Things happened in my life

During the past few months, I have chances to work in projects which have the requirement to monitor a specific folder on the filesystem for changes and act accordingly (e.g import structured file from a FTP upload folder). The following solutions I found useful to fulfill the above requirements:

jNotify

Pros: easy to use, cross-platform (kind of, because jNotify author has developed native code implementation for 3 platforms: Windows, Linux, and MacOS X), can be used with Java 1.4+

Cons: there are a few limitations on each platform as noted on its website.

I've got a problem when I tried to run it on a Ubuntu 64bit machine, but after searching on their forum on sourceforge, I've found my solutions (by recompiling the native code). I've used it in production since then and it never failed me. But there is a catch: when you move files/folders into your monitored folders, it won't detect the event as new files/folders created. I think it's because Linux only updates (and not creates) the inodes of those files/folders on the ext4 filesystem. In order for your Java application to pickup the native code, you have to put the native library (.so for Linux, .dll for Windows) in the folder specified by the system property java.library.path

JPathWatch

Pros: cross-platform (including Windows, Linux, MacOS X and FreeBSD), the native library is bundled inside the distribution jar so all you have to do is including its jar in the classpath.

Cons: the API is a little hard to use (the author do not use the observer pattern due to the scalability), requires Java 5+

WatchService in JDK 7

Pros: built-in in standard distribution of Java 7, no other external libraries or native code is required.

Cons: you have to use Java 7, of course. This sometimes is a showstopper in various projects, although Java 6 is about to reach EOL (in November this year)

You can download all the sample code at https://bitbucket.org/tinhtruong/blog-samples/src/6d3ce2ac58a8/filesystem-monitoring

I just released the first public version of Spring Security Acl with Mongodb as the data store at https://bitbucket.org/tinhtruong/spring-security-acl-mongodb. Inspired from Spring implementation based on JDBC, I leverage the power of the Spring Data Mongodb to implement the retrieve strategy for ACL stuff. The implementation is based on Spring Security 3.1.0.RELEASE (the latest version at this time).

It is released under Apache 2.0 license, so you can use it in your commercial projects.

I use Gradle as the build system. So if you want to build from the source, make sure you have an up-and-running Gradle 1.0. To create a jar:

gradle jar

Feedback are welcome!

In integration tests, I often need to setup the database to a determined state before every test case. In order to do that, we need to create database schema which reflect your current models and then populate with proper data. In today post, I will show you how to generate the database schema from your Hibernate models by using the Hibernate Maven Plugin (3.0). In my shop, we are still using the "old-school" Hibernate mapping: XML!. The reason is we want to keep our models pristine (no annotations or whatever). We use Maven as our build tool, so here is the plugin configuration that works:

    org.codehaus.mojo
    hibernate3-maven-plugin
    3.0
    
        
            create-schema
            process-test-resources
            
                run
            
            
                
                    
                    
                
            
        
    

We use Spring to configure Hibernate directly, the hibernate.cfg.xml used here is just for this purpose only (ie. generating the database schema). Here is it just for the sake of completeness:

    
        org.hibernate.dialect.HSQLDialect
        org.hsqldb.jdbcDriver
        jdbc:hsqldb:mem:sample_db
        sa
        
    

There are other operations that you can utilize with Hibernate Tools. For more information, please refer to Hibernate Tools documentation and the hibernate3-maven-plugin site

.

I have a chance to use Lucene (3.6.0) to implement a full-text search in one of my recent projects. One of the requirements is to highlight the matched text in the result. The highlighted text should be displayed in the whole paragraph (not just a small text fragment). Here is my snippet to achieve this:

private String getHighlightedField(Query query, Analyzer analyzer, String fieldName, String fieldValue) throws IOException, InvalidTokenOffsetsException {
    Formatter formatter = new SimpleHTMLFormatter("", "");
    QueryScorer queryScorer = new QueryScorer(query);
    Highlighter highlighter = new Highlighter(formatter, queryScorer);
    highlighter.setTextFragmenter(new SimpleSpanFragmenter(queryScorer, Integer.MAX_VALUE));
    highlighter.setMaxDocCharsToAnalyze(Integer.MAX_VALUE);
    return highlighter.getBestFragment(this.analyzer, fieldName, fieldValue);
}

• By creating a SimpleSpanFragmenter with a very big fragment size, we can display the highlighted text in the whole paragraph or document. Lucene also does a nice thing here for free, by merging all the highlighted text fragments into one big chunk (or our original paragraph/document)
• query: is the Lucene query you constructed to do the search
• analyzer: is the Lucene analyzer are used to analyzed the field when you create the index for that field

Happy highlighting!

In my previous post, I successfully managed to put all the output of Maven (and Eclipse) into RAM to reduce the disk write. In today post, I will do the same thing but with Gradle. Here is my build.gradle

apply plugin: "java"
apply plugin: "war"
apply plugin: "eclipse"

sourceCompatibility = 1.6
targetCompatibility = 1.6

group = "me.tinhtruong"
version = "1.0"

repositories {
    mavenCentral()
}

dependencies {
    compile "com.icegreen:greenmail:1.3.1b"
    compile "jstl:jstl:1.2"
    providedCompile "javax.servlet:servlet-api:2.5"
}
// Change the buildDir of the current project to /tmp/ (which is mounted on RAM using tmpfs)
buildDir = new File("/tmp/gradle/" + project.name);

// First, create a linked resource pointing to our Gradle buildDir, then hook into the .classpath file generation of Gradle to change the output of Eclipse to that linked resource.
eclipse {
 project {
  linkedResource name: 'build', type:'2', location: buildDir.absolutePath
 }
 classpath {
  file {
   whenMerged { classpath ->
    classpath.entries.each { entry ->
     if (entry.kind == 'output') {
      entry.path = 'build/eclipse-output';
     }
    }
   }
  }
 }
}

It has been a while since I finished porting the Monokai color theme from TextMate to Eclipse. Thanks to the original author of Monokai (I believe this is the guy) and the authors of the awesome Eclipse color theme plugin which enable everyone to create an Eclipse color theme easily. Here is the Monokai theme in action (click on the screenshot for larger preview)

I've recently bought an Intel 520 SSD, so my top priority for my laptop is to reduce the disk write during daily use. In the Java world, Maven is one of the most common build tool out there. My most used Maven command is:

mvn clean package

This command will do a fresh build of the project, then copy all the compiled classes/resources to the build directory of Maven (which is the folder 'target', next to your source directory). If you have a large Maven project at hand, this command will perform thousands disk write (clean compiled classes and resources, then compile Java classes and assemble the resources into WAR/JAR file). By changing the build directory of Maven to /tmp (which is a tmpfs mountpoint (residing on RAM on my ArchLinux)), I can now rebuild my project as many times as I like without the write amplication concern. Here is now I do it in Maven:

• Define a property name target.directory with the default value is target
• Under the build section in your POM file, add this tag <directory>${target.directory}</directory>
• Modify your settings.xml file to include this snippet in your active profile:

  
      /tmp/maven/${project.groupId}-${project.artifactId}/target
  
  

With the above configuration, maven will do all the build in /tmp, my SSD will thanks me for this :), but we have a small problem! Eclipse does not allow the build directory to be located outside the project directory, but I have a workaround for that. By defining a linked resource, Eclipse will happily with our setup! Here is the complete sample pom.xml file:

 4.0.0
 com.abc
 abc
 war
 1.0
 Sampe Webapp
 http://www.abc.com
 
  target
 
   
  
   junit
   junit
   3.8.1
   test
  
 
 
  abc
  ${target.directory}  
 
 
   
     maven-eclipse-plugin
       2.8
       
         true
         true
         target
         
           
             target
             2
             ${project.build.outputDirectory}
           
         
       
   
 

And the sample settings.xml file:

  
  
    
      development
      
        /tmp/maven/${project.groupId}-${project.artifactId}/target
       
    
    
      public-snapshots      
  
  
    development
  

Happy coding!

This how-to is served as my memo to import a Bitbucket project into Eclipse. First things first, make sure you have an up-and-running Eclipse with EGit (in this how-to I am using Eclipse Indigo 3.7.2 with EGit 1.3.0)

• Create a folder to hold the workspace for your new project (You can skip this step if you want to use an existing workspace), I will use one of my project on Bitbucket in this how-to. It's GreenMail web application. So I create a folder name greenmail-webapp
• Open the Git Repositories view (it's under Windows -> Show View -> Other..., then under the Git category).
• Click 'Clone a Git repository' and paste in the whole Git repo URL to the URI textbox, EGit will automatically fill in other textbox with the right value (so cool!), all you need is to enter the correct password. Then click 'Next'.
• At the Branch Selection screen you are able to choose a branch to clone, my project has only one branch (master). If you clone an empty project from Bitbucket, just click next.
• At the Local Destination, browse to your workspace directory for this project. I usually don't check-in the .project and .classpath files into the VCS, I just generate them locally right after I checkout/clone instead. If your build tool does not generate those files for you or you are the only one in this project, it'd better to check-in those files.
• Now, let's generate the Eclipse project files. I use Gradle as my build system, so it's just a command away.

  gradle eclipse
  

  Before running this, make sure you've already applied the 'eclipse' plugin in your build.gradle. If you use Maven, please include the maven-eclipse-plugin into your pom.xml and run the command

  mvn eclipse:eclipse
  

  The above commands will generate/overwrite the .project and .classpath for you.

• Returning to your Eclipse, you can now Import your new cloned project into your workspace by right clicking on your in the Git Repository view, select Import Projects...

  
• Select Import existing projects
• Select the project you want to import.

Happy coding!

In my recent project, I have a need to monitor a directory for changes such as a new file created, modified or deleted. Surprisingly, JDK (up to JDK 6) does not have APIs to do that! JDK 7 does support it out of the box but I am stuck with the version 6

After looking around for an existing solution, I finally found that jNotify seems to fit the need. It supports all the three major platforms (Windows, Linux, and MacOS).The project has not released any update for nearly two years, so it's a little bit tricky to compile the native code. In this post, I will show you how! (at least on a Ubuntu Server 64bit)

In order to build anything serious on Ubuntu, you should install the package build-essentials

sudo apt-get install build-essential

Then download the source code of jNotify from its homepage. Extract it, open a terminal and issue the command: (assuming that you are at the directory jnotify-native-linux-0.93-src)

cd Release
make

On ArchLinux 64bit, the compilation will succeed and you will get a libjnotify.so in the same folder. But on my Ubuntu Server 10.04 LTS, I've got this:

/usr/include/asm-generic/fcntl.h:96: error: expected specifier-qualifier-list before ‘pid_t’

After Googling around I found this thread. Basically, you have to change the file net_contentobjects_jnotify_linux_JNotify_linux.c by moving up the "unistd.h" up above the "sys/time.h". Now issue the command "make", the compilation will work as expected. Happy coding!

I've been exposing to Gradle for a couple of weeks and really like it. As a Maven user, I found that Gradle is a refreshing methodology on how a build tool should be. But one thing I miss from the Maven land: the ability to generate the initial structure for a project (I mean the 'archetype' plugin of Maven). Luckily, due to the plugin architecture of Gradle, a Gradle user has developed a plugin just for that purpose. It's called the 'tempaltes' plugin. The wiki page of the plugin gives you all the details you need to install, but the global installation section is confusing and did not work. Here is how I did to make it work:

• Create a file called templates.gradle in the ~/.gradle/init.d/ folder (actually you can name it whatever you want as long as it has the extension .gradle)
• Edit that file and paste this little code snippet:

gradle.beforeProject { prj ->
   prj.apply from: 'http://launchpad.net/gradle-templates/trunk/latest/+download/apply.groovy'
}

Now, you can create a project structure event without a build.gradle with the command:

gradle createJavaProject

Happy coding!

The title says it all. This post will show you how to configure Spring Security with Digest authentication and encoded password. But first, here is the Spring Security documentation about digest authentication and encoded password:

The configured UserDetailsService is needed because DigestAuthenticationFilter must have direct access to the clear text password of a user. Digest Authentication will NOT work if you are using encoded passwords in your DAO. The DAO collaborator, along with the UserCache, are typically shared directly with a DaoAuthenticationProvider. The authenticationEntryPoint property must be DigestAuthenticationEntryPoint, so that DigestAuthenticationFilter can obtain the correct realmName and key for digest calculations.

It's not all true. Let's take a look at the source code of DigestAuthenticationFilter

public void doFilter(ServletRequest req, ServletResponse res, FilterChain chain) throws IOException, ServletException {
    // omit code for simplicity
    
    // Compute the expected response-digest (will be in hex form)
    String serverDigestMd5;

    // Don't catch IllegalArgumentException (already checked validity)
    serverDigestMd5 = DigestAuthUtils.generateDigest(passwordAlreadyEncoded, username, realm, user.getPassword(),
                    request.getMethod(), uri, qop, nonce, nc, cnonce);
    // omit code for simplicty
}

Digging a little more into the method generateDigest of the class DigestAuthUtils, we have:

static String generateDigest(boolean passwordAlreadyEncoded, String username, String realm, String password,
                                        String httpMethod, String uri, String qop, String nonce, String nc, String cnonce)
            throws IllegalArgumentException {
    String a1Md5 = null;
    String a2 = httpMethod + ":" + uri;
    String a2Md5 = md5Hex(a2);

    if (passwordAlreadyEncoded) {
        a1Md5 = password;
    } else {
        a1Md5 = DigestAuthUtils.encodePasswordInA1Format(username, realm, password);
    }
   
   // The rest of the code

A little more into encodePasswordInA1Format of DigestAuthUtils:

static String encodePasswordInA1Format(String username, String realm, String password) {
    String a1 = username + ":" + realm + ":" + password;
    String a1Md5 = md5Hex(a1);

    return a1Md5;
}

From here, if we set the property passwordAlreadyEncoded of DigestAuthenticationFilter to true and create a suitable password encoder and salt source, we can make Spring Security digest authentication work with our tailored password encoder, salt source. Let's create them:

package com.mycompany.myproject;

import org.springframework.beans.factory.InitializingBean;
import org.springframework.security.authentication.dao.SaltSource;
import org.springframework.security.core.userdetails.UserDetails;
import org.springframework.util.Assert;

public class DigestAuthenticationAwareSaltSource implements SaltSource, InitializingBean {
 
    private String digestRealm;
 
    @Override
    public void afterPropertiesSet() throws Exception {
        Assert.hasText(digestRealm, "A Digest Realm must be set");  
    }

    @Override
    public Object getSalt(UserDetails user) {
        return String.format("%s:%s:", user.getUsername(), digestRealm);
    }

    public void setDigestRealm(String digestRealm) {
        this.digestRealm = digestRealm;
    }
}

package com.mycompany.myproject;

import java.security.MessageDigest;
import java.security.NoSuchAlgorithmException;

import org.springframework.dao.DataAccessException;
import org.springframework.security.authentication.encoding.PasswordEncoder;
import org.springframework.security.core.codec.Hex;

public class DigestAuthenticationAwarePasswordEncoder implements PasswordEncoder {

    @Override
    public String encodePassword(String rawPass, Object salt) throws DataAccessException {
        String a1 = salt + rawPass;
        return md5Hex(a1);
    }

    @Override
    public boolean isPasswordValid(String encPass, String rawPass, Object salt) throws DataAccessException {
        String calculatedPass = md5Hex(salt + rawPass);
        return calculatedPass.equals(encPass);
    }
 
    private String md5Hex(String data) {
        MessageDigest digest;
        try {
            digest = MessageDigest.getInstance("MD5");
        } catch (NoSuchAlgorithmException e) {
            throw new IllegalStateException("No MD5 algorithm available!");
        }

        return new String(Hex.encode(digest.digest(data.getBytes())));
    }
}

And now you can do Digest Authentication with encoded password in Spring Security!

First, please make sure you've already installed JDK on your Centos box. There are tons of tutorials on how to do that :). Now we will download and install Jetty, at the time of this writing, the latest version of Jetty is 7.4.5.v20110725. Every command listing below must be run as user root.

cd /opt
wget http://download.eclipse.org/jetty/stable-7/dist/jetty-distribution-7.4.5.v20110725.tar.gz
tar xzvf jetty-distribution-7.4.5.v20110725.tar.gz
ln -s jetty-distribution-7.4.5.v20110725 jetty

OK, we've just installed Jetty to /opt/jetty. Now we will setup Jetty to run on startup

cd /etc/init.d
ln -s /opt/jetty/bin/jetty.sh jetty
chkconfig --add jetty
chkconfig --level 345 jetty on

It's a good practice to run Jetty as a non-root user, let's create a normal user called jetty

useradd -m jetty
chown -R jetty:jetty /opt/jetty

Now edit the startup script of Jetty located at /opt/jetty/bin/jetty.sh to include the following lines at the beginning (right after a bulk of commented lines describing various variables for Jetty)

JAVA_HOME=/opt/java
JAVA=$JAVA_HOME/bin/java
JAVA_OPTIONS=" -server -Xms256m -Xmx1024m -XX:+DisableExplicitGC "
JETTY_HOME=/opt/jetty
JETTY_USER=jetty
JETTY_PORT=7070
JETTY_HOST=0.0.0.0 # If you don't set this to 0.0.0.0, jetty only listen on localhost
JETTY_LOGS=/opt/jetty/logs/

Now start up it:

service jetty start

Open your browser and navigate to http://:7070/ Hope this help!

I've maintained the ibus-unikey package in Arch User Repository (AUR) for a while and suddenly my package disappeared. While I was wondering what happened to my package, I received an email from a Trusted User (TU) of Arch saying that my package had been moved to the community repo of Arch. Kudo to the ibus-unikey team for such a great piece of software which are so useful for Vietnamese Arch Users!

A quick and dirty way to do this is using a built-in authentication mechanism of Tomcat (I'm using Tomcat 6.x).

 
  EntireApp
  /*
  GET
  POST
 
 
  client
 


  client



 BASIC

If you use Spring Security, remember to remove Spring Security filter in the web.xml file. Now in your context file (the context.xml file in your META-INF of your web archive or in ther conf/Catalina/localhost/ context file), adding the following:

Finally, edit the tomcat-users.xml (usually located at /conf/tomcat-users.xml) to include all the roles you've specified in the security constraint of you web application, my tomcat-users.xml looks like this (yours may be different):

Now, every time the clients access the demo site, the browser will pop up a dialog asking for username and password. After successfully authenticated, the clients will use the demo site as normal.
This authentication mechanism is the Basic access authentication. As the title said, this is a quick and dirty way to secure a web application on Tomcat 6.
Hope this saves you some time.

I've just come across a Grails plugin called PrettyTime and I find it's really interesting library. Basically, it's utility library to format Date/Time value in a human-friendly form (eg: "right now", "two days ago", "3 months from now"...), it also supports localization for many languages but Vietnamese. So I decide to write a small resource bundle class for it.

package com.ocpsoft.pretty.time.i18n;

import java.util.ListResourceBundle;

public class Resources_vi extends ListResourceBundle
{
   private static final Object[][] OBJECTS = new Object[][] {
   { "CenturyPattern", "%n %u" },
   { "CenturyFuturePrefix", "" },
   { "CenturyFutureSuffix", " sau" },
   { "CenturyPastPrefix", "cách đây " },
   { "CenturyPastSuffix", "" },
   { "CenturyName", "thế kỷ" },
   { "CenturyPluralName", "thế kỷ" },
   { "DayPattern", "%n %u" },
   { "DayFuturePrefix", "" },
   { "DayFutureSuffix", " sau" },
   { "DayPastPrefix", "cách đây " },
   { "DayPastSuffix", "" },
   { "DayName", "ngày" },
   { "DayPluralName", "ngày" },
   { "DecadePattern", "%n %u" },
   { "DecadeFuturePrefix", "" },
   { "DecadeFutureSuffix", " sau" },
   { "DecadePastPrefix", "cách đây " },
   { "DecadePastSuffix", "" },
   { "DecadeName", "thập kỷ" },
   { "DecadePluralName", "thập kỷ" },
   { "HourPattern", "%n %u" },
   { "HourFuturePrefix", "" },
   { "HourFutureSuffix", " sau" },
   { "HourPastPrefix", "cách đây " },
   { "HourPastSuffix", "" },
   { "HourName", "giờ" },
   { "HourPluralName", "giờ" },
   { "JustNowPattern", "%u" },
   { "JustNowFuturePrefix", "" },
   { "JustNowFutureSuffix", " khắc sau" },
   { "JustNowPastPrefix", "cách đây " },
   { "JustNowPastSuffix", " khắc" },
   { "JustNowName", "" },
   { "JustNowPluralName", "" },
   { "MillenniumPattern", "%n %u" },
   { "MillenniumFuturePrefix", "" },
   { "MillenniumFutureSuffix", " sau" },
   { "MillenniumPastPrefix", "cách đây " },
   { "MillenniumPastSuffix", "" },
   { "MillenniumName", "thiên niên kỷ" },
   { "MillenniumPluralName", "thiên niên kỷ" },
   { "MillisecondPattern", "%n %u" },
   { "MillisecondFuturePrefix", "" },
   { "MillisecondFutureSuffix", " sau" },
   { "MillisecondPastPrefix", "cách đây " },
   { "MillisecondPastSuffix", "" },
   { "MillisecondName", "mili giây" },
   { "MillisecondPluralName", "mili giây" },
   { "MinutePattern", "%n %u" },
   { "MinuteFuturePrefix", "" },
   { "MinuteFutureSuffix", " sau" },
   { "MinutePastPrefix", "cách đây " },
   { "MinutePastSuffix", "" },
   { "MinuteName", "phút" },
   { "MinutePluralName", "phút" },
   { "MonthPattern", "%n %u" },
   { "MonthFuturePrefix", "" },
   { "MonthFutureSuffix", " sau" },
   { "MonthPastPrefix", "cách đây " },
   { "MonthPastSuffix", "" },
   { "MonthName", "tháng" },
   { "MonthPluralName", "tháng" },
   { "SecondPattern", "%n %u" },
   { "SecondFuturePrefix", "" },
   { "SecondFutureSuffix", " sau" },
   { "SecondPastPrefix", "cách đây " },
   { "SecondPastSuffix", "" },
   { "SecondName", "giây" },
   { "SecondPluralName", "giây" },
   { "WeekPattern", "%n %u" },
   { "WeekFuturePrefix", "" },
   { "WeekFutureSuffix", " sau" },
   { "WeekPastPrefix", "cách đây " },
   { "WeekPastSuffix", "" },
   { "WeekName", "tuần" },
   { "WeekPluralName", "tuần" },
   { "YearPattern", "%n %u" },
   { "YearFuturePrefix", "" },
   { "YearFutureSuffix", " sau" },
   { "YearPastPrefix", "cách đay " },
   { "YearPastSuffix", "" },
   { "YearName", "năm" },
   { "YearPluralName", "năm" },
   { "AbstractTimeUnitPattern", "" },
   { "AbstractTimeUnitFuturePrefix", "" },
   { "AbstractTimeUnitFutureSuffix", "" },
   { "AbstractTimeUnitPastPrefix", "" },
   { "AbstractTimeUnitPastSuffix", "" },
   { "AbstractTimeUnitName", "" },
   { "AbstractTimeUnitPluralName", "" } };

   @Override
   public Object[][] getContents()
   {
       return OBJECTS;
   }
}

Hopefully they will integrate this snippet into the next release of PrettyTime.

A couple months ago, I moved my blog from Blogger to Wordpress, that's because Wordpress offers some nice themes. Recently, I've bought a new domain (my first domain actually) and I want to use that domain for my blog but it costs me 10$/year to use custom domain in Wordpress! Searching around and I found that Blogger let you use your custom domain for free with hassle-free setup. Horay!
Then I moved back to Blogger, but the problem arises, the ugly theme! I think it's the right time for me to learn CSS and start hacking Blogger API. And after 2 weeks of hard working, I've finished my custom theme for Blogger based on Natural Power theme for Wordpress.
I'm quite happy with the result I've got so far. Not too bad for a CSS newbie.
From the normal user perspective, Wordpress provides many useful functionality out of the box, but if flexibility is your top priority, Blogger is your best choice so far.

I've just finished my first Grails project. To be honest, it's a quite good mind refreshment for me as a Java developer. Grails simplify alot of complex stuff of Java ecosystems like IoC(Spring), ORM(Hibernate), XML configuration with convention over configuration. It takes you less than 10 minutes to create a full working web application with Spring Security for authentication and authorization, Quartz for job scheduling and Hibernate for data persistence when you're familiar with it. Here are some Pros and Cons:
Pros:

• Almost "zero learning effort" for Java developers (although it almost took me a week to get productive with it)
  
• No XML configuration
  
• GORM is much easier to work with comparing to Hibernate (both XML-based and annotation-based), dynamic methods for domain class is definitely a good selling point for me here. It saves me tons of DAO code.
  
• Lots of plugins to extend Grails functionalities

Cons:

• Grails heavily uses convention-over-configuration but the documentation about conventions is not enough. Sometimes it get frustrated when you reading someone's code without knowing the convention his/her using.
  
• Bugs. Lots of it, just navigate to JIRA of Grails and you will know what I mean :)
  
• Lacking of good IDE support. First I've tried Spring STS, then NetBeans, then IntelliJ IDEA Ultimate (free trial) and finally finishing my first Grails project in Gedit!
  

Grails is a great framework and I'm still learning it.

I've uploaded a PKGBUILD for ibus-unikey to the ArchLinux AUR. You can install it with yaourt like this:

yaourt -S ibus-unikey

If you don't want to build from the source, you have to install all the dependencies first:

pacman -Sy gcc gconf gtk2

And then download and install the binary package
Then add the following to your .bash_profile

export XMODIFIERS=@im=ibus
export GTK_IM_MODULE=ibus
export QT_IM_MODULE=ibus

PS: Ok, there is no spyware/malware/rootkit in this binary package! If you don't trust me, don't use this method and build the package yourself (this is the way I recommend)

I've just joined a project as a freelancer to customize a commercial e-commerce solution (so let called it EP). OOTB, the email sending of EP does not support authentication, which is a weird thing for such a cool product like that. Actually, EP does declare an JNDI mail session in the context, but in the code they don't use it. With the source code of EP in hand, things get much easier. First, we declare a JNDI resource in our application context (either in your context file in the conf/Catalina/localhost/ or in your context.xml under META-INF inside your war):

<context>
<Resource name="mail/Session" auth="Container"
    type="javax.mail.Session"
    username="my.smtp.user"
    password="my-secret"
    mail.debug="false"
    mail.transport.protocol="smtp"
    mail.smtp.host= "xxx.xxx.xxx.xxx"
    mail.smtp.auth= "true"
    mail.smtp.port= "25"
    mail.smtp.starttls.enable="true"
    description="Global E-Mail Resource"
/>
</context>

And put those in your web.xml file:

<resource-ref>
    <description>Email Session</description>
    <res-ref-name>mail/Session</res-ref-name>
    <res-type>javax.mail.Session</res-type>
    <res-auth>Container</res-auth>
</resource-ref>

And now, in your Java code, you will use this email session like this:

Context initContext = new InitialContext();
Context context = (Context) initContext.lookup("java:comp/env");
Session mailSession = (Session) context.lookup("mail/Session");
//Your code to send the email goes here

If you get a java.lang.ClassCastException, don't worry, this is a known issue. It's because you have multiple Java Mail jar file and Java Activation Framework jar file in your CLASSPATH!
Hope that may help!

I've just imported all my post and comments from blogger.com to my wordpress.com account. So far so good! I've also cleaned up all of my posts on wordpress.com to make it *cleaner*. Wordpress.com does come with a cool code syntax highlighter as described here http://support.wordpress.com/code/ , the number of programming language it supports is quite impressive.
By the way, I've just found out some cool tricks to use the built-in code highlighter of wordpress. Instead of typing [sourcecode language='css'] code here [/sourcecode], you can use one of the following:

• [source language='css']code here[/source]
  
• [code language='css']code here[/code]
  
• [sourcecode lang='css']code here[/sourcecode]
  
• [source lang='css']code here[/source]
  
• [code lang='css']code here[/code]
  
• [sourcecode='css']code here[/sourcecode]
  
• [source='css']code here[/source]
  
• [code='css']code here[/code]
  

Enjoy!