Blok blogs

I have a new blog using the static site generator Pelican instead of Wordpress.

There appears to be no way to update the PLAY HOME in IntelliJ / Idea from the UI (version 11.1.2). So if you need to change the configuration of the “Play with Playframework” feature then you need to do it from the config file. On Linux the PLAY HOME is stored in this file:

~/.IntelliJIdea11/config/options/other.xml

And this is the entry you need to change:

  <component name="PlayframeworkConfiguration" home="$USER_HOME$/programs/play-2.0.1" />

Note that you need to update the file after you closed IntelliJ, otherwise the old value will be rewritten…

On starting IntelliJ 11.1 I got the following exception:

Java HotSpot(TM) 64-Bit Server VM (build 20.1-b02, mixed mode)
ERROR: Failed to create temporary file for jnidispatch library: java.io.IOException: Permission denied
java.lang.Error: Failed to create temporary file for jnidispatch library: java.io.IOException: Permission denied
    at com.sun.jna.Native.loadNativeLibraryFromJar(Native.java:762)
    at com.sun.jna.Native.loadNativeLibrary(Native.java:697)
    at com.sun.jna.Native.<clinit>(Native.java:127)
    at com.intellij.openapi.util.io.FileSystemUtil$JnaUnixMediatorImpl.<init>(FileSystemUtil.java:280)
    at com.intellij.openapi.util.io.FileSystemUtil$JnaUnixMediatorImpl.<init>(FileSystemUtil.java:264)
    at com.intellij.openapi.util.io.FileSystemUtil.<clinit>(FileSystemUtil.java:62)
    at com.intellij.openapi.util.io.FileUtil.delete(FileUtil.java:594)
    at com.intellij.ide.startup.StartupActionScriptManager$DeleteCommand.execute(StartupActionScriptManager.java:225)
    at com.intellij.ide.startup.StartupActionScriptManager.executeActionScript(StartupActionScriptManager.java:47)
    at com.intellij.ide.ClassloaderUtil.initClassloader(ClassloaderUtil.java:127)
    at com.intellij.ide.Bootstrap.main(Bootstrap.java:38)
    at com.intellij.ide.Bootstrap.main(Bootstrap.java:34)
    at com.intellij.idea.Main.main(Main.java:93)
Exception in thread "main" java.lang.AssertionError: Failed to create temporary file for jnidispatch library: java.io.IOException: Permission denied
    at com.intellij.openapi.diagnostic.DefaultLogger.error(DefaultLogger.java:50)
    at com.intellij.openapi.diagnostic.Logger.error(Logger.java:69)
    at com.intellij.openapi.util.io.FileSystemUtil.<clinit>(FileSystemUtil.java:66)
    at com.intellij.openapi.util.io.FileUtil.delete(FileUtil.java:594)
    at com.intellij.ide.startup.StartupActionScriptManager$DeleteCommand.execute(StartupActionScriptManager.java:225)
    at com.intellij.ide.startup.StartupActionScriptManager.executeActionScript(StartupActionScriptManager.java:47)
    at com.intellij.ide.ClassloaderUtil.initClassloader(ClassloaderUtil.java:127)
    at com.intellij.ide.Bootstrap.main(Bootstrap.java:38)
    at com.intellij.ide.Bootstrap.main(Bootstrap.java:34)
    at com.intellij.idea.Main.main(Main.java:93)

It appeared that this dir needs to have write rights:

/tmp/jna

which in my case was not so. After doing a chmod on this directory IntelliJ started without a problem.

This is a post that summarizes how you can integrate Liftweb in an existing Spring-based web application.
The description is based on an actual example project of which the code is available at the Scala Spring project at Google code.

Before we start with the recipe, first some notes about the assumptions for this integration:

• This recipe assumes Lift 1.0.2 and is about putting Lift’s webkit module to use (Lift also has modules for data access for example, but we will still use Spring for data access).
• It also assumes that all Spring objects that the Lift-based code needs to access are available in the root Spring ApplicationContext. So in the example project they are available in applicationContext-jdbc.xml as opposed to the servlet specific ApplicationContext petclinic-servlet.xml.

Ok, the recipe:

1. Specify the necessary dependencies.

           <properties>
                   <scala.version>2.7.5</scala.version>
                   <lift.version>1.0.2</lift.version>
           </properties>
   
           <dependencies>
                   <dependency>
                           <groupId>org.scala-lang</groupId>
                           <artifactId>scala-compiler</artifactId>
                           <version>${scala.version}</version>
                   </dependency>
                   <dependency>
                           <groupId>org.scala-lang</groupId>
                           <artifactId>scala-library</artifactId>
                           <version>${scala.version}</version>
                   </dependency>
                   <dependency>
                           <groupId>net.liftweb</groupId>
                           <artifactId>lift-webkit</artifactId>
                           <version>${lift.version}</version>
                   </dependency>
                   <dependency>
                           <groupId>net.liftweb</groupId>
                           <artifactId>lift-util</artifactId>
                           <version>${lift.version}</version>
                   </dependency>
           </dependencies>
   

   See: pom.xml

2. Add a filter and filter-mapping to the web.xml. For example:

   	<filter>
   		<filter-name>LiftFilter</filter-name>
   		<filter-class>net.liftweb.http.LiftFilter</filter-class>
   	</filter>
   
   	<filter-mapping>
   		<filter-name>LiftFilter</filter-name>
   		<url-pattern>lift/*</url-pattern>
   	</filter-mapping>
   

   See: web.xml

3. In Lift’s Boot class configure where Lift should look for snippets. For example:

   LiftRules.addToPackages("org.springframework.samples.petclinic.lift")

   See Boot.scala

4. Create a Lift template

   <lift:surround with="default" at="content">
     <head><title>Owner</title></head>
   
     <lift:OwnerSnippet.add form="post">
     <h2><owner:newText/>Owner:</h2>
     <table>
       <owner:firstName />
       <owner:lastName />
       <owner:address />
       <owner:city />
       <owner:telephone />
       <owner:submit />
     </table>
     </lift:OwnerSnippet.add>
   </lift:surround>
   

   See ownerForm.html. This particular template uses a default template via the lift:surround element to wrap it in some standard html, but you can organize it any way you like.

5. And finally a snippet:

   class OwnerSnippet {
   
     // Set up a requestVar to track the object for edits and adds
     object ownerVar extends RequestVar(new Owner())
     def owner = ownerVar.is
   
     def add (xhtml : NodeSeq) : NodeSeq = {
   
       def doAdd () = {
         // Reusing the spring validator code
         val result = new MapBindingResult(new HashMap(), "owner")
         new OwnerValidator().validate(owner, result)
         LiftUtils.toLiftErrors(result)
         if (!result.hasErrors()) {
           try {
             LiftUtils.clinic.storeOwner(owner)
             S.redirectTo("/owner.do?ownerId=" + owner.id)
           } catch {
             case e : DataAccessException => error("DataAccessException")
           }
         }
       }
   
       // Hold a val here so that the "id" closure holds it when we re-enter this method
       val currentId = owner.id
   
       val submitText = if (owner.isNew) "Add Owner" else "Update Owner"
   
       bind("owner", xhtml,
        "newText" -> (if (owner.isNew) "New " else ""),
        "id" -> SHtml.hidden(() => owner.id = currentId),
        "firstName" -> LiftUtils.field("First Name: ", owner.firstName, owner.firstName = _, "size" -> "30", "maxlength" -> "80", "id" -> "firstName"),
        "lastName" -> LiftUtils.field("Last Name: ", owner.lastName, owner.lastName = _, "size" -> "30", "maxlength" -> "80", "id" -> "lastName"),
        "address" -> LiftUtils.field("Address: ", owner.address, owner.address = _, "size" -> "30", "maxlength" -> "80", "id" -> "address"),
        "city" -> LiftUtils.field("City: ", owner.city, owner.city = _, "size" -> "30", "maxlength" -> "80", "id" -> "city"),
        "telephone" -> LiftUtils.field("Telephone:", owner.telephone, owner.telephone = _, "size" -> "20", "maxlength" -> "20", "id" -> "telephone"),
        "submit" -> <tr><td><p class="submit"> { SHtml.submit(submitText, doAdd) } </p></td></tr>
       )
     }
   
   }
   

Note that in this example we are reusing the validation code that was originally written for Spring (line 10-14).

Also note that it is up to you whether you define your code as a template or snippet. E.g. it would be easy to translate the ownerForm template into a snippet, or vice versa to move more code into the template.

Furthermore, check out the nifty way how Lift enables you to write the add method. What I especially like is the way the type safe way the owner is constructed (lines 4 and 5), how binding works (lines 29 and further ) and the way Scala and Lift allow you to implement a form submit (see line 37 for the call to doAdd and the method definition of doAdd starting at line 9).

On the other side, a Spring controller allows you your dependencies to be injected whereas in Lift you need to call out to obtain the objects you require. For example, in the Spring controller AddOwnerForm.scala we have a setClinic method that injects a Clinic. On the other side, in the Lift snippet we need to call out to the ApplicationContext via the ServletContext to obtain the Clinic object (see line 16 LiftUtils.clinic and it’s implementation).

By the way, there has been an interesting discussion about dependency Injection in Lift. Read that and decide what your opinion is.

Of the many ways of writing a DSL in groovy I prefer the approach that is taken in
a Groovy DSL from scratch in 2 hours. That article describes a general way to create a DSL and you don’t have to write too much code for it. Moreover it takes the groovy Script as a base of the DSL which is really quite powerful. But there’s still too much unnecessary complexity that you will have to deal with if you want to create a DSL (see the code dealing with ExpandoMetaClass, closures and delegates in the article).

The goal of my experiment is to further simplify creating a DSL. I want to adapt groovy Scripts in such a way that it becomes easy to access the parts of a script and use them to construct your DSL.
For this reason I choose to combine scripts with Expando’s in such a way that closures become accessible as Expando’s in a recursive way. The class that adapts scripts like this is ExpandoDsl.

This post shows what you have to write to create a DSL using (the current version of) ExpandoDsl. I will simply take the architecture example from a Groovy DSL from scratch in 2 hours so it’s easy to compare the code you have to write. Note though that I didn’t rewrite the architecture DSL completely, but it’s enough to make Steven’s example working and it should give you a good idea of what you need to write to create a DSL.

Here is the code:

package nl.rintcius.example.dsl.architecture

import com.seventytwomiles.architecturerules.configuration.Configuration
import com.seventytwomiles.architecturerules.domain.SourceDirectory
import com.seventytwomiles.architecturerules.domain.Rule
import com.seventytwomiles.architecturerules.services.CyclicRedundancyServiceImpl
import com.seventytwomiles.architecturerules.services.RulesServiceImpl

import org.apache.log4j.*

import nl.rintcius.groovy.ExpandoDsl

public class GroovyArchitecture {

    static void main(String[] args) {
        BasicConfigurator.configure()
        LogManager.rootLogger.level = Level.INFO

        runArchitectureRules()
    }

    static void runArchitectureRules() {
        runArchitectureRules(new File("architecture.groovy"))
    }

    static void runArchitectureRules(File dsl) {

        Expando root = new ExpandoDsl(dsl.text).create()

        Configuration configuration = createConfiguration(root.architecture)

        new CyclicRedundancyServiceImpl(configuration).performCyclicRedundancyCheck()
        new RulesServiceImpl(configuration).performRulesTest()
    }

    static Configuration createConfiguration(Expando expando) {
        Configuration configuration = new Configuration()
        configuration.doCyclicDependencyTest = true
        configuration.throwExceptionWhenNoPackages = true

        expando.properties.each() { key, value ->
            if (key == "jar") { configuration.addSource( new SourceDirectory(value.value, true)) }
            else if (key == "rules") { createRules(value).each() { rule -> configuration.addRule(rule)} }
        }
        configuration
    }

    static List createRules(Expando expando) {
        expando.properties.collect() { key, value ->
            createRule(key, value)
        }
    }

    static Rule createRule(String ruleName, Expando expando) {
        Rule rule = new Rule(ruleName)
        expando.properties.each() { key, value ->
            if (key == “comment”) { rule.comment = value }
            else if (key == “violation”) { rule.addViolation(value.value) }
            else if (key == “package”) { rule.addPackage(value.value) }
        }
        rule
    }
}

As you can see, the code is quite easy to follow. All you have to do is build the root Expando via
new ExpandoDsl(text).create() and then traverse the Expando’s via expando.propertes and build your structure.

Finally, apart from ExpandoDsl there’s also ExpandoBuilder if you prefer the more restrictive builder syntax.

Let me know what you think of it!

Need to do custom RPC serialization for a class in GWT?

Here’s how. E.g. for the class my.example.MyObject create a class named my.example.MyObject_CustomFieldSerializer. And then implement

    public static void serialize(SerializationStreamWriter streamWriter, MyObject myObject)
        throws SerializationException {}

    public static void deserialize(SerializationStreamReader streamReader, MyObject myObject)
        throws SerializationException {}

    //This method is optional
    public static MyObject instantiate(SerializationStreamReader streamReader)
        throws SerializationException {}

For more info, see CustomFieldSerializer (somehow this page didn’t show up when I googled for it, only after I knew the term was CustomFieldSerializer).

Just played around a bit with WolframAlpha.

I can see it has added value especially for scientists and engineers. It’s a kind of Mathematica with a web front end and a large dataset to obtain the information from. Its use will largely be depending on the quantity (and quality of course) of the underlying dataset. To give you an impression of its current state, the data it has of Balkenende (Prime Minister of the Netherlands) is still limited (what is his salary?) and nothing compared to the data that is provided on Wikipedia.

On the other hand, it’s quite nice to be able to quickly see some math plots and it’s also good in doing conversions and quick calculations such as 3 feet or time in tokyo.

All in all, I think I will use WolframAlpha every now and then as a companion to Wikipedia.

After the old NKS, Wolfram is about to present another (and better?) NKS version to the world. A new kind of search that is surely making me curious! Hope it will live up to my expectations when it’s getting live.

See this preview (the video is low quality but gives an impression).

Next to searching from the Lenzcape application, you can also use the search plugin to search directly from the browser’s searchbar. This is how I am using Lenzcape myself usually.

The searchbar is particularly easy when you know the area that you want to search (that’s the lenz in lenzcape’s terminology). Further, the more you need to search in a specific area, the more it will pay off to create a lenz for that area.

Lenzcape’s searchbar in Firefox

To illustrate what I mean, let’s take grails as an example. I used grails to develop Lenzcape, so I am often searching grails related stuff. The base lenz for grails is “dev/grails” and following the search syntax you separate the lenz part of the query and the terms that you want to search for with a “)”. In this way you can type dev/grails ) redirect in the searchbar and it will search for redirect just in the grails area.

But you can also search more specific grails topics such as:

• grails issues: dev/grails/issue ) redirect
• grails reference documentation: dev/grails/doc ) redirect or if you want to search a particular version dev/grails/doc/1.0 ) redirect or dev/grails/doc/1.1 ) redirect
• grails api: dev/grails/api ) redirect, dev/grails/api/1.0 ) redirect or dev/grails/api/1.1 ) redirect
• grails on git: dev/grails/git ) redirect

These lenzes all work for you as well because they are predefined. But you can also extend lenzes by integrating them with your own bookmarks. For example, I have included a few grails blogs with tag dev/grails/blog to my personal bookmarks. In this way I can search my bookmarked grails blogs via dev/grails/blog ) redirect. Note that if you type this using the search bar you will get no results unless you have also added bookmarks with tag “dev/grails/blog”.

In a former post I wrote that I would show how you can build your own lenzes in Lenzcape.

Initially I intended to show it in a blog post, but later when I got the idea to make a live demo, I thought it would be clearer (and cooler) to include it in the live demo. So check out the live demo if you want to see how you can build your own lenzes!