Last night Google announced Java support on Google App Engine.

After a bit of toying around, here are my findings.

The Eclipse plugin is pretty slick. Deploying and build is simple.

The dev server that you spin up locally looks to be jetty under the hood.

Objects intended for storage are JDO annotated and after compiling, you run the .class files through the DataNucleus Enhancer which adds additional metadata so Google can map it to BigTable. The Eclipse plugin automatically performs this step for you after compiling. The examples provide a bunch of ant macros to help facilitate building/deploying.

One issue that I had was that the project was building with Java 1.6 and I would get an error after compiling:

Caused by: java.lang.UnsupportedClassVersionError: Bad version number in .class file

Even though they say they support Java 1.5 and 1.6, I guess this doesn’t work on the Java 1.6 for the Mac. Switching the build to 1.5 allows the DataNucleus Enhancer to run successfully.

Even though they are using JPA, some features have not yet been implemented or supported ( see http://code.google.com/appengine/docs/java/datastore/usingjpa.html#Unsupported_Features_of_JPA)

Overall, I like what I see so far and think this would be great for quick prototypes of web apps/services.

Going through the tutorial, my awesome Guestbook application has been created and deployed.

Posted in geekery, java at April 8th, 2009. View Comments.

The defaults in Rails with ActiveRecord is beautiful when you are just getting started and are created everything for the first time. But once you get into it and your database schema becomes a little more solidified, the things that would have been easy to do by relying on the conventions of Rails require a little bit more work.

In my case, I had a form where there was a database column named “num_guests”, representing the number of guests. When the field fails to pass validation, the error messages is something like

Num guests is not a number

Not quite the text that we want. It would be better if it said

Number of guests is not a number

After doing a little bit of digging, I found the human_attribute_name method. You can override this method in your model class to provide alternative names for fields. To change our error message, I did the following

class Reservation < ActiveRecord::Base
  ...
  validates_presence_of :num_guests
  ...
  HUMAN_ATTRIBUTES = {
      :num_guests    => "Number of guests"
  }

  def self.human_attribute_name(attr)
      HUMAN_ATTRIBUTES[attr.to_sym] || super
  end
end

Since Rails 2.2, this method is used to support internationalization (i18n). Looking at it, it reminds me of Java’s Resource Bundles and Spring MVC’s error messages. Messages are defined based off a key and there’s a chain of look ups that get applied to resolve an error’s message.

Although, I don’t see myself doing any i18n work in the near-term, it is cool that we have that option now in Rails.

Posted in geekery at April 6th, 2009. View Comments.

When building dynamic web sites with lots of javascript UI components being created on the client, understanding how the web framework you’re using will process the request and what must be done to update fields accordingly is even more important.

Specifically, checkboxes have always been a pain to deal with. The gotcha with checkboxes are if a checkbox isn’t checked, the request doesn’t send the parameter so it requires some additional checks to detect that the user deselected something that was there to update the field accordingly. I’ve been playing around with the Stripes framework and ran into this issue.

With Stripes, you can render a checkbox using their JSP tag:

<stripes:checkbox checked="true" name="property1" value="yes"/>
<stripes:checkbox checked="true" name="property2" value="no"/>

When the “checked” value is equal to “value” value, Stripes will render the checkbox as checked. So with the code shown, two checkboxes will be shown with the first checked and the second unchecked.

If a user reverses this by unchecking the first, checking the second, and submit the form, the HTTP request will only see that property2=no. Before the form was submitted, “property1″ had a value of “yes”. Now, “property1″ won’t even appear in the request parameters, so we have to do special handling to check for the absent of the parameter to update “property1″ to whatever value it should be when it is not checked.

In Spring MVC with form binding, checkboxes are dealt with a little differently. Using Spring MVC’s form JSP tag, you can do:

  <form:checkbox path="property1" value="yes"/>
  <form:checkbox path="property2" value="no"/>

Assuming your command bean is named “person”, this will generate the following HTML:

        <input name="person.property1" type="checkbox" value="yes"/>
        <input type="hidden" value="1" name="_person.property1"/>
        <input name="person.property2" type="checkbox" value="no"/>
        <input type="hidden" value="1" name="_person.property2"/>

As noted by the docs,

What you might not expect to see is the additional hidden field after each checkbox. When a checkbox in an HTML page is not checked, its value will not be sent to the server as part of the HTTP request parameters once the form is submitted, so we need a workaround for this quirk in HTML in order for Spring form data binding to work. The checkbox tag follows the existing Spring convention of including a hidden parameter prefixed by an underscore (“_”) for each checkbox. By doing this, you are effectively telling Spring that “the checkbox was visible in the form and I want my object to which the form data will be bound to reflect the state of the checkbox no matter what”.

Spring MVC also provides a “checkboxes” tag which allows you to render a list of checkbox boxes without having to wrap the “checkbox” tag around a JSTL forEach.

Hopefully, that gives you some insight into how to work with checkboxes in Stripes and Spring MVC.

Posted in geekery, java, tutorials, web at March 29th, 2009. View Comments.

If you’ve been following my blog for awhile, you might have noticed I moved this blog from blog.notedpath.com to theodorenguyen-cao.com as it was more fitting domain. I originally just registered the domain, added the DNS record, and updated my apache config to have theodorenguyen-cao.com to be an server alias to blog.notedpath.com.

This allowed requests to blog.notedpath.com/* and theodorenguyen-cao/* respond with the same content. I thought I was done. I discovered this wasn’t the case when I saw blog.notedpath.com as a direct traffic source in my google analytics for theodorenguyen-cao.com. To fix the screwed up analytics, I needed to make it so that all requests that go to blog.notedpath.com are permanently redirected (301) to theodorenguyen-cao.com.

To do this I had to apply an Apache mod_alias redirect directive as such:

<VirtualHost *:80>
        VirtualDocumentRoot /var/www/blog
        ServerName blog.notedpath.com
        Redirect permanent / http://www.theodorenguyen-cao.com/
        ErrorLog /var/log/apache2/wp-error.log
        TransferLog /var/log/apache2/wp-access.log
</VirtualHost>

The virtual host for theodorenguyen-cao.com looks like:

<VirtualHost *:80>
    VirtualDocumentRoot /var/www/blog
    ServerName www.theodorenguyen-cao.com
    ServerAlias theodorenguyen-cao.com
    CustomLog /var/log/apache2/theodorenguyen-cao.com_access.log Combined
    ErrorLog /var/log/apache2/theodorenguyen-cao_error.log
</VirtualHost>

At first I thought this would only fix the simple case of blog.notedpath.com redirecting to theodorenguyen-cao.com, but blog.notepath.com/foobar not being translated to theodorenguyen-cao.com/foobar. However, this does exactly what I want. All blog.notedpath.com URLs will be replaced with theodorenguyen-cao.com URLs. Old bookmarks will simply redirect to a theodorenguyen-cao.com URL and not 404.

Success!

I’m still waiting to see if Google will update the search result links that point to blog.notedpath.com to be theodorenguyen-cao.com URLs.

Posted in geekery at January 21st, 2009. View Comments.

rsvp.theoandpat.com had boolean flag to marked whether or not a visitor was going to be able to make it to our wedding. Unfortunately, if you selected you were not able to make it and submit the form, the application would return saying it could not process your submission because you have to say that you are going to make it. I argued, this is an RSVP form so you have to accept if you are RSVPing. That’s the point of the RSVP! Only people RSVP would bother submitting the form!  Pat wasn’t too happy about that and ask/told me to fix it. 

Digging into it, it turns out the way  for validates_presence_of relies on Object#blank which of course when sent

false.blank? # returns true

Reading up on the documentation, it is suggested to use validates_inclusion_of when dealing with booleans.

The one line change solved the problem:

validates_inclusion_of :accepted, :in => [true, false]
Posted in geekery, ruby, tutorials at January 12th, 2009. View Comments.