displayed.”

This section required a number of re-reads before I understood that it was talking about auto-populating the form with existing model data. To me it would make sense that you introduce the concept and then explain why it is useful:

We also might want to auto-populate a form with existing model data. There are a couple of reasons for this. Firstly, we might want to create a model with default values, and have these values visible when the form first loads. Secondly, the user might have tried to enter an order, but their data failed the validation procedures, and on reloading the form we don’t wish to lose the information they have already entered.

"However, simply knowing the name of the object isn

"Before we move on to that new action, though, let

We are told to edit the index, but that action was not specified when we generated the login controller.

“The template is in the file index.rhtml
in the directory app/views/login.”

In the code listing for depot_c/db/migrate/003_add_test_data.rb,
the definition for self.down begins before self.up ends. In other words, it should read:

1. . . .
   end

def self.down
Product.delete_all
end

Using Rails 1.1.2, I couldn’t get the add_to_cart.rjs to work as written:
page[:cart].replace_html :partial => ‘cart’, :object => @cart

It did work when I changed to
page.replace_html ‘cart’, :partial => ‘cart’, :object => @cart

(Dave says: I checked here, and with the core team. The first form should work with 1.1.2)

“It falls through to the end, and renders the form associated with add_user”

would it not be clearer to say that it renders the view associated with add_user?

you’re passing more than one parameter to the assert statement:

assert product.valid?, product.errors.full_messages

yet you don’t explain what these parameters do. It would help a lot if you could spend some time explaining the various assert statements, and the parameters they can take. :-)

Testing the Cart: fifth paragraph “…a cart, it check to see” -> “…a cart, it checks to see”

Perhaps you should note that ‘@content_for_layout’ has been deprecated in favor of ‘yield’?

Dave says: Actually, it hasn’t. Marcel mispoke in the blog entry.

assert_success() is deprecated so it should not be used anymore…

in my case it’s located in /actionpack-1.12.1/lib/action_controller/deprecated_assertions.rb

When I did this, the scaffolding generated code using

@products_pages

not

@product_pages

(Dave says: Perhaps you typed ‘products’ on the command line?)

Throughout this section (Task D) of the tutorial, “Empty Cart” button remains untouched. This seems to be a serious deficit in our application, as it can leave the notice “Your cart is currently empty” and have the content of the cart shown in the side bar. This problem should become clear quite early (as early as at the end of the iteration D2). I suppose you can “leave it as an exercise for the reader”, but at least you should acknowledge this shortcoming.

"we

As well as having a description of the asserts, would it be at all possible to have simple examples of them?

validates_format_of :image_url,
:with => %r{\\.(gif|jpg|png)$}i,
:message => “must be a URL for a GIF, JPG, or PNG image”

running the browser shows:

app/models/product.rb:14: parse error, unexpected $, expecting kEND

Line 14 is the regex line.

No idea what the correction is. I’m stuck?

(Dave says: I don’t see that problem here: are you running the code downloaded from our site, or did you type it in? If the latter, I suspect a typo)

It’s not the idiom of “Rails it just so easy”, it’s the typographical error: “is”, not “it”.

Change
[…]
cart only appears with it has some content?
to
[…]
cart only appears WHEN it has some content?

you include login_controller_test.rb, yet you don’t include everything. In my stub i have the following extra lines:

1. Replace this with your real tests.
   def test_truth
   assert true
   end

It would be great if you could either mention that the actual file has more lines, or include them (my preference would be the latter).

“depot> ruby test/functional/login_controller_test.rb -n test_index”

is there any chance that you could briefly explain what you’re doing here with the -n flag, and why you need to pass it test_index?

Footnote 5, first sentence.
“…, given that it’s value comes ….” should read
“…, given that its value comes ….”

It’s late and maybe I’m being dumb, but I don’t get this sentence:

We

Next to the line
# (create the session table with ‘rake create_sessions_table’)
There’s a margin comment, “see if they update env.rb with new task names”

They did. The line in my rails project is
# (create the session table with ‘rake db:sessions:create’)

Isn’t content_for_layout being replaced by yeild?

- <%= @content_for_layout %>
+ <%= yield %>

(Dave says: no)

after rake db:session:create, I had to rake db:migrate to let the changes make effect. Didn’t see this in the book.

Iteration F1 should be E1, etc. now that chapter is renamed Task E: Administration.

Technically speaking, a Ruby method isn’t invoked. Instead, a message is sent to an object which may be handled by a method. Because Ruby takes this to the extreme, it may be worth mentioning to Nuby’s lest they gasp at the missing 1-to-1 link present in so many other languages.

(Dave says: that’s really just semantics: a method is indeed invoked, and the message sending metaphor (which I believe I introduced in the PickAxe) is actually more confusing than not to newcomers)

<%= render :partial %> or <%= render(:partial) %> (see preceding pages)? Sticking to one notation would help overall consistency.

“(using javascript_include_tag?” misses closing paranthesis

“methdod” should read “method”

Another note on consistency: this is the first time that I’ve seen you use single-quoting, which may strike some readers.

“facailities” should read “facilities”

The URL to the ActiveMerchant library isn’t hyperlinked.

“of the form_for, it is automativally associated with the data”

automativally = automatically.

It is not "Tous les jours,

once youve chosen an editor => should read once you have chosen or once you’ve chosen

if request.post? and user.save flash[:notice] = "User #{user.name} created"
@user = User.new
end

should be

if request.post? and user.save flash[:notice] = "User #{user.name} created"
else
@user = User.new
end

(Dave says: no, this way it lets you create multiple users)

Well, I figured out one way to ensure the hiding of the “Your cart is currently empty” flash message when a customer adds a book to the cart. Be sure to add the “page[:notice].hide” line Jason suggested to the add_to_cart.rjs file (see comment above) AND modify the store.rhtml file in the layouts directory. Add the two lines indicated below:

This ensures that the application will work the first time a user adds a book to the cart, bypassing the error I mentioned earlier. I would have preferred to add an if clause to the .rjs file, but nothing I tried worked. Any suggestions?

(Dave says: Yup: look at the end of the checkout chapter for an example.)

In the footnote (6) from the previous page you say "With other databases, the value might come back as a string, so you

just deleted the only administrative used from the system
should be
just deleted the only administrative user from the system

The scaffold-generated controller code requires a POST to destroy content, but your sample list.rhtml code generates a GET URL. It took me a while to figure out why it was silently not deleting the table entries.

And, as we saw, startign with a conventional application mkes

should be “…starting… makes”

but their data hase failed validation

in the login action of login controller, how to redirect to the original referrer?

i.e. when try to access “admin/list” I should be taken to “login” and back to “admin/list”

If I login from “login/login” then I should be taken to the list?

I have stored the original referrer information in session and used this information to redirect. But the code looks ok…is there a better way?

In paragraph starting “Using your favorite editor..” file admin_controller is referenced without extension “.rb” - extension should be there for clarity.

In the footnote: “Remember.. cents..” Actually, I don’t remember any mention of the price being stored in cents. Either point this out when the column is added or motivate here why you want to use cents rather than dollars. [cents will not allow micro payments, though :-)]

Hmmm…let

“See the discussion starting on
page 380 on page ?? for some more background” -> on page is repeated.

class OrderController; def rescue_action(e) raise e end; end

should be;

class StoreController; def rescue_action(e) raise e end; end

ORIGINAL:
Rails went from being an unknown toy to being worldwide
phenomenon.
CORRECT (Change between <<>>):
Rails went from being an unknown toy to being <> worldwide phenomenon.

It allows us to specify one path through the action is the exception is thrown, and other if it isn

“delcarations” should read “declarations”

The “depot-form” selectors should be IDs instead of classes for them to work properly (at least in Firefox). i.e. replace all occurances of “.depot-form” with “#depot-form”.

Underneath “What We Just Did”, the first bullet item has a trailing dot but the second and third haven’t. Also, “we did the following” might use a trailing colon instead of a dot.

ORIGINAL
It turns out that we can: migrations to the rescue.
Problem:The colon next to can isn’t right.

There’s a fair bit in here that wasn’t in the previous pages. safe_delete wasn’t introduced before, nor were create_new_salt and encrypted_password declared private.

The code for safe_delete is already shown on PDF page 151, so the user may have typed this in already meaning that they wouldn’t be able to delete the last admin.

“rendering is the one that was just changed.” -> “rendering is the one that just changed.”

"If it comes from an link, we

The first footnote in 12.1 could benefit from some pointers as the reader may also be interested in reading more about the mentioned four topics, namely SOAP/XML-RPC client/server.
So we could have a few page numbers here.

(Dave says: It’s likely we’ll be losing the web services chapter…)

‘l33t’ on the 5th or 6th line. Perhaps it is just personal opinion, but l33tsp34k(leetspeak) doesn’t exactly reek of professionalism. I wouldn’t use it in an official email, much less in a book; but that’s just me. :)

(Dave says: ah, I don’t write that kind of book :)

Section “The Desktop”, final paragraph, fourth line from the bottom: s/can can/can/

Replace the “wicked which” in the next-to-last line: “a Ruby library that conntects”.

Line 5 from the bottom: s/50 Mb/50 MB/

The footnote begins with “Rather that”

It should be “Rather than”

The spacing in the text below the screenshot is off. I see “The cart’s inthe sidebar” when it should be “The cart’s in the sidebar”

“with an object names ”order." " should be “with an object named ”order." "

depot> script/generate model user
should be
depot> ruby script/generate model user

In the list of editors, jEdit would be a good choice to include. With the addition of its SuperAbbrevs and Console plugins, jEdit can do a lot to help Rails developers be more productive. See, for example, http://saimonmoore.net/permalink/jedit-for-ruby-rails-development for details.

The locomotive link runs off the right side of the page and is outdated. I’m using Preview under Mac OS X 10.4.6.

“a use at a browser interacting” -> “a user at a browser interacting”

After test code, first sentence, "Here we

You may want to add a footnote about the text "It isn

While I don’t know if you’ve got plans to change this section I think the “You might be able to use the -s option when you run breakpointer to connect to an application on
another machine.” is very vague. It would be nice to give a better explaination of why you “might” be able to.

“… sidebar on this public-facing pages.” should be “… sidebar on the public-facing pages.”

One extra period in sentence "We can implement one using Ruby

permentently => permanently

first line: appplication

Hashes and Parameter Lists

Add a converse line of code to show a call made with a hash instead of showing one without.

(Dave says: this call already shows a hash being passed)

Perhaps it might be worth mentioning that digest/sha1 is required for the snippet to work. Otherwise, users may overlook the require in the code listing later on.

Shouldn’t the database columns be set NOT NULL? After all, you do perform a validates_presence_of later on. (Same goes for all later migrations and database tables too.)

(Dave says: possibly, but I wanted to keep it simple for now)

About the reference to Rails Recipes: it’d might be interesting to note why a reader would also consult RR when a fine authentication recipe is provided here already.

(Dave says: because RR also has a role-based recipe)

It might make sense to rename the login action in user.rb to authenticate. Technically, no session login occurs and it also prevents confusion with the login action in the LoginController later on.

Aracho Ruby -> Arachno Ruby

The sentence “It you now look at the database…” should read “If you now look at the database…”

A previous errata already identified the typo for the session create command to correctly be

rake db:sessions:create

However, this only creates a migration file and does not automatically create the session table as the book implies. The session table only comes into existences once

rake db:migrate

is executed.

The cursor changes to a pointer finger when mousing over the top third of the page. The underlying link reference is invalid.

(Dave says: Could someone email me a screenshot of this problem: I just dan’t see it here. dave at pragrog.com)

… if you’re decided to rebel …
should be
… if you’ve decided to rebel …

… once you’ve go Rails …
should be
… once you’ve got Rails …

… so that in looks like the following …
should be
… so that it looks like the following …

The explaination of how a partial assign an element of a collection to a variable is very confusing.

"That

Shouldn’t we see the code that actually uses the safe_delete helper at this point?

The command “rake db:migrate” is failing with this error:

No such file or directory - db/schema.rb

Is it possible that we have to do “script/generate migration” before running that command?

BTW,
marklar:~/src/learn/agilebook/depot max$ uname -a
Darwin marklar.local 8.6.0 Darwin Kernel Version 8.6.0: Tue Mar 7 16:58:48 PST 2006; root:xnu-792.6.70.obj~1/RELEASE_PPC Power Macintosh powerpc

(Dave says: this is probably a database configuration issue, as it seems to work fine here)

the command “ruby script/generate scaffold product admin” terminates with the following error:

No such file or directory - script/../config/../app/views/layouts/admin.rhtml

The scaffolding seems to work, although the CSS stylesheet doesn’t seem to exist. I’ve been building the project by hand and there’s not app/views/layouts folder

(Dave says: the ‘rails’ command creates it)

… for an array of atems. When …
should be
… for an array of items. When …

Section 8.3 is called Iteration C2 and Section 8.4 is also called Iteration C2

on the footnote where it reads “application from Chaptera 4.” it should read “application from Chapter 4.”

At the very botton, it should say “…open up index.rhtml in the app/views/store directory…”

see typo “…on the transition between and cart that’s empty”

In the first paragraph, the sentence “We’d use it in the store layout like this.” coupled with "For now, we’ll keep

Perhaps use ActiveRecord::Errors.default_error_messages[:taken] instead of “has already been taken” in the following line of code:

assert_equal “has already been taken”, product.errors.on(:title)

Footnote says: “Rails provides good support for generating
selection lists in this context too”. Maybe an example?

Second paragraph: There’s a spurious space character between the reference to “Figure 5.2” and the following comma.

Third paragraph: There’s a spurious space character between the reference to “Figure 5.3” and the following fullstop.

The list of possible errors of “rake db:migrate” should be extended for the message “No such file or directory - /tmp/mysql.sock”. On OpenSUSE (10.0), this file is /var/lib/mysql/mysql.sock. After setting a soft-link to /tmp, the DB test succeeds. (Answer found in the ROR wiki.)

I guess that your customer will bark at the “integer test” for the price after having to enter, say “2395” instead of “23.95”. How could she let this slip through Iteration A3?

(Dave says: she was probably drunk)

First bullet point: We created a development database […] to access it.

The two links that appear on this page cross the margin right border. Also, the first line of the 3.4 section has grown to more than the width of the complete text area.

(Dave says: all layout issues are address just before final printing)

Undefined acronym “RESTful”

Also, ???
- It turns out the best way
+ It turns out that the best way

duplicate word ‘on’
"We

Footnote: Should “assert_directed_to()” be “assert_redirected_to()”?

Did I miss the step where DAVES_DETAILS was defined?

“Now for the template itself.”
=> “Now, for the template itself.”

“Maybe we want to edit an existing model. Or maybe the user has
tried to”

The sentence just ends and the next line begins mid-sentence.

“The first parameter, :order tells…”
=> “The first parameter, :order, tells…”

Section 15.1 - “allow them to be send” should be “allow them to be sent”.

“so after freezing your application …” should be “so after freezing, your application …”

"Let

In ‘Joe Asks…’ section.

"Joe

“If the save succeeds, we delete the cart from the session ready for the next order and redisplay the catalog, using our redirect_to_index( ) method to display a cheerful message.”
=> “If the save succeeds, we do two things. First, we ready ourselves for this customer’s next order by deleting the cart from the session. Then, we redisplay the catalog using our redirect_to_index( ) method to display a cheerful message.”

“If instead the save fails, we…”
=> “If, instead, the save fails, we…”

(Dave says: much better wording. Thanks!)

Could you elaborate why components are considered deprecated? As a former Rails 1.0 user and buyer of Agile Rails 1.1 I am quite interested in the changes of the framework.

(Dave says: I’ve added a section to Action Pack)

As I figured out the rake-rdoc is not generated for code living inside the lib/ directory. Maybe you could explain the reason for that (or give a solution to it)?

(Dave says: It qwill include lib/ as of 1.1.3)

Is there a way to make classes or modules reloadable in development mode if they are located in the lib/ directory?

(Dave says: I believe not)

In the rhtml:

Time to say
<%= link_to “GoodBye!”, :action => “goodbye” %>

<<link_to “GoodBye!”>> s.b., <<link_to “Goodbye!”>>, to match the (correct) spelling elsewhere.

- Overspecify it upfront
+ Over-specify it up front

(Dave says: I’ll let the copy editor worry about this one :)

In figure 6-1,

- database: ttt_development
+ database: depot_development

- …you shound

- It should be the same as the name of the database
you created using mysqladmin.
+ It should be the same as the name of the database
you created using your database administration tool.

This sentence got scrambled with another:

If you see the error Before updating scaffolding from
new DB schema, try creating a table for your model (Product), it may well be because Ruby (and hence Rails) can

Footnote: s/ad nauseum/ad nauseam/

The reference to “Figure 7.1” has a spurious space character before the following fullstop.

The reference to “Figure 7.2” has a trailing space character before the following fullstop.

In the footnote, you said “Remember that we’re storing the price in the database in cents”, but I didn’t find any reference to that election before this page.

(Dave says: It’s in the sidebar two pages befoee)

The footnote is splitted in two pages. Probably the best way is to have all the lines in the same page.

(Dave says: we do layout at the end)

Normally, this doesn

If you didn’t stop the webrick server after adding “scaffold :product” line at the app/controller/admin_controller.rb file, the new list.rhtml view won’t be rendered, because webrick seems to use still old scaffold functions.

(Dave says: seems to work OK for me. Does anyone else see this?)

“Rails complains. Let’s make it happy by writing…”

In "So far, the only code we

First paragraph: “adds some hhelper methods to strings” - “hhelper” should be “helper”

Last paragraph, first line. “…,you want want to consider…” should read “…, you may want to consider…”

Last paragraph, third line. “…, six months from they, …” should read “…, six months from now, …”

In section “Linking Your Project to the Rails Repository”, the last word on the line just below the code example “svn up vendor”: “From now one, …” should read “From now on, …”

Third paragraph in “Runtime Environments” section (and following command-line examples): "If you

The reference to “Figure 8.4” has a trailing space character.

Another trailing space after “Figure 8.5”.

…“The scipt.aculo.us effects library”…

sc*r*ipt.aculo.us

duplicate then “It then then reopens the PaymentGateway” 7 lines from the end of the page.

The unexpected line break looks like you were looking at my suggestion for the previous edition:

“However, consider the general case. Maybe we want to edit an existing model. Or maybe the user has tried to enter an order, but their data hase failed validation. In these cases, we want any existing data in the model shown to the user when the form is displayed.”

This section required a number of re-reads before I understood that it was talking about auto-populating the form with existing model data. To me it would make sense that you introduce the concept and then explain why it is useful:

We also might want to auto-populate a form with existing model data. There are a couple of reasons for this. Firstly, we might want to create a model with default values, and have these values visible when the form first loads. Secondly, the user might have tried to enter an order, but their data failed the validation procedures, and on reloading the form we don’t wish to lose the information they have already entered.

My suggestion was in the errata for the previous version, yet I can’t see any change, hence I’m suggesting it again. Ignore it if you think it’s too pernickity!

ust a consistency point.

In add_to_cart.rhtml you use the variable `cart_item` as a placeholder for all the cart_items in the cart:

<% for cart_item in @cart.items %>

yet on page 99 you use `item` as the variable name:

<% for item in @cart.items %>

Is there any real reason for this? If not it might be nice to keep them the same - either one is nice, although I personally prefer the more verbose `cart_item`. :-)

(Dave says: I don’t see this: it seems consistent (cart_item) here. Please e-mail me if I’m missing something)

Just after, the hide the

I did not get the exception, “depressingly illustrated in Figure 8.2”, instaed everything just worked. The old session was apparently already deleted when restarting application.

(Dave says: it wouldn’t be. Perhaps you hadn’t previously added something to the cart, in which case the session wouldn;t have out-of-date data)

In the “Assertions” list, this assertion appears:
assert_response(type, message=nil)
For consistency, this should probably simply read:
assert_response(type, message)
… with “message” in italics (like the other assertions in the list)

In line:

redirect_to(session[:original_url] || { :action => “index” })

“:original_url” should be “:original_uri”

line:
redirect_to_to(:action => :list_users)

should be:
redirect_to(:action => :list_users)

In the add_product listing there is one extra ‘end’ which should be deleted

The RJS example should do a conditional page.hide on the flash #notice div if it exists, otherwise you add something to the cart and the “your cart is empty” notice is still there.

Also it would be a good idea to discuss the fact that offloading the add_to_cart functionality into an RJS file breaks the app if javascript is turned off - instead of degrading gracefully, it renders the javascript onscreen in the browser.

(Dave says: See later for both…)

… lets revisit out layout.
to
… lets revisit our layout.

The current CSS file in both the book and source files has the depot-form id set to classes, as in the following:

.depot-form fieldset { … }

The source code that is generated gives the depot-item div an id, not a class. I had to go and change the CSS file to reflect that the div contained an ID and not a class.

(Dave says: I believe the latest edition has them as classes (except for depot_r’ add_user form, which I’ve fixed)

The code for the links to the Previous page and Next page is equivalent to, but different from, the scaffolding code that’s generated automatically. It should match, unless there’s a good reason not to (which should be mentioned in the text).

(Dave says: I prefer to have shorter lines when I type stuff manually)

Footnote 9, to the images, results in a “Forbidden” page—I didn’t have access to the directory.

The find_products_for_sale method has “self.” in front of the method name, which is different from all other methods up to this point. Either it’s extraneous/wrong, or it should be explained.

(Dave says: the annotation in the margin next to the example points to the explanation)

In 2nd paragraph you give example subdirectory lib/pdfstuff; after 4th paragraph the example of importing from that directory has an underscore: require “pdf_stuff/receipt_writer”.

* puts 20.minutes.until(“2006-12-25 12:00:00”.to_time)*
I think that the word “until” should be blue.

(Dave says: yes, it should, but our syntax coloring tool ain’t that smart… :)

About the footnote: it might be useful to elaborate why a reader should also take a look at Rails Recipes. Other than a shameless plug, what is the advantage on a technical level?

Dave says: because it show role-based authenitcation, too)

Halfway the page, you mention unit and functional tests. It might be useful to also prepare the reader for the bit about integration tests.

Not actually a typo, but an ending colon after “The first is the line” might be appropriate.

“Let’s write the full test” misses an ending colon.

In the listing of depot_q/app/controllers/application.rb seems to be missing a return “false”:

class ApplicationController < ActionController::Base
private
def authorize
unless User.find_by_id(session[:user_id])
flash[:notice] = “Please log in”
redirect_to(:controller => “login”, :action => “login”)
false # <—- HERE
end
end
end

Otherwise, the rendering of the page continues and and error is thrown

(Dave says: From the documentation: "Additionally, it

When talking about foreign keys, suggesting this [http://www.redhillconsulting.com.au/rails_plugins.html] would be a good idea i think. Not sure why it isn’t part of rails anyway?

(Dave says: I don’t like the way that plugin automatically adds constraints: I’d prefer always having to use :references. As to your second point: I agree)

An alternative to using package management on Linux could be ROROX (http://rubyforge.org/projects/rorox), which puts Ruby, Rails + dependencies + useful gems over a XAMPP (http://apachefriends.org/en/xampp.html) installation. It supports running Rails under SCGI within subdirectories of localhost, and supplies scripts for doing the necessary configuration. (I maintain the package, by the way.)

(Dave says: this is very cool. I’ll include it if it seems to have traction in a year)

The testdata contains para tags which show up in figure 6.5 Tidied up Pruduct Listing (p.85).

(Dave says: Tht is correct.)

The rest of book uses double instead of single quotes. For reasons of consistency it may be appropriate to stick to a single notation throughout the book.

“either Comma-Separated Value (CSV) format or YAML format.” The first “format” is superfluous.

“it’s title” should read “its title”

“compart” should read “compare”

"Remember that underneath the covers, button_to( ) generates an HTML

“This form is different to ones we’ve seen earlier” should be something like “This form is different FROM ones we’ve seen earlier”

the linked list_users.html is much different than the one in the PDF.

(Dave says: sorry—that code didn’t get reuploaded)

call me a dummy, but when i saw that we were writing a “helper ” (safe_delete), i headed straight for the helpers folder. is “helper” a bad word choice here?

“These Rake tasks take a version … and freeze them …” Is that grammatically correct?

Last line of first paragraph plus footnote: “partial templates” and the associated footnote should be reformatted. I guess, what you want to have is something like: “\\textit{partial templates}\\footnote{Another way […], see section \\ref{sec.dont.use.components}, on page \\pageref{sec.dont.use.components}.)}.”
(a) The fullstop after “partial templates” should go /after/ the \\footnote insertion.
(b) “, on page …” should go /into/ the \\footnote.
© The final sentence of the \\footnote should get a closing parenthesis.

(Dave says: this is an artifact of the formatting system: it doesn’t like missing cross references in footnotes :)

Last line of the third paragraph from the bottom: Remove the space character between “HTML” and “.”.

Trailing space after reference to “Figure 10.1”

How do I add the price column in a different order (e.g. after the title)?
I usually expect that one book can’t answer all my questions, but I really appreciate it when the author(s) add pointers to other references with more details, so I don’t have to spend days (sometimes weeks) looking for answers to questions I might have after reading a book.

(Dave says: I don’t believe you can. I also don’t really see why it would matter..)

version of our code,x too.

Running this code gives me a “NoMethodError in Login#add_user”, undefined method ‘name’ for #
If I delete the <%= form.text_field :name, :size => 40 %> line from my code, the form gets at least loaded…

(Dave says: that means your users table doesn’t have a column called name)

Two lines above the second code inset: s/eihter/either/

If you have added a product that matches the name of one of those in add_test_data, it won’t be updated when rake db:migration is ran. I used the title from the example back on page 70 when I was told to add a product earlier in the book. I think a little more detail on the add_test_data and migration process works would be very helpful here (or at least a reference to where to find detailed info on migration).

(Dave says: isn’t this covered when we talk about vlidation: the validations we wrote check that you cann’t have duplicate names..)

has_many :orders, :through => :line_items
end

should be:

has_many :orders, :through => :line_items

1. ….

In index.rhtml:
<%= product.price %>
needs a:


inserted after it to get the same layout as Fig. 7.2

(Dave says: Not if you use the supplied test data, as it has

tags around the description)

2nd paragraph: “Because we added the belongs_to
declarion to LineItem” should be “Because we added the belongs_to declaration to LineItem”.

The two “cart.add_product rails” statements can be DRY’d by a little Ruby magic like “2.times { cart.add_product rails }”

(Dave says: They could, but I think it’s clearer laying them out in this case)

It could be real useful to provide a listing of Active Record’s built in error message table. I couldn’t find it documented anywhere but by looking at the source code.

“built in” should read “built-in” (look who’s talking :)

“…waiting for us in the test/functional directory.” Readability might be improved by ending the line with a colon instead.

About the “setup()” notation: many Rails developers are new to Ruby and might be struck by this notation. After all, where are the parameters in the method definition? Just writing “setup” might be less ambiguous and not necessarily wrong. (I’ve seen you use this on page 189 too.)

In the sidebar "What

change “thhow” to “how”.

“The assert_response() … was successful” misses a closing dot.

“logged-in” should read “logged in”.

“The parameters are as follows.” I know it’s a matter of style and I’ve whined about it previously, but it might be more correct to end that line with a colon instead of a dot.

(Dave says: I tend to leave these to the copy editor, but I’ve changed this (and others))

“functional-test assertions” should read “functional test assertions”.

I don’t think there is such a thing as “a browser field”. There’s a though, so perhaps this could be worded better.

“mime-encoded” should read “MIME-encoded”.

The “dave = users(:dave)” statements can be DRY’d in the setup method.

(Dave says: they could, but then I’d have to show it… ")

On page 207, you instantiate the products fixtures as ruby_book and rails_book. It would probably be more consistent and less ambiguous if you would here as well.

“but its still pretty readable”: “its” should read “it’s”

“integation” should read “integration”

A note on consistency: here you’re referring to the symbol :index whereas all your previous code uses the string ‘index’ for assertions.

(Dave says: let’s sneak in both uses… :)

Second line of next-to-last paragraph: “you enter the URL” — only one “the”

Since the source files are named dsl_ as well, it might be nice to note that the “minilanguage” in fact is a domain-specific language. Also, it might be more consistent to write either “minilanguage” or “mini-language” but not both (e.g. page 211).

Extra space after reference to “Figure 12.2”

Should “Which we do in the Pragmatic Programmer store” read “Which we have done for the Pragmatic Programmer store”?

“or WEBrick, We’ve been using this…” should read something along the lines of “or WEBrick, which we’ve been using…”

“Rail’s” should read “Rails’”

Now that Rails has a stable branch, linking to that one might be another option worth mentioning.

“Figure 6.1, on page 62 shows…” should probably read either “Figure 6.1 on page 62 shows…” or “Figure 6.1, on page 62, shows…”

Continuing my nitpicking :) - “dbfile” under “development_sqlite” is aligned with the entry above it, but no such alignment is present at “development_mysql”. Also, it might be more correct to end the line “Environment files typically do three things” with a colon.

“Then the following happens” should probably end with a colon (and I’ll now quit my enduring whining about colons by noting that lines signaling the start of a bulleted list should generally end with a colon).

in the last sentence of the paragraph following the graphic, change “tart” to “start.”

context:

Let

Page 213, last sentence: If the response is a redirect, follow it and any subsequent redirects, until a response that isn’t a redirect is returned.

First paragraph of section P&B: Use an em-dash (—-) in the next to last line, please.

2nd paragraph - “it simply tells your each of your application…” should probably be something like “it simply tells your application…”

Hiding the cart as directed did not work for me in IE7, while it did in Mozilla. Normally I wouldn’t have such a blind eye, but the problem occurred exactly as described on page 131, which led me to believe I had coded wrong and wasted time debugging: “… we see a flicker in the browser as the cart is first displayed, and then hidden and slowly revealed by the blind_down effect.”

(Dave says: this is an issue with IE quirks mode, and is fixed by including the correct DOCTYPE in the layout (which I didn’t have)

user_test.rb is missing from the unit listing and login_controller_test.rb is missing from the functional listing.

There’s a missing closing parenthesis near the bottom of the page: change
“(saying something like (in /Users/dave/work/depot), or”
to
“(saying something like (in /Users/dave/work/depot)), or”

The sentence starting with “How often have you implemented…” should probably end with a question mark.

“Figure 16.1 , shows”: there’s a space leading the comma, but I believe the comma can be omitted.

“Or maybe the user has tried to enter an order, but their data hase failed validation.”
=> ‘hase’ to ‘has’

End of page…

"Before we start on our final form, let

End of page…

"you

For consistency throughout text…

“…dropdown…”
=> “…drop-down…”
(Also in the footnote)

“So Where’s Rails?”
=> “So, Where’s Rails?”

(Also change linked-to section header).

Perhaps a footnote that the \\ character in the generator statement is a shell command. Thought it might be part of the statement at first, which happily created a “\\.rhtml” file and a “\\” method in the login controller to no ill effect until several pages later.

Perhaps explain the presence of “wibble” in

string_to_hash = password + “wibble” + salt

which seems obscure.

You suggest the convention for migration names as follows:
depot> ruby script/generate migration add_price

But that doesn’t indicate which table we are adding the price column too.

Wouldn’t it be better to use:
depot> ruby script/generate migration products_add_price

so that the name is more descriptive of what is actually happening?

(Dave says: yes, that’s a good convention)

In IE6, the default display of the red borders around non-validated controls does not display correctly… the right hand red border is spread right across the page showing up as a big fat red block, instead of just a simple border. The stylesheet shows:

.fieldWithErrors {
padding: 2px;
background-color: red;
display: table;

Which isn’t interpreted by IE6 very well.

It may be beneficial to show a work around / fix for this, or at least make a comment about it so we don’t feel like we are heading down the line of a non-IE friendly web development platform.

(Dave says: this sounds like a bug in Rails: perhaps it should be reported to the core team)

Downloading depot.css is suggested, but I think it should mention changing the layout to use this new stylesheet rather than the scaffold stylesheet.

(Dave says: that’s on page 84)

Section “lib/”: s/fit fit/fit/

Second text paragraph: Replace the en-dash (—) with an em-dash (—-) in the fourth line: Gems—-it

(Dave says: good eye!)

First paragraph: “… for Rails’ internal use.”

Instead of

On the ‘Beta Book’ information at the very start of the book, paragraph 4:

In the meantime, we

Near bottom of page: "using LighTTP (if it

The logo.png image is not on your server to download.

redirect_to_index() in store_controller.rb doesn’t seem to work after each “Add to Cart”. “Add to Cart” does not update the page automatically. However, “Empty Cart” seems to work fine.

(Dave says: it seems to work here. Could you try it with the downloadble code and see what happens?)

The image of the web page with the “Add to Cart” button shows the button adjacent to the item price. However, Rails renders the button with a

around it, so it’s going to appear below the item price (which is how it looks in FireFox 1.5.0.3 on OS X 10.4.6).

(Dave says: this is how it looks in Safari. It may look different in your browser. Got to love standards…)

This bit of code line doesnt work..

page[:cart].replace_html :partial => ‘cart’, :object => @cart

It works only if this format is used.

page.replace_html ‘cart’, :partial => ‘cart’, :object => @cart

(Dave says: could you update to the latest Rails and let me know if it is still not working?)

Middle of page sentence is missing ‘that’: “Now we have our tables and our models,” should be “Now that we have our tables and our models,”

(Dave says: I leave this kind of thing to Kim, our copy editor..)

Re “Making Development Easier” (auto-reload in development mode):
It doesn’t always work! In particular, it appears that it you have a session open and modify the Cart class code, the edits don’t take effect. This is probably obvious to the experienced, but the tutorial in the book runs straight into this problem.

(Dave says: but, it then shows you how to fix it… :)

“it’s built in” should be “its built-in”

“if not it adds” should be “if not, it adds” to keep parallel structure with the beginning of the sentence (“If so, …”)

“Ruby book’s price” should be “the Ruby book’s price”

When I got to page 141, my code (see p.104) had the “current_item” object called “existing_product”. So it appears you have changed from calling it existing_product to current_item.

(Dave says: that’s right—it changed name because it serves a different purpose now)

Near the end of the 2nd paragraph of the sidebar, “They technical term is idempotent…” should be “The technical term is idempotent….”

foot-note “We’re tried” -> “We’ve tried”

The delete_user method as-is results in an error:
“undefined local variable or method `user’”

I found that adding a line:
user = User.find(id)
before the “if id && user…” line solved the situaton for me, but I’m uncertain if this is the correct fix for this problem.

(Dave says: my guess is you mistypes the if statment to be "if id && user == User.find)

Section 8.3 jumps into creating the CartItem class without describing where the class file belongs. (Though its location can be inferred from the link to the download of the file.)

The assertion “test_add_duplicate_product” fails for the price.

the code:
assert_equal 2*rails.price, cart.total_price #failure?
error message:
<20> expected but was <10>

I thought this is may be because the inject method is not invoked properly and changed the assert to verify total items, but this passes the test.
“assert_equal 2, cart.total_items #cart.items[0].quantity”

Any idea what could be the error?

(Dave says: It’m afraid it works fine here. Bring it up under script/console and check cart.total_price. Also, make cure that cart_items.price is correctly defined)

WHY are you using SQL in your migrations for foreign keys? This seems like a very non-RoR thing to do! ;o) How about this instead:

http://www.redhillconsulting.com.au/rails_plugins.html

(Dave says: I don’t like the defaults of this plugin, but I do plan to plug it in the new Migrations chapter).

The integration test runs fine, but I couldn’t find the (new) order in the database? Does the test deletes the entries after running them?

(Dave says: all tests do a rollback at the end)

When I run the performance test I get the following error:

NoMethodError: undefined method `stringify_keys!’ for #

Think this is something to do with MySQL, any suggestions?

(Dave says: hmm.. seems to work here. Perhaps try copying the source down from the web site and see if that runs for you)

Formatting of the store/index page is incorrect. The dotted bottom border of the

"If you refresh your browser, you should see no dif ference in the page that

Is the comment line “# …” in the listing for any particular reason?

In the sidebar “A Friendlier Login System”, the change you make using session[:original_url] does not match the above session[:original_uri]

(Dave says: I believe they both say _uri)

I made the mistake of adding the columns to the 001_create_products.rb file but not saving the file before running “rake db:migrate”. The result was the products table only had an “id” column. Saving the file and re-running “rake db:migrate” didn’t add the columns. I fixed things by running “rake db:migrate VERSION=0”, followed by “rake db:migrate”. Seems like a common enough error that you might want to mention it.

Oops just read the page numbering note.

ignore #3813:

1st paragraph, 3rd line
currently -> this content a many
change to -> at this content many

I get a different error from the 2 solved around page 118:
NameError in Store#add_to_cart
uninitialized constant Reloadable
/usr/local/lib/ruby/gems/1.8/gems/activesupport-1.2.5/lib/active_support/dependencies.rb:200:in `const_missing’
#{RAILS_ROOT}/app/models/cart.rb:11

where do i go to fix these problems?

(Dave says: update your version of Rails)

for windows environment, need to specify -u option as follows:
depot>mysql -u root depot_development

Otherwise, the following error will result:
… Access denied for user ‘ODBC’@localhost’ (using password: NO)

Ph 106 says, "[The client] points out that we

In the last paragraph you tell us to clik “Checkout” button, but that should be “Place Order” button right?

On my system, the flash message “User dave created” still hangs around above the Welcome message when I point at http://localhost:3000/login. How can this be removed?

(Dave says: I don’t see that problem)

I think you wanted a space before the “×” in the line in add_to_cart.rhtml: "

<%= cart_item.quantity %>×

At the top of the page:

"the main page doesn

In the section “Locomotive Mac Installation” the last line of the third paragraph states:

Access it from the Rails > Open Terminal menu option.

it should read:

Access it from the Applications > Open Terminal menu option.

The definition of total price invokes a method on CartItem that does not take into account the quantity, it should be changed to:

def total_price
@items.inject(0) { |sum, item| sum += item.total_price }
end

and a total_price method should be added to CartItem:

def total_price
price * quantity
end

(Dave says: have a look at the existing CartItem model…)

in the first listing you have
dave[code/depot_q 19:20:09] script/console

which should include the ruby command and for consisency with previous command line examples should read

depot> ruby script/console

Since we have already written the delete_user action suggest:

And then write a delete_user action in the login controller.

should be

And then modify/re-write the delete_user action in the login controller.

Add after the line

session[:user_id] = nil

a line to remove the original_uri in session also, otherwise they don’t always get redirected to login/index to see the orders.

session[:user_id] = nil
session[:original_uri] = nil

In the models/order.rb code snippet, I had to put the “validates_inclusion_of” line after the definition of the “PAYMENT_TYPES” array, to avoid a run-time error. If that is indeed necessary, there should be a mention in this area.

Another hint if the Ajax magic doesn’t happen: - Did you remember to delete the add_to_cart.rhtml file?

It is probably a formatting error. Footnote 6 is truncated.

6A newly created Rails application enters the world containing 44 files spread across 36

(Dave says: (it continues on the next page, but we’ll fix that during layout)

Suggest then that the 006_create_line_items.rb migration be ammended to cater for this situation during forward and backward migrations.

The following will ensure an “admin” user is inserted with a password of “password”

class CreateUsers < ActiveRecord::Migration
def self.up
create_table :users do |t|
t.column :name, :string
t.column :hashed_password, :string
t.column :salt, :string

end

execute “INSERT INTO users (`name`, `salt`, `hashed_password`) VALUES ('admin', '982014440.257920467993245', 'c0d230ce90af6ed0e00b38961d5b76a39bbbe156')”

end

def self.down
drop_table :users
end
end

(PS: This may be deemed a security issue, so a better solution may be possible, such as allowing navigation to add_user if there are 0 users in the database).

(Dave says: added as an exercise)

After deleting my session information, the first time I navigated to http://localhost:3000/admin/list, I got to the list successfully. After forcing a refresh in my browser (FireFox 1.5.0.4, OS X 10.4.6), I did get to the login page—I think the first time, I got a cached page.

(Dave says: strange—I don;’t see that)

In the footnote, ‘layout’ is used as a verb (it’s a noun). Change to ‘lay out’ as in:

…to make them lay out properly.

Prices are stored as integers, right? So, if a product has a price of $56.43, the admin user would need to type in “5643”, right? Wouldn’t it be easier for the admin to type in “56.43” and have the model convert that into an integer?

(Dave says: yes, but that would muddy the explanation at this point. We’ll cover that in the second part of the book)

In the description of the cookies attribute:

“read values form the hash”

should be:

“read values from the hash”

“caLlout id” should be “callout id”

I’m just learning rails and ruby. Read book one or twice thru. I give it an A-.

Most of my problems when working on my first application involve understanding what gets passed into what automatically. And what I have to send in. In Java, this would be a pure scope issue. But with Rails doing a lot behind the scenes. It would be nice to understand just what information is available say from the controller to the template and from the template to the partial. How things get passed from a template into a partial. Granted I’m new to ruby as well. Some of my challenges have been when to use the @ sign or not. ;-). But the ruby scope is defined fairly clearly in the Pickaxe.

In summary, perhaps a scoping summary table of what’s exposed to what in the framework. And another table of how things get passed into helpers, templates and partials.

Most of the information is spread throughout the book. But not all of it in one easy place. Sometimes in the APIs, but not always consistently (for example, there is no mention of :object hash in the API for partials). If you know of one, let me know.

This knowledge has just about clicked for me (but only with some pain).

(Dave says: Rails has a lot more magic than Ruby, and it doesn’t really have tidy scoping the way RUby does)

Third-to-last paragraph: “All of the
testig assertions accept an optional trailing parameter…” (Notice “testig” != “testing”)

Once again, third-to-last paragraph: “All of the
testig assertions accept an optional trailing parameter containing a string.
This will we written out along with the error message if the…”

This should be: “…This will [[be]] written out along with the error message if the…”

Second paragraph first sentence. “seem” should be “seemed” because it is past tense.

(Dave says: hmm.. I meant it to be present tense. Let’s see what the copy editor says…)

It will be great to have an example on how to link two different tables associated using a lookup table. e.g. in the diagram 12.1, the arrows point outside.

(Dave says: there is a lot of data modeling stuff that I don’t really have space to document here)

Dont know if this might be intended, but in the source below “private” the find_cart function isn’t visible, mabe add it, or a comment eg. #rest of controller

"We

Line 3: “… We discuss builder templates starting on page 403.”

This seems disconnected to the rest of the paragraph: ‘Templates’ have been discussed so far, not ‘builder templates’. Either remove the word ‘builder’, explain the difference between ‘templates’ and ‘builder templates’, or explain why the name has changed.

A question I also asked on the forum: why implement things like safe_delete() and delete_user() when there are callbacks like before_destroy() available? The latter could potentially prevent a lot of footshooting in case one forgets to actually use our custom delete_user(). Are there any disadvantages here that I’m missing? I’d like to suggest a throwback to the first edition, where you indeed used before_destroy().

(Dave says: you can’t use before_destroy, because it isn’t transactional. You can do the test in the filter, and then someone else could delete another user, and then your own delete takes place)

There are code chunks showing how files should look in numerous places, I find it fairly irregular. Sometimes only the new change is listed, sometimes its the change and #rest of model and sometimes its the entire model. I think you should stick to one of these three (preferable “new code” + #rest of model/controller/view).

In the sidebar, 2nd paragraph, 4th sentence “…A GET request is defined by HTTP to be used to retrieving data:…”, ‘retrieving’ should be ‘retrieve’.

says: “…code base would have to do the same. And, if we were working on a team on this project…” which seems just a tad unnatural. I’d suggest “with a team on this project”.

Modification needed for MySQL4.0.13 for rake db:migrate

I received a rake abort, MySQL error when I tried rake db:migrate. Apparently the Create table statement in MySQL 4 is slightly different than MySQL 5 in that MySQL 5 has ‘ENGINE=InnoDB’ while MySQL4 has ‘type=InnoDB’. Next I tried:
rake db:migrate —trace
And identified the file in activerecord mysql_adapter.rb where I changed from ‘ENGINE=InnoDB’ to ‘TYPE=InnoDB’.

Now the products table is created successfully.

(Dave says: this sounds like a bug in Rails…)

About the footnote: it might be useful to elaborate why a reader should also take a look at Rails Recipes. Other than a shameless plug, what is the advantage on a technical level?

Dave says: because it shows role-based authentication, too

Roderick says: then that may be worth mentioning, because it now seems to come from nowhere

with add_to_cart.js visual effect only works when the item has already been added to the cart. If being added for the first time I get $(“current_item”) has no properties. I fixed it by surrounding the visual effect expression with an if statement. There is probably a better way to do this I imagine.

(Dave says: I’ve tried it here with both Safari and Firefox, and it seems to work OK. Have you tried using the downloaded code?)

Sometimes you use line-numbers in the code examples, making it hard to just copy-paste.
EG;
Line 1 def save_order

- cart = find_cart - order = Order.new(params[:order])

- order.add_line_items_from_cart(cart)
5 if @order.save

- session[:cart] = nil
- redirect_to_index(“Thank you for your order”)

(Dave says: that’s why we have a download link, and why you can download the whole chunk of code from the site. I add line numbers to help folks cross reference)

I believe this sentence, "Bruce doesn

The UnixUtils package referred to in the footer is old and unmaintained. I’d recommend pointing to the GnuWin32 project at: <http://gnuwin32.sourceforge.net/>

Specifically, the CoreUtils package contains the ‘tail’ program: <http://gnuwin32.sourceforge.net/packages/coreutils.htm>

I’m not sure if this is an error in my understanding. Near the bottom of page 99 when discussing find_cart you say:

“We can achieve that by creating a helper method,”

However, this method then goes on to be put in the store controller.

When i read the line and it said create a helper method, my mind immediately thought back to the previously discussed helper files - so if it is indeed a helper method i was expecting to go and edit a helper file for store. Instead, the method is put in the store controller?!

Is there any reason why you’re using the ERB <% -%> form on this page, when there doesn’t seem to be any potential problem with end of lines? You don’t use this form on page 80 for a similar example…

(Dave says: I didn’t write the code on page 80—the scaffold generator did)

Middle of page:
“OK. You Windows users are done: you can skip forward to on page 24. ”

Specify “forward to” what or omit “on”.

The yellow triangle indicators that show changed/added lines in the examples are hard to see in the pdf. Could they be made to stand out a bit more?

“You can render a partial from another template or from a controller
and the partial will return the results of rendering itself.”

You introduced partials as ‘a kind of method for views’. I really like this analogy, but I think you could perhaps detail the mapping between the two a little more explicitly. For example, you use the verb render, when perhaps it would make sense to initially use the word invoke, and then explain that invoking a partial is the same as rendering it. i.e. you invoke a method, but you render a partial.

Also, a very mild point. The phrase “..and the partial will return the results of rendering itself” is slightly ambiguous. It could mean that the partial itself will return the results of rendering, or that when invoked, a partial will render itself and return the results of that rendering.

The examples use .at_beginning_of_xxx except for one .beginning_of_quarter and there is no mention of the aliasing (i.e. all the methods are available with and without leading "at". Either mention the aliasing in the text or don’t confuse the reader by using a single method without the leading at.

“Within this block, you can put normal template stuf f (such as
the tag)”

I’m sure that most people will know why the tag is normal, and what you’re doing with it, but unfortunately that’s not me! Would it be at all possible to spare a quick sentence to explain what the tag is doing, and in particular what the “order_name” for field represents?

(Dave says: I changed the text to talk about

insted. I don’t want to divert into an HTML markup discussion here)

“We’ll simply copy the boilerplate from the product_test.rb file…”. Since product_test.rb has already been significantly modified at this point, wouldn’t it make sense to suggest copying the boilerplate from any of the other unit test files?

“…between multiple test cases…” -> “…among multiple test cases…”

Suggest changing “…these tend be be used infrequently when testing Rails applications, we won’t discuss them here…” to “…these tend be be used infrequently when testing Rails applications—we won’t discuss them here…”

The text on ‘created_at’ doesn’t tell me whether I should define it as a datetime or a timestamp. The docs for ConnectionAdapters::TableDefinition likewise don’t distinguish between these types, except that they both have names.

MySql has both types but Postgresql has only a few varieties of ‘timestamp’ (with and without timezone).

(Dave says: thanks for nudging me—it looks like you can store the timestamo into a date, datetime, or string)

"If the image path doesn

user.rb :

Which is the difference between “password" (instance variable) and "self.salt", "self.hashed_password"? Why not to use "self.password" or "salt”?

I think it is not clear enough.

(Dave says: @passwrod is an instance variable. self.password is a call to a method)

“DDL” is mentioned in the second to last paragraph, but no where in the pdf is this acronym defined.

Why is the model name “product” in lower case on the command line here, but uppercase on page 66?

Would prefer both to be lower case for uniformity, so the reader does not spend time doing research to find that it doesn’t matter.

This isn’t exactly an error, but on page 15, you generate the Login with several methods. You do not include the method index() among those, yet it is created in the pages that follow, even before delete-users, list_users, or logout are implemented, so it might cause a little less confusion if you included it on the command line.

The images repository linked from this page (URL below) doesn’t contain “logo.png” which is needed for the layout introduced on page 91 of the PDF.

http://media.pragprog.com/titles/rails2/code/depot_c/public/images

Re :null — “If true, the underlying column has a not null constraint added” — s/true/false/.

“For example, you might notice that once you application has a large number of
orders in the database…” s/once you/once your/.

“Rails takes responsibility for incrementing the value of this column for each new row added toa table.” s/toa/to a/.

The example migration to create the books_authors join table contains two errors: create_table :books_authors(:id => false) should not use parenthesis; it should read instead: create_table :books_authors, :id => false. Also, the column definitions are missing a data type (:integer?) before :null => false.

AddUsers.down reads “drop_table = :users” but I think should be “drop_table :users”.

It might not be obvious from the Forms Containing Collections side bar
that the select form helper in the following sections can be used in a
similar manner:

assuming that products has a product_type_id and that product_types has
a name column:

controller:
product_types = ProductType.find_all product_type_select = @product_types.collect { |type| [type.name, type.id] }

view:
<%= select (‘product[]’, ‘product_type_id’, product_type_select, :selected => product.product_type_id) %>

Product.update(params[:product].keys, params[:product].values)

would then work as with the text_field shown in the side bar.

“Install the annotate_models plugin using the following command (which has been spolit onto two lines…” s/spolit/split/.

OrderHistory.create(:order = order, :notes => “test”)
should be OrderHistory.create(:order => order, :notes => “test”)
note the hash after :order

Second paragraph, first sentence: “Let’s start by look at migrations at an abstract” should be ‘looking at’ or ‘examining’.

populate a table seem to not run under postgresql

code
User.create(:name => “yo”, :password => “#{Digest::SHA1.hexdigest(‘yo’)}”).save!

stacktrace
yo:$ rake migrate
(in /Volumes/Data/Repository/rst)
Admin: migrating =
— create({:name=>“yo”, :password=>“c41975d1dae1cc69b16ad8892b8c77164e84ca39”})
rake aborted!
undefined method `create’ for #

(See full trace by running task with —trace)

(Dave says: I don’t see this code in the book. However, create automnatically saves the record)

In the first sentence under the “Renaming Tables” heading, change “we sometimes fine ourselves” to “we sometimes find ourselves”

In the sentence just before the “16.6 When Migra…” heading, change “that authomatically handles” to “that automatically handles”

First para second line. Says: “Wouldn’t it me more convenient …”. Should be “Wouldn’t it be more convenient …”. Change me to be.

The second last para begins with “If you specify force => true, the migration will …”. I think it should use the symbol :force and not just force. So the correct line should be: “If you specify :force => true, the migration will …”.

The para just above the section “Primary Keys” towards the end mentions about “passing an array of colum names to create_index”. There is no other mention of create_index in the book. I think it should be add_index.

Last para just before the code which defines run_with_index. The last line begins as “This could be a private method in the model, or or could…”, has the word “or” twice.

Third last para, second statement: “To keep track of the schema as it evolves, you can uuse the..”. The word use is miss-spelt as “uuse”.

I realize the 2nd edition of the Form Helpers section hasn’t been released yet. Having read the 1st edition text, this seems to be a part of the text that I found a good deal less clear than the rest of the book. I would especially like to see addressed editing a many to many relationship with checkboxes, or the case where the join table is actually a model itself. Thanks.

(Dave says: I believe the Recipes book covers this)

After Ch 9, there is a minor issue with the flash[:notice]. After you enter an invalid item via an explicit url, the flashed message is never removed after new valid items are selected. It continually displays “Invalid product”.

(Dave says: great idea for an exercise)

I don’t recall if it has been suggested before, but I would love if the “Building an Application” section contained, at the end of each task or sub-task, a series of suggested exercises to extend the sample application in small ways to help reinforce the learning process. It’s not always easy when your learning a framework to think of ways to extend examples that fall within the limits of what you have learned.

Since there are 1000, not 100, products in the performance products fixture, should “…the block creates 100
orders using each of our 100 products…” be changed to something like “…the block creates 100 orders using the first 100 of our 1000 products…”?

“…customs documentation for shipping…” s.b. “…customizes documentation for shipping…”

D (Dave says: customs doc. is required for international shipping)

“…to include of PDF receipt writer…” s.b. “…to include the PDF receipt writer…”

Last paragraph: One scheme is to keep …

Second paragraph: they edit the existing migration

"databases just don

You created footnote 3 based on the erratum suggestion #3810, but I hope you might take another look at improving it.
You mention that runing migrate VERSION=0 is only for those feeling “frisky”, but I find that “rake migrate VERSION=0; rake migrate” Is much more convenient than dropping and recreating the database in another tool then rerunning the migration. I usually have a migration along the lines of 001_initial_setup.rb and as I’m specing out the app, I will add columns to it then run migrate VERSION=0 to drop all the existing tables then migrate to get my fresh schema.

(Dave says: I’m glad you experience is good, but othher find VERSION=0 very frustrating, particularly if they have had an error in a migration which leaves the database in an inconsistent state. My experience from teaching Rails is that it’s easier to drop and recreate the database)

About halfway down it says ‘foreigh_key()’ instead of ‘foreign_key()’

In the last sentence of the footnote “aplugin” should be “a plugin”

last paragraph before section 7.3, you mention the depot.css file, but the link is broken.

The closing > on all the labels are blue instead of red.

(Dave says: we must get cleverer code formatters… :)

The arrow to the line_items table does not have an arrow, but the arrow from it does.

“specific” is spelled wrong in “You may have
noticed that the Time class has a similar extension for formatting datetime
fields in a database-sepcific format.”

Where is says “(For strings, it first strips leading and
trailing whitespace before testing.)” many programmers will be left wondering why it strips both leading and trailing whitespace instead of just leading (or trailing) whitespace prior to testing for a string being empty? If there exist non-whitespace characters which can compose a string that is considered “empty”, what are these characters?

(Dave sys: because that’s what the code does… It calls str.strip.empty?, and strip removes leading and trailing whitespace. Agreed, it could be optimized…)

Bottom line: “This is less risky that having your project…” should read “less risky than” (p230 on page, p242 in PDF)

depot> ruby script/generate migration add_test_data
Create db/migrate
create db/migrate/003_add_test_data.rb

I think “Create db/migrate” should be “exists db/migrate”

> Should be:
> @cart = (session[:cart] ||= Cart.new)
>
> - —
> Peter Harkins - http://push.cx

Otherwise, page 115 won’t display, instead displaying an error that item is nil

(Dave says: it seems to work fine here. The code on 115 seems to set @cart when needed)

ActiveSupport::CoreExtensions::Date::Conversions::DATE_FORMATS.merge!(
:Ordinal => “%Y-%j”
:Ordinal —> should be :ordinal (lowercase)

for the validate method, change it to “if !price.nil? && price <= 0”

There is already a check for nill with validate_presense_of, and you’ll get the ‘must be positive’ message when it’s not needed.

(Dave says: try it without and see :)

"If it

third line of first paragraph in section 16.8 - Managing Migrations (page 271) - potentially misspelt as potenially

Just below the heading “Putting Sessions in the Database”, you have the following paragraph:

“”Rails makes it easy to store session data in the database. First, we need to create a database table with the correct layout. We

When I use the h function in list.rhtml as included in your file to download, it simply shows the

inside the page. When I take the h() out, then it shows the actual html?

http://dusty.name/rails/with_h.tiff
http://dusty.name/rails/without_h.tiff

Thought it would be easier to show you.

Perhaps I have something wrong elsewhere? Right now I’m following along and downloading the pages I haven’t purposely changed anything.

You can email me at pragmatic@dusty.name if you want more info on it.

(Dave says: that’s correct—in the product listing, HTML is escaped)

I find the book’s typeface difficult to read, and suggest you consider changing it to something that appears less “heavy.”

In particular, the letter heights are too similar, so a word like “edit” has the dot of the ‘i’ at the same height as the crossbar on the ‘t’. Even worse is the contraction “it’s” where the dotted ‘i’, crossed ‘t’, and apostrophe are all at the same height.

Also, the letter spacing is too close, so one letter’s serifs might be nearly touching adjacent letters.

Maybe the printed book will be easier to read. But I’m reading it with Adobe Acrobat 6 Reader at 110% magnification on a 17" monitor, and it’s very difficult. Even at 125%, it’s still hard.

Please choose a better typeface. Thanks for listening.

Mike

(Dave says: I’m afraid Bookman is a house style. We actually receive quite a few complements on it…)

Hi Dave, Can you provide a sample deployment scenario with lighthttpd as well as with Apache? There is lots of discussion on running RoR apps in Apache 1.3 and not in Apache 2.0…so some pointers will be great.

(Dave says: lighttpd deployments seem to be falling out of favor, so the book concentrates on apache/mongrel)

Having links to the full code is fantastic, except that I’m finding that the code displays VERY poorly. Something that would be very nice is if the code was highlighted, and formatted on a webpage with the ability to comment it. The commenting alone would be near priceless.

(Dave says: the problem is that this code gets refreshed every time the book gets built, and so comments on it might become out of date. I understand what you’re saying, and I’m thinking about how we could work it. In the meantime, I fixed that annoying content-type bug on the site…)

The text talks about creating the depot_development database, but does not say to create depot_test and depot_production (even though those entries are present in database.yml). Then the final paragraph says, “…and if you’ve created databases with the suggested names, you can now skip forward….”

There should be instructions to create the test and production databases.

Mike

Missing the space after depot> in the second command prompt example.

Follow-up to float: left on price — the formatting error is fixed when the “Add to Cart” button is added. The screenshot on this page (figure 7.2) cannot be accurate.

(Dave says: indeed—the stylesheet is the one file shared between all versions of the application, and the float: left was needed when the button is added).

“Constructs a url given a set up options.” -> “Constructs a url given a set of options.”

First full paragraph, "…code that

Aargh! It took me 10 minutes to discover why the refreshed browser page didn’t have alternating background colors, as stated in the first bullet. It’s not until page 85 do we learn how to use the depot.css file.

It would be helpful to note that the CODE in list.rhtml implements these features, but is incomplete now, and refreshing the browser won’t display everything mentioned.

I’ve been using the code found here: http://media.pragprog.com/titles/rails2/code/rails-code.tgz and I don’t see any documentation comments in the code…

For example, in depot_q/app/controllers/login_controller.rb I would expect to see some comments above the class and all def lines. I don’t see any. This would be helpful as it would be a good guide for proper formatting of comments to allow rDoc to work.

(Dave says: This code is included directly into the book, and having a lot of comments would make the text a lot longer. I should, though, produce a fully commented final version at some point)

The sidebar for adding a redirect to the original uri after
login is a welcome addition, but I note that it has one
remaining nit: the login action doesn’t clear session[:original_uri]. I suggest the following modest change to LoginController#login which cleans up the session before the redirect:

# Clear the original_uri before redirect
original_uri = session[:original_uri]
session.delete(:original_uri)
redirect_to(original_uri || { :action => “index” })

The delete_user action will perform the delete via a GET. It would be to encourage good HTTP method practices and wrap the body of delete_user in a request.post? guard:

def delete_user
if request.post?
user = User.find(params[:id])
user.destroy
end
redirect_to :action => :list_users
end

Of course, this means that users without Javascript enabled
can no longer delete users. On the other hand, if the page
gets scraped for any reason we won’t delete all of the users.

From what I understand, @content_for_layout is deprecated in favor of just plain yield()?

(Dave says: no—Marcel was somewhat overly enthusiastic in that blog post)

DRY.
When adding the id=“current_item” I’d suggest
<tr <%= cart_item == @current_item ? “id=\\”current_item\\“” : “” %>>
in case we later want to add any more attributes to the tr tag.

(Dave says: true—if we ended up doing that, we’d probably want to make the change you suggest. But as we don’t need it now, I prefer the clearer code)

“Building Your Application to a Gem Version” callout, second paragraph “When you application starts” s/b “When your application starts”

In the first paragraph: “This works as long as your model classes do not contain any addition functionality” s/addition /additional /

In paragraph 8: “It then decrements the version, looks for the matching files, undoes it…” The case of “files” and “it” disagree.

The 17.3 section: “Primary Keys and IDs” section seems redundant with the side-bar on page 262: “Why Use These Artificial Primary Keys?”

Deleting Rows.First Paragraph… "An exception is not thrown if the row doesn

Bullet #2 is pointed to twice in the body (by the paragraphs discussing the database and username/password fields). The second instance (discussing the username/password field) should point to bullet #3.

The color of the server output is highlighted in Ruby style - the text following a ‘#’ is green.

(Dave says: yeah—that’s hard to fix…)

missing closing tag:

(Dave says: it split onto the next page. That kind of thing gets fixed just before we go to press)

In the footnote: “(the standard warnings APLY)”

You mention that using migrations causes the problem of potentially having your database definition for any table spread over multiple files. One thing to note is that the schema.rb file contains the ruby ddl for each individual table all in one place in alphabetical order. If I’m doing a lot of model work I keep it open all the time. This also holds the definition of non-model tables; I’ll let you decide if that’s an advantage or a disadvantage. The plugin sounds cool, I’ll have to give it a go.

It seems a bit strange to have two download links for the same store.rhtml layout file on the same page. The first link seems a little premature since the discussion that follows the flash code snippet motivates the placement of that snippet within the layout code.

It would be nice to know which rails-version is required for respond_to.

(Dave says: the book assumes the latest rails (1.1+)

… Alter natively, specifying
add_column :orders, :placed_at, :datetime, :default => “now()”
sets the default to the MySQL now( ) function

This doesn’t actually work, at least with MySQL 5.0.18:

— create_table(:payments)
rake aborted!
Mysql::Error: Invalid default value for ‘date’: CREATE TABLE payments (`id` int(11) DEFAULT NULL auto_increment PRIMARY KEY, `user_id` int(11) NOT NULL, `amount` int(11) DEFAULT 0, `date` datetime DEFAULT ‘now()’) ENGINE=InnoDB

“if msg” is not really necessary in redirect_to_index. A null flash :notice will not display in the store’s layout.

(Dave says: true, but I still like to be explicit here)

Echoing the previous comment, there should probably be an arrow pointing to the added “has_many” line in the Product class, or the remainder of the class elided.

End of second last paragraph:

possibly -> possible.

I would second the suggestion for a blind_down on empty_cart. empty_cart looks clunky compared to the blind_up’d add_to_cart. Hopefully the reviewer’s suggestions shouldn’t require too many changes to the existing text as no new concepts seem to be introduced.

(Dave says: great idea for an exercise…)

I like the idea suggested by erratum #4023 for clearing the original uri, but the code presented in the erratum does not work for me. The session object does not appear to offer a single-argument delete method. I find I have to do something clunky like the following:

original_uri = session[:original_uri]
session[:original_uri] = nil
redirect_to(original_uri || { :action => “index” })

right under figure 6.4, where the protected method appears, it is not entirely clear that the method should be contained within class Product.

I’m using rails via cygwin on Windows XP. When testing my database connection I was seeing an error saying “No such file or directory…” with a file name of mysql.sock, as explained in the 6th bullet on page 64. I eventually managed to fix it after looking at the following articles:

http://wiki.rubyonrails.com/rails/pages/database.yml/versions/9
http://blog.labnotes.org/tag/quickhacks

The latter explains, “It turns out that connecting to localhost in the cygwin environment attempts to use Unix sockets. I run MySQL … as a Windows service, using NT pipes for local connections [note: as do I]. Similar concept, but not interoperable. The quick fix is changing localhoust to 127.0.0.1. With an IP address (or any name other than localhost), the client library uses TCP/IP instead.”

I found these articles very useful in helping me to get up and running with rails, so it may be worth including this information in the “list of things to try if rails can’t connect to your database.”

“Developers (or database administrator) make” should read “Developers (or database administrators) make”

Though PostgreSQL is often abbreviated Postgres, it really should be PostgreSQL. (PostgresSQL stems from Postgres95 which stems from Postgres.)

(Dave says: it probably should, but I’ll stick with the common usage. PostgresSQL is unpronouncable… :)

Does Rails always take responsibility for incrementing the value of the primary key? I’m pretty sure that it delegates any such responsibility to the database whenever possible (auto_increment in MySQL, serial sequence in PostgreSQL). In that case, it may be better to write “Rails ensures that the value of this column is incremented”.

Dropping the entire database just isn’t a feasible option in production environments — and it’s not entirely unthinkable that a certain migration did just fine in development trials but failed at the moment supreme. “You’ll have lost nothing” then seems a bit out of place. Simply mentioning that it only goes for development might prevent raising a couple of eyebrows.

“nondatabase” — should that have a hyphen?

Download depot_c/app/view/admin/list.rhtml is missing the ruby code to insert product.image_url into the tag. It’s in the text on the page, but not in the download.

There should be a download link for the first use of depot.css for Iteration A4. I was able to create it via the appendix.

(There’s a download link at the top of the file in the appendix. I really don’t want to include the file inline at this point)

Sample code of depot_c/db/migrate/003_add_test_data.rb - in the description section, there’s a portion of italicized text that probably should not be. It’s in the line:

…
version control: it’s a foundational piece of any project’s
…

The text between the two single quotes is italicized.

(Dave says: yes—our syntax highlighter has a problem with %{…} style strings. We’re all going to have to live with this, I’m afraid)

This page contains a reference to “sec.consolidate”, which doesn’t work. I hope you’re still writing that section and that’s why it doesn’t refer to anything.
I’ve been working with migrations and I must say that I’ve often hated ending up with so so many of them. The automatically generated schema.rb file helps (you should mention that when you introduce the annotate_models plugin), but some automatic way to concatenate the migrations would be wonderful. This link seemed to promise an explanation of hoow to do that…

I read about the Data Migrations with interest, but I’m left wondering a bit when to use fixtures, when to use data migrations, and when to load fixture data in migrations. Maybe you could paint three scenarios that highlight when to use which, or just give some rules of thumb. I would assume one should use fixtures for test data and data migrations for “constant” data. For example, if there is a table that lists the types of credit cards that the system accepts, then that’s really “constant” and needs to be pre-loaded in all environments (dev/test/prod). When adding code to handle a new type of credit card the data migration going with it can add a new row to that table. Cheers & many thanks for a great book!

Please provide description on how to create test suite that links all integration tests together. Also, needed is the description to route the test results to a file so the user may access to this test log file at later time.

(Dave says: just run rake test:integration and it runs them all. You can redirect outputusing your OS facilities for doing that)

Hi Dave:

My understanding is that Rails is moving toward using plug-ins in favor for components. It would be very helpful if your book has a chapter devoting to how to use plug-ins.

(Dave says: that’s a big topic, and the book has already grown far too big… :)

Hi Dave:

In the testing chapter, it would not be complete if it does not include a detailed section in using the Rails debugger.

(Dave says: I cover debugging later—testing is really a different topic)

Add a class method to the Product model to return salable items.

(Dave say: I’m not sure what this report means)

You should consider mentioning which file you expect to insert who_bought.

The title ‘Dynamic Fixtures’ is missing the ‘t’

Hello! I think it would be great if you suggested Markaby as a template-handler (available as a plugin downloadable from http://code.whytheluckystiff.net/svn/markaby/trunk/ and documented at http://markaby.rubyforge.org/) in this second edition of AWDwR. Markaby enables to write views templates in pure-Ruby, thus providing a very clean and smart markup compared to eRuby. It might be quick but useful to explain how to write HTML templates with this wonderful plugin (although Markaby can be used in any Ruby script). Thank you for the brilliant books that you write.

(Dave says: added as an exercise)

“If refactoring leads us to rename variable and columns” -> “If refactoring leads us to rename variables and columns”

depot>curl would look more consistent if there was a space between teh prompt and “curl”

maybe it was mentioned earlier in the book (I don’t think so), but it’s unclear if _field_tag functions will automatically do the necessary html escaping on the parameters they’re passed. ie, are malicious items passed via the query string going to get rendered as is in the elements? It would be nice to see this question addressed here, since it is the first time the_field_tag method is used.

(Dave says: they don’t which why we call h() when displaying values later)

Have some grammar issues:

“We show our customer and are rewarded with a big smile and a request. Could we add a sidebar and put links to the user and product administration stuff in it.” would read a lot better as something like “We show our customer and are rewarded with a big smile and a request: Could we add a sidebar and put links to the user and product administration stuff in it?”

So how do I get to RDoc? (footnote)

(Dave says: changed it to say "Rails api documentation)

The line:
cart = (session[:cart] ||= Cart.new) seems like it's either tricky or buggy. As far as I can tell, it both sets the session[:cart] to a new value if it's empty and assigns the new/existing cart to cart. If this is intended to set session[:cart], it seems like a pretty hacky (idiomatic?) way to do it. It seems like it begs an explanation.

(Dave says: it is indeed idiomatic, and explained earlier in the book)

A footnote indicating why you don’t use the name “hash” for the hashed password field would be nice. I fell into this trap with the first ed. too!

(Dave says: I’m not sure there’s a problem: hashed_password seems like a reasonable name for a hashed password. You could call it ‘hash’, and it would override the built in method with that name)

I’m building a cart and on page 118 it suggests that the ‘product’ method is being called inside the ‘add_to_cart()’ method which is causing a problem.

‘add_to_cart()’ never calls ‘product()’. I think ‘add_to_cart()’ should actually be ‘add_product(product)’ from the Cart class in cart.rb.

Is there meant to be an inconsistency in the way the generator is invoked on these two migration examples? The first example explicitly invokes the ruby interpreter, the second does not. The explicit invocation doesn’t seem to really be needed…

(Dave says: it is for windows)

The sentence “The most common templating scheme, called rhtml embeds snippets…” should have a comma after “called rhtml.”

In the “Installing on Mac OS X”, it should be noted that all problems with ruby were fixed on 10.4.6, and a simply installing Gems an doing a “gem install rails” will do.

(Dave says: despite this, I still install Ruby using DP, simply because it gives me more control)

There is no

I spent plenty of time tracing why ajax-changes didn’t work on my computer. Finally I solved problem by setting IE to check new version of every visit to page. I think you should add this list on that page.

My problem was that development.log wasn’t really helping me.

“Processing StoreController#add_to_cart (for 127.0.0.1 at ActionController::MissingTemplate (Missing template ./script/../config/../app/views/store/add_to_cart.rhtml):”

(Dave says: I believe the new Rails timestamps address this)

The form_tag example on this page isn’t valid Ruby. E.g. type the given code into irb or script/console:

>> form_tag { :action => :save }, { :class => “compact” }
SyntaxError: compile error
(irb):2: parse error, unexpected tASSOC, expecting ‘}’
form_tag { :action => :save }, { :class => “compact” }
^
(irb):2: parse error, unexpected ‘,’, expecting $
form_tag { :action => :save }, { :class => “compact” }
^

A Ruby method call with multiple hashes like the above requires parentheses around the parameters, so the example on this page should be:
<%= form_tag( { :action => :save }, { :class => “compact” } ) %>

Here’s a link that breaks this issue down with a bit more detail:
http://muness.blogspot.com/2006/04/multiple-hashes-as-parameters-in-ruby.html

‘ActionMailer does not deliver e-mail in the test environment. Instead, it appends each e-mail it generates to an array.’

It might be worthy to note that in the development environment, delivery is set to :smtp by default. I find it convenient to use the :test delivery and pop the most recent email off ActionMailer::Base.deliveries in my view.

Missing closing parenthesis from the last sentence on the page: “The fix is either to relax the rule (making the leading zero optional in the requirement with [0-3]?\\d or to make sure you pass in two-digit numbers.”

:datetime cannot use “now()” as a default in MySQL 5.0. A MySQL column of type Timestamp can use CURRENT_TIMESTAMP as the default, but Rails turns :timestamp into a DateTime column.

Is the only / best way around this by manually adding an execute to the migration? This seems like something that will need to be addressed.

Second-to-last line has

I’m getting an error
—————-
ActiveRecord::HasManyThroughAssociationNotFoundError in InfoController#who_bought
Could not find the association :line_items in model Order
—————-
when I try to access the Action who_bought at

myserver.com/info/who_bought/1

(Dave says: do you have ’has_many :line_items" in your order model?)

The iteration Task E: Shipping present in the 1st ed is missing from the 2nd edition. Does this reflect a missing chapter in the process of a re-write? Or did our “client” decide that shipping orders was no longer important? ;)

(Dave says: it’s gone from this edition. It doesn’t teach enough new stuff)

Iteration D4: To know when to reveal the cart with the blind down effect, we’re querying cart.total_items. We need to write this. The implementation you show calculates the total of the items. Instead, return items.length.

(Dave says" No, you need the total number of items, not items.length, because you only reveal when the first product is added to the cart. Add the same product twice, and there’ll still be just one item)

The test on this page (test_unique_title) would also fail because the image URL is invalid. Perhaps it would be clearer (and more appropriate for reader experimentation) if the example contained a valid image URL (i.e. one beginning with http://).

(Dave says: the model only looks at the suffix when validating)

The change to the _cart.rb partial should come with a link to source code just like the others.

(Dave says: this isn’t in the code base. It’s bad code)

Just before the section entitled “Creating the Order Capture Form,” the last sentence should read “relationships starting in Section 17.6….”

First sentence of last paragraph should read, "… but rjs doesn

(And if you

“Sure enough, our user is removed, and an empty list is displayed.”

The login page is displayed instead. After the last user is deleted, session[:user_id] does not match any user in the database, and so the redirect to list_users is not authorized.

(Dave says: good catch)

End of 2nd paragraph: “…useful to diagnosing what went wrong” should be “…useful for diagnosing what went wrong.”

On page 159 of the pdf (or page 171 depending) you show the resulting ‘add user’ form, also on page 158 you make a reference to this form as well. But in both cases the form looks very different because just a stub exists at this point.

It may be helpful to let the user know that this code can be found on page 165, as I was left wondering and had to skip ahead from 11.1 to 11.4 in order to find it.
(Dave says: I’m not sure what’s missing: it comes up looking exactly like this for me)

Figure 9.1 doesn’t match the text at the bottom of the page. The text mentions that upon adding something to your cart you should see a page like in Figure 9.1, but Figure 9.1 shows the index page rather than the add_to_cart page which is then fixed by changing store_controller.rb on the following page. Either don’t mention the figure until the following page or fix the figure.

(Dave says: “if you display the catalog after adding an item to your cart”)

<%= @content_for_layout %> should be <%= yield %> according to http://weblog.rubyonrails.com/2006/04/25/use-params-not-params/

(Dave says: That weblog entry is incorrect)

The footnote on this page states that because it uses an HTML form, button_to() can’t be used in-line the way link_to() can. However, with the proper CSS setup, it IS possible to run this type of element in-line.

Very bottom of PDF-page 99:
[Original]
in the store controller. A simple implementation would be
[Suggested]
in the store controller. An example of a simple implementation would be
[Reason]
While reading the text I suddenly panicked, not knowing with file was being edited. Only after reading further did it occur to me that this was just an example.

Test::Unit does not have ‘assert_not_match’. It is ‘assert_no_match’. I guess this could be called a typo, but it’s a pretty important one, so I classified it as a “Technical Error”. :) It only cost me about 3 minutes, so not too bad.

Last full sentence on page begins, “For this reason, its normally a bad idea…;” “its” should be “it’s.”

Under point #2, you write, “A common example of this kind of problem are database-related performance issues.” “are” is wrong here; use “is” instead, since the verb modifies “example” and not “issues.”

Fourth paragraph after “The Moral of the Tale”…

“include a include Reloadable directive” should be “include an include Reloadable directive”

Under “Testing” you say that some of the testing code from the 1st edition was broken by the release of Rails 1.0. Do you rather mean Rails 1.1? I don’t know if the 1st edition was published before Rails 1.0 or not, but prior to this statement, you were discussing changes in the 2nd edition pertaining to Rails 1.1.

(Dave says: no, it was broken by 1.0. The book went to press before 1.0 came out, and they made the change after)

In this chapter, we

I am writing unit tests for the first time. I went to the back of the book to see the complete Depot app but the tests aren’t there :( Please include them.

When using the example I run into trouble with my images being truncated. It turns out that mysql’s BLOB is limited to 2^16, which wasn’t enough for my images.

I had to use something like:
add_column :products, :image, :binary, :limit => 1.megabyte

in my migration to make rails use MEDIUMBLOB for binary instead of BLOB.

This is not an error in the text, but a simple note might save a lot of time for people like me. It took me a while to figure out where was the mistake.

(Dave says: great tip!)

The logout method is listed as:

def logout
reset_session
flash[:notice] = “Logged out”
redirect_to(:action => “login” )
end

This will correctly log the user out, but the flash is never seen. Resetting the session kills the old session. With no session available, I believe that flash[:notice] will generate a new FlashHash outside of any session. But then a new session is created with the redirect. And the flash is never seen by the browser (same behavior with Firefox and IE6).

Changing this to “session[:user_id] = nil” will log out the user but keep the rest of the session intact.

order = Order.find(:first,
:conditions => “name = ‘Dave Thomas’”,
:order => “id DESC”)

one the last example in the middle section, find(:first) will only ever return one value, so including and :order => “id DESC” won’t have any effect. Right? Consider changing to find(:all) ?

(Dave says: nope—it makes a difference. It determines what the database considers to be the first row)

I think that you should emphasize the “rake db:migrate” command that should be run by formatting it the same as previous commands. It is easy to skip over because the reader is used to seeing all command formatted the same way.

You use this line on this page: find(:all, :order => “title”)
This would be a good time to mention that :order => “title” means order by title. I’m not sure if this has appeared yet in the text, and it may be obvious to many readers, but is still worth mentioning.

just for nit-picking, the error image in figure 8.1 shows different code than what you’ve told the reader to enter (cart_item vs. item). I was getting a different error and I didn’t know why (undefined variable or method because of an “item” in the for statement instead of a cart_item)

In the lib/ section, first sentence: “The lib/ holds application code…” should read “The lib/ directory holds application code….”

Your test_index_without_user includes:

assert_redirected_to :action => “login”

This fails for me, due to a difference in the returned redirect. Earlier on the page you show that the login_controller uses:

redirect_to(:controller => “login” , :action => “login” )

In order to get this test to pass, the asserted redirect had to match:

assert_redirected_to :controller => “login”, :action => :login

(Dave says: it works fine here: looking at the source of assert_redirected to, it seems only to check the controlled if you specify one)

Hi,

you call a method named “self.find_by_name”. Also you call in “controller/application.rb” the method: User.find_by_id

But in the user model there is no method created called find_by_name or find_by_id. Did you miss something? please correkt me if i’m wrong. But the session won’t work without these methods.

(Dave says: it’s a dynamically constructed finder method. I’ve added some text to describe what’s happening)

The string “Download work/demo1” amid references to “demo” without the 1. Apologies in advance if my noob is showing.

(Dave says: the 1, 2, 3 are verions of the code, and demo is the overall name)

those model’s
should be: those models’

“Looking at the .rthml file in Figure 20.2…” should be “Looking at the .rhtml file in Figure 20.2…”

“Next, on line 5, we tell the order object to save itself (and its children, the
line items) to the database.” After reading this, I tried following the code and it is unclear to me where or how the array of line_items is being held for saving. I was expecting that the result of add_line_items_from_cart should be put somewhere like @order.line_items, but it just looks like magic. It would be helpful to clarify how these model child items work.

(Dave says: I’ve added a note to say it’s defined in the next paragraph)

While the response to Joe’s inquiry into credit card processing seems reasonable for a short intro to rails, it sure would be nice to have a brief tutorial in the appendix on how one would go about using either of the two suggested methods. (the developers skimped a little on their readme’s)

(Dave says: to be honest, it’s a book in its own right. I don’t have the space to do this here)

fieldset should be inside the form tags.

(Dave says: I checked with the W3C spec, and I don’t see that for HTML4 strict)

“We moved some application-wide controller helper methods into the
ApplicationController class in the file application.rb in app/controllers.” So far there is only one method in application.rb, but the text hints at plural methods. Also, the method was just created there, not moved.

No matter what I try, I can’t get the user deletion transaction to roll back properly. safe_delete throws an error (I set User.count == 1 to test it, deleted a second user, and received the error message in the view), but the user row gets deleted anyway. Any thoughts?
I am using:
XP Professional (for development)
Ruby 1.8.4
Rails 1.1.4
WEBrick 1.3.1
MySQL 5.0.21

(Dave says: are you using a transactional database engine inside MySQL: MyISAM isn’t)

The download for the add_to_cart.rjs file is from the “depot_r” set. However, previous downloads in this Chapter 10 are from the “depot_p” set, as are downloads in the first part of the next Chapter -> this could be confusing if you’ve downloaded the full source code set (rather than individually).

(Dave says: sorry—that’s just where the source is…)

The p method in the example for in_groups_of is unclear (esp. to those of us just coming to Ruby — I thought it was a variable). It would be good to have a footnote about it or a short explanation, since this is a tricky term to try and google.

You use the idiom ‘self.attribute=’ when setting instance attributes in the password and create_new_salt functions. Setting instance attributes in this way is not, to my knowledge, common among Ruby programmers; it’s certainly not discussed in the Pickaxe book (2d ed.). It’s my understanding that the accepted way to set instance variables is via ‘@attribute=’. If there’s a difference between the two idioms, here would be a good place to explain it.

(Dave says: @fred = assigns to the instance variable. self.fred = invoked the attribute setting method)

In index.rhtml: I found I needed to add a   after the <%= product.price %> code in order to get the same layout as Fig 7.2. I was using the supplied test data, and the generated html looked correct without the  

Without the   my browsers (Firefox 1.0.7 & Konqueror 3.4.3-0.fc3.3) both put the price below the entry’s border-bottom causing the entry below to be offset from the beginning of the line.

CSS/HTML problem or browser rendering bug? I don’t know…

(Dave says: it is formatting fine here. I wonder if you have

tags around your product descriptions (I have in this code))

find_by is not an instance method; it is a class method. You’ll need to change self.find_by_name in the validate_on_create function to User.find_by_name.

Using a model named Order has the potential to interfere with the reserved SQL word ORDER. I didn’t encounter an issue in the context of the code presented in the book, but in some work I’ve done in an app that leverages the depot models and table relationships. I ran into an issue where I started getting SQL errors from MYSQL 4.1.20 while trying to sort a nested Scaffold table my product name (Order table, expand Line Items under order, sort Line Items by product name) using the Product has_many :orders, :through => line_items relationship. I can’t fully explain what exactly triggered it, but once I changed the model Order to StoreOrder, it worked fine.

(Dave says: the word ‘order’ should never leak into the SQL—it will be ‘orders’ by the time it gets there. I suspect you have some other problem)

Because the code for test_unique_title1 wraps onto page 191 and the “Figure 13.1 Standard Active Record validation messages” begins on page 191. It looks like the validation messages are part of the test_unique_title1 code.

It would be great if the seperation were more clear. (maybe move the figure to a different page?)

(Dave says: we do all layout at the end)

In the file show.rhtml (bottom of page), I almost fell off my chair when I saw the argument to ‘src’; there were unquoted double quotes inside the string! There were random thoughts concerning the inclusion of untested code etc. Oh well, I’ll do it right and send an errata later. Really? Yes, really!

However..
Quoting the double quotes did not work, but changing them to single quotes did. This bothered me, so when I had the rest of the upload working (including changing blob to longblob to accomodate modern size images), I tried the example and it worked. I assume this is due to the expansion inside "<% .. %> taking place first.

Suggestion:
Either change to single quotes so parser brains like mine don’t balk or add a footnote explaining why this works. Reading it still bugs me :-)

Add a reference to caching techniques to the last sentence “…by caching the picture action.” I’m reading the PDF version (just love it) and was disappointed to find no reference and no continuation if this on next page.

You make the point a few times that changes to rails or the database show up instantly when we refresh the browser, but this is a common trait of applications written in scripting languages. Is there some sort of caching that would normally take place with rails? The reason for the impressiveness is unapparent.

(Dave says: only true if you’re writing CGI scripts, which reloads on every request. Rails is not a CGI environment)

I had some difficulty understanding the add_user action. Specifically, the line: user = User.new. After some playing around I figured out (I think) that this line clears the form if the user.save succeeds, but it took me a long time of thinking that I was just missing something that I had already been told. A line in the text that explains this would be great!

“Finally … make sure we’ve created a order” - should be AN order

(Ruby Nuby suggestion) Explain the use of “self” in variables. Previously self was used to define class methods. Here it appears that self.salt is an instance variable; is that so? and if so how is it different from @salt?

(Dave says: nope—it’s still a method call)

“After the last migration has been applied, the version in the
schema_info table is updated to its sequence number.” — Really? It doesn’t update the schema_info table after each migration is applied? Your description of backward migrations is incremental, but forward migrations only logged when the entire process is complete?

(Dave says: you’re right—the number is updated for each one)

All this stuff about how to fix the shortcomings of Rails plurals is great, but I had to go to the Rails Wiki to find out where to actually put it to make model generation go smoothly (environment.rb in the config directory). (It seems that Rails thinks the plural of “cactus” is “cactus” rather than either of the two accepted English plurals.)

I had to set my fixture path to:
self.fixture_path = File.join(File.dirname(FILE) + “/../../fixtures/performance”) to get the YAML file to work, and i also had to change,
require File.dirname(FILE) + ‘/../../test_helper’

(Dave says: you have a ‘+’. It should be ‘,’ as in the book)

I bumped into the same problem with the page[:cart].replace_html call not working in rails v 1.1.4. That format yields code like this: ‘$(“cart”).update(etc…)’ whereas the page.replace_html(‘cart’) call returns this js code: ‘Element.update(“cart”,etc…)’ and the changes are displayed correctly in the browser.

I was able to get the highlighting effect to work with this code (rails 1.1.4):

page.visual_effect(:highlight, “current_item”, :startcolor => “\\”#88ff88\\“”, :endcolor => “\\”#114411\\“”)

the hex values have to be quoted or firefox 1.5 will complain about the # sign being an illegal character

(Dave says: this works with the Rails I’m using, so I assume this was a Rails bug)

When working on rhtml files I love to use dreamweaver.

As a suggestion http://www.ridingtheclutch.com/entry.cfm?id=58
is a plugin into dreamweaver to have it support the rhtml commands.

Was very nice during som applications.

The name of the french doctor who you have taken the sentence “Tous les jours, a tout point de vue, je vais de mieux en mieux” is not spelled "Coul

If typing in all the code, watch out for adding an extra ‘,’ after “#114411”. It was very hard to debug this…:)

Page 205 at the section about @assigns. I think this is too vague and it needs some more work. It could benefit from some more examples here. Here’s my suggestion below.
NOTE: that i have not tested the code, but I expect you to use these as examples for improvement.

assigns(key=nil)
Instance variables that were assigned in the last action.
assert_not_nil assigns[“items”]
assert_equal 2, assigns[“items”].size
The assigns hash must be given strings as index references. For example, assigns[:items] will not work because the key is a symbol. To use symbols as keys, use a method call in stead of an index reference.
assert_not_nil assigns(:items)
#=> symbol :items set in controller
assert_not_nil assigns(:products)
#=> symbol @products set in controller

It’s then possible to test further on the contents of your variables in the code:
assert_equal 3, assigns(:products).size
#=> 3 books for sale

Please give examples (note the plural form!) for ALL functional testing helpers, including:

find_all_tag(conditions)
follow_redirect()

Please give examples (note the plural form!) for ALL assertions, including:

assert_dom_equal(expected_html,actual_html,message)
assert_dom_not_equal(expected_html,actual_html,message)

Please give examples (note the plural form!) for ALL HTTP request methods, including:

xhr(method,action,parameters,session,flash)
xml_http_request(method,action,parameters,session,flash)

post()
especially because post is very different from get in this context. show how parameters are posted (and also include an example with nested parameters (we use it all the time in controllers):
post :search, :employee => {:pattern => ‘programming’}

A full table of all test assertions would be much appreciated.
It turns out that it’s hard to find anywhere else. Rails api documents only the added assertions, and the Ruby core api has no documentation of them.

So a collection here would be much appreciated.

(Dave says: I believe I give all the useful ones)

Please give examples (note the plural form!) for ALL assertions in the section “Unit Testing Support”

The link to file depo/app/controllers/storecontroller.rb is one version too refined. In other words, it has the next update to this file’s code already included.

(Dave says: that happens sometimes…)

Line 2: “…let’s load it into a load variable.” — I think you just mean “let’s load it into a variable”.

(Dave says: changed to “local”)

One-to-Many Relationships diagram says:
has_many :invoice
instead of:
has_many :line_items

"It know that it

In the footnote you have:
Unless you use ask your database to enforce foreign key contraints.

Should be:
Unless you ask your database to enforce foreign key contraints.

Sans ‘use’

Currently:
However, we

wrpas should be wraps (2nd line of last para)

Paragraph 2, line 4: “withing th” should be “within the”

Under :from, “let’s” should be “lets”

with_scope.rb:1: Is it deliberate that object dave is created with name Ned?

(Dave says: no… :)

knopwledge

Last word of first para under Counting: “paramters” should be “parameters”

Near bottom of page, “incompatibilities between difference databases” should be “incompatibilities between different databases”.

You are talking about the installation of different database adapters. To install firebird, you say do a “gem install fireruby”, and in the section about installing frontbase you say to do a “gem install fireruby” which is, according to the text, the “ruby-frontbase package”. :)

(Dave says: thanks. It’s gem install ruby-frontbase)

User.login in the sidebar for Friendlier Login System should probably be User.authenticate. user.rb does not have a login method.

Bill

Tried to enter the tutorial code, got the following:
rake —trace —verbose db:migrate
(in /Users/albee/Development/Languages/Ruby/RubyOnRails/trunk/work/depot)

Invoke db:migrate (first_time)

Invoke environment (first_time)

Execute environment

Execute db:migrate
AddPrice: migrating ==
— add_column(:products, :price, :decimal, {:precision=>8, :scale=>2, :default=>0})
rake aborted!
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.[]
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/connection_adapters/abstract/schema_statements.rb:259:in `type_to_sql’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/connection_adapters/abstract/schema_statements.rb:122:in `add_column’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:272:in `method_missing’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:256:in `say_with_time’
/usr/local/lib/ruby/1.8/benchmark.rb:293:in `measure’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:256:in `say_with_time’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:270:in `method_missing’
./db/migrate//002_add_price.rb:3:in `real_up’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:210:in `migrate’
/usr/local/lib/ruby/1.8/benchmark.rb:293:in `measure’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:210:in `migrate’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:332:in `migrate’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:327:in `migrate’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:294:in `up’
/usr/local/lib/ruby/gems/1.8/gems/activerecord-1.14.3/lib/active_record/migration.rb:285:in `migrate’
/usr/local/lib/ruby/gems/1.8/gems/rails-1.1.4/lib/tasks/databases.rake:4
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:387:in `execute’
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:387:in `execute’
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:357:in `invoke’
/usr/local/lib/ruby/1.8/thread.rb:135:in `synchronize’
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:350:in `invoke’
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:1906:in `run’
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/lib/rake.rb:1906:in `run’
/usr/local/lib/ruby/gems/1.8/gems/rake-0.7.1/bin/rake:7
/usr/local/bin/rake:18

Looking at the ActiveRecord::Migration docs, decimal is not an allowed type.
Replacing it with float at least works around the error.

I’m using Rails version 1.1.4, and ActiveRecord version 1.14.3.

(Dave says: you need edge Rails to use the Decimal stuff)

Maybe you can write a sentence on the difference between find_or_create_by and find_or_initialize_by. One of them does save the object back to the database, the other only creates the object, right?

save, save!, create, and create! :
“saves” instead of “saved”

2nd sentence in 1st Para: “Thi” instead of “This”

The last paragraph before 16.7 starts with “All this discussion suggests that migrations are dangerous to us on pro-
duction databases.” Should it be “… dangerous to use”?

In the instructions for setting up sessions in the database, the code from environment.rb changes from

1. Use the database for sessions instead of the file system
2. (create the session table with ‘rake db:sessions:create’)
3. config.action_controller.session_store = :active_record_store

to

1. Use the database for sessions instead of the file system
2. (create the session table with ‘rake create_sessions_table’)
   config.action_controller.session_store=:active_record_store

(note the difference in the second line, the rake command)

I DL’ed 003_add_test_data.rb from your website and tried to use it to load the test data. That version of the script has the prices in decimal dollars instead of cents. It (silently) fails the “numericality_of” validation so no data gets loaded. I changed the prices to integers and everything worked fine.

(Dave says: you need edge Rails to use the Decimal stuff)

In the intstructions to download the images and stylesheets, you say to put them in their “corresponding places” but there’s never been a discussion of the public directory and its subfolders which is where they go.

Paragraph 9.4, the _cart partial listing displays ‘format_price(cart.total_price)’ it should be ‘number_to_currency(cart.total_price)’

“Rails also extends Enumerable with two other methods. index_by( ) It takes a collection and converts in into a hash where the values are the values from the original collection. The key referencing each value is determined
by passing that element to the block.”

The “index_by() It takes…” is a little awkward. And “converts in into” has a typo.

the header “DB2 Adapter” is on the section describing the installation and configuration of the Openbase adapter

In the first sentence of the paragraph beginning “The :precision argument” I think you want “two” instead of “thwo”

orders.length is described twice—before and after orders.empty?

In description of orders.size:
“If you’ve already loaded the association (by access it)”
the “by access it” part doesn’t look right (maybe that’s just my limited knowledge of English).
Should it be “by accesssing it” or “by access to it”?

Second to the last paragraph:
“tells Active Record to go though”

Shoud be:
“tells Active Record to go though”

though -> through

Second sentence in the “Sharing Association Extensions”:
“…in a Ruby module and passing that model…”

model -> module

In the footnote: ontent_for_layout should be content_for_layout

Under Playtime, the second bullet begins “Pass this counter to you template” that should be “your template”

We

The various linkage declarations (belongs_to, has_one, and so one)
*and so one) -> and so on)

bottom of page:
(which is Ruby is treated as false). This let

“… is falling into disue are the basic …” - are should be as

2nd para “… the way THE report errors.” - last the should be they

create method - “… regardless of WHTHER …” - should be whether

In talking about the save!() method, you first say it raises a RecordNotSaved exception when unsuccessful, then later on the page you mention it raises a RecordInvalid exception if the object could not be saved - which one is it, or be more specific about particular exceptions if they’re both possible (I’m not sure for now as a newbie)

(Dave says: different exceptions can be raised depending on circumstances. I’ve had to punt on this)

Last paragraph of the page:

last sentence on the page : “polymorhpic associations” should be “polymorphic associations”

It says “Our performance test runs fine…” but there’s missing how to run the test. It should say (just above “Our performance test…”. (example text feel free to change):

Obviously, the performance test doesn’t run with “rake test”, so we have to start it manually from the command line:
$ ruby ./test/performance/store_controller_test.rb

:password => “optional”, #defaults to ‘root’, should be on the above line :username => “optional”, #defaults to ‘root’

“has_one many more options” needs a HAS between has_one and many

“… letting our caller detmine the …” detmine should be determine

2nd para - “… applies to all the association declations” - last word should be declarations

Ist para for Section 18.4 “Relational database allow us …” should use databases (i.e. plural form) or reword sentence so singular form is OK

Last paragraph suggest making it start with "Now THAT we have …

First sentence has “… how to connection to …” - should be connect

2nd para “(Make sure you database …” should have YOUR not you

Regarding your suggestion of setting up a CI system. This is the first time I’ve ever heard of such a system, and so it would be nice if you could provide a link or two detailing what such a system is (in greater detail) and also maybe a few recommendations. Thx…Bob

(Dave says: I’m nervous about doing this, as the area is changing fairly fast)

in the code block starting with “class AddPrice < ActiveRecord::Migration”, I first typed this in verbatum and got an error when subsequently running rake db:migrate. After checking the Rails API, I believe the third argument to add_column should be :float not “:decimal”. If I use :float, the example works.

(Dave says: you need edge Rails to use the Decimal stuff)

I’m not too sure of the phrase “we can retrieve its order by name”. Technically, it’s retrieved by its ID, not its name, which, btw, is the customer’s name. I think the sentence should drop the “by”, becoming something like “we can retrieve its order’s name” or “…order’s customer name”.

Footnote: “and make the orders column a foreign key” doesn’t make sense. “orders” is a table, not a column anywhere. I guess you meant to say “and add a column with a foreign key… to the orders table.”

(Dave says: it should say "make the payment type column….)

Code blocks lack the line numbering formatting on the left margin; this is not uncommon through the book, and sure is not very useful for small code blocks, but since the text below refers to “line 4” explicitely, it wouldn’t hurt to add the formatting directive for it in the book’s source.

In the paper there is displayed:

redirect_to_index, but it should get a parameter like redirect_to_index(“”)

(Dave says: the parameter is optional)

On page number 103 when talking about the template its written “(We’ll tart it up…)” It seemed like tart should be start.

(Dave says: to “tart something up” is to decorate or make more attractive)

Hi, the DB2 Adapter provided by IBM should be referred to as IBM DB2 in order to distinguish it from the community one. This is reflected in the connection parameters too where it’s ibm_db2 and not db2 only.

You mention compound primary keys are not supported, but there is a module that attempts to do so here: http://rubyforge.org/projects/compositekeys/ - perhaps you could add a footnote linking to that page

2nd line, explcitly should be explicitly

Figure 18.3: Methods Created by Relationship Declarations seems to be out of place. It appears out of context in the middle of an Acts As discussion.

If you think about it, the
original file contains a end-of-line character…

should be

If you think about it, the
original file contains an end-of-line character…

with “an” instead of “a”

“…and thn write the collesponding…”

should be

“…and then write the corresponding…”

First sentence of sec 15.4: “Integers gains…”. Use “Integers gain …” or “Integer gains …”

Many applications impose contraints on what.

Many applications impose CONSTRAINTS on what. (just a missing s)

The with_scope machanism lets

The with_scope MECHANISM lets

When using validation helpers for the User model, isn’t it redundant to include validates_presence_of :password and :password_confirmation when you already have validates_confirmation_of :password? I can’t think of a situation where you would get a different result by leaving out validates_presence_of :password_confirmation, unless it’s simply for a specific error message. DRY?

Include the driver ibm_db2 (the official one produced by IBM).

http://www.alphaworks.ibm.com/tech/db2onrails?open&S_TACT=105AGX01&S_CMP=TILE

I’m working with this, and if you wish can offer some hints and tips for building and using it.

(Dave says: referenced now)

First line of text on page: “The key here are…” should be “The key here is…”.

In description of “:limit => size” I suggest “This basically appends the string (size) to the column type when creating the column.”

I second the suggestion of having Index as a top-level bookmark. In fact, I don’t see it in my bookmarks at all!!

(Dave says: it’s listed under Appendices)

Code listing here has line numbers which makes copy/pasting harder, other code listings (so far) don’t seem to have that

(Dave says: click and link and download the code)

2nd last line of page “…rather that…” should read “rather than”

Section 7.3, 4th paragraph:
"Instead, let

You show how to use the variable @page_title in views, but it’s not showed how it’s set in the controller.

(Dave says: the rest of the book shows that)

Prior to this page you should remark that the CSS and the description markup still need to be “turned on”. I spent some time at this point trying to figure out why my List page did not reflect these items, only to find out on subsequent pages that these items would be enabled later.

In “Playtime”, tracing the flow : “enw fidls” should probably be “new fields”, in “if you change something, for example by adding a enw fidls to a cart item, does that change ripple through the code?”

In “A Friendlier Login System”, login method : I think “redirect_to(session[uri] || { :action => ”index" })" should be : “redirect_to(uri || { :action => ”index" })", since we defined uri = session[:original_uri]. This redirection works for me now.

But if instead of using “wibble” as a bad product id you use “3wibble”, the program acts as if you used "3 as the product id. You might want to catch this error.

(Dave says: true, but not really what I want to illustrate at this point)

Second last paragraph mentions “precision and :scale” parameters, but precision lacks a leading ‘:’

In the “Dave Says” box
Person.find(:first,:condition=>"name=

In the run_with_index example at top of page, you call connection.add_index but the ActiveRecord connection object is missing in the call to remove_index.

Two friends of mine who are learning rails through the beta book have both gotten stuck on the fact that they need edge rails to run the samples. In both cases, they were getting frustrated because the samples didn’t work and they didn’t realize that that needed edge rails (actually neither one knew what edge rails was until I explained it to them). Please make this clearer.

In line 2 it should read “controller”

“Adding a test to the controlled finds this case and
ignores the request.)”

In line 8 it should read “authentication”

“When the next request came in, the authenication failed,…”

“Note that before we compare the price to zero” should really talk about less than one cent. There is another reference to zero at end of the paragraph.

In first paragraph:

…’here we

- If you like the progrmmatic generation of HTML responses…

Two “the”s in this context

Add credit card… content management, and so on the the Depot application.

line 4 - divine the relationships “divine” should be “define”

(Dave says: divine is correct)

particular Subversion revsion number -> revision

Footnote has two typos:
1) “ontent_for_layout" instead of "content_for_layout”.
2) “resonably” instead of “reasonably”

Maybe you can mention the fact that incorrect field names in yaml fixtures will cause the test to fail with an exception (in the current Rails relase 1.1.4).
It cost me nearly two hours to trace the exception back to my fixture file because the debug trace does lead to the core Rails libraries

Page 6 mentions that Rails 1.1 is described.

Question/suggestion: will the CRUD/resource-based features as keynoted during RailsConf and mentioned on http://weblog.rubyonrails.com as to be included in release 1.2 (release “soonish”) also be part of this second edition?

If not, you can already prepare to write a third edition :-) since that 1.2 release will make a significant enhancement to Rails that will be of great interest to a lot of people.

BTW, what a great medium, such a beta book! Great job.

(Dave says: SImplyRestful shoul dmake it into this edition)

There appear to be problems with the default mysql component in Ruby/Rails which result in an error:

Mysql::Error: Lost connection to MySQL server during query: SHOW TABLES

when you do the first “rake db:migrate” command. Judging from web searches, a lot of people have this problem but the book doesn’t currently cover it and the existing tests given in the book don’t help. I suggest adding an extra item referencing:

http://wiki.rubyonrails.com/rails/pages/Mysql+Connection+Problems/

which describes both the problem and how to fix it.

The last sentence at the bottom of the page (the footnote), the sentence reads “We discuss rjs on on page ??”, where “on” has been doubled.

The first paragraph reiterates the problem with GET versus POST in a way that looks like the writer doesn’t remember what they wrote about :post => true one page 85 (PDF). This is quite annoying. Consider rewriting the paragraph on page 95, and/or mention button_to in the sidebar on 85.

It would be nice to point out potential race conditions when using “validates_uniqueness_of”

See http://www.ruby-forum.com/topic/65257

(Dave says: good point)

In the Playtime section, one of the items is to add a helper method for currency conversion, however we aren’t told how to add helper methods until page 133. At the very least, a reference to the helper method section might be worthwhile.

The first sentence reads “… use respond_to() to vector to the correct …” which cannot be right.

(Dave says: seems OK to me)

“quatity” instead of “quantity”

Last sentence in last paragraph begins: “You cn use it…”, it should probably read “You can use it…”

Re: "If we were rolling the Depot application into production, we

last sentence “with deployment your application” - make deployment => deploying

Last sentence before section 15.4 begins “In a Raails application”, which should probably be “In a Rails application”

On this (and the next page) you write that when we use display:none for the cart, we don’t get any flicker in the browser when the cart blinds down. this works in firefox but in the ie6 I still get a flicker.
knoebi [] gmail.com

(Dave says: I added a note to this effect)

First word (continued from previous page) - “establish these con-traits - constraints is missing the ”s".

Cheers.

The oracle adapter is listed as taking a
:database => “railsdb”
parameter. This should instead be:
:host => “db.company.com/railsdb”

(Dave says: according to the source, :host is obsolete)

"… It installs as a plugin, so you

In the next second last paragraph, the second sentence contains the word polymorhpic which should surely be polymorphic.