大漠驼铃

置身浩瀚的沙漠,方向最为重要,希望此blog能向大漠驼铃一样,给我方向和指引。
Java,Php,Shell,Python,服务器运维,大数据,SEO, 网站开发、运维,云服务技术支持,IM服务供应商, FreeSwitch搭建,技术支持等. 技术讨论QQ群:428622099
随笔 - 238, 文章 - 3, 评论 - 117, 引用 - 0
数据加载中……

A Blog Application with Warp (continued)(3)

A Blog Application with Warp (continued)

First check out the previous tutorial on creating a Blog application to get yourself going. In that article, you saw how to add forms and a bit of interactivity. In this one, I'll demonstrate integration of Warp-persist, the JPA/Hibernate module. It is worthwhile reading the introduction to the Warp-persist module first. You will learn the following Warp concepts:

The first thing to do is modify our data model POJO (Blog) and make it a mapped JPA entity:

@Entity
public class Blog {
private Long id;
private String subject;
private String text;
 

@Id @GeneratedValue
public Long getId() {
return id;

//rest of the getters/setters + equals() + hashCode()

}

The most obvious addition here is that of a surrogate key, a Long field id that represents the identity of the object. We have also marked its getter as the identity and the class itself as a JPA entity with the @Id and @Entity annotations respectively.

Remember that you absolutely must override equals() and hashCode() -- this is done by simply returning the equals and hashCode of the id field. It is a poor programming practice to create data model objects and not define equals() and hashCode() for them, it will lead to all kinds of problems if you do not.

OK, let's drop in a simple persistence.xml into our /META-INF directory so that Hibernate knows our configuration. I've stolen this from the Warp-persist guide:

<?xml version="1.0" encoding="UTF-8" ?>
<persistence xmlns="http://java.sun.com/xml/ns/persistence"
	xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://java.sun.com/xml/ns/persistence
http://java.sun.com/xml/ns/persistence/persistence_1_0.xsd" version="1.0">

<!-- A JPA Persistence Unit -->
<persistence-unit name="blogJpaUnit" transaction-type="RESOURCE_LOCAL">
<provider>org.hibernate.ejb.HibernatePersistence</provider>
 
		<!-- JPA entities must be registered here -->
<class>com.wideplay.misc.blog.Blog</class>

<properties>
			<!-- vendor-specific properties go here -->
</properties>
</persistence-unit>

</persistence>

Not much to notice here except that Hibernate is our persistence provider and we're using RESOURCE_LOCAL transactions. Inside the <properties> tag, however, you need to place all your configuration options (for whatever JPA vendor you choose). For Hibernate, mine looks like the following:

<property name="hibernate.connection.driver_class" value="org.hsqldb.jdbcDriver"/>
<property name="hibernate.connection.url" value="jdbc:hsqldb:file:C:/temp/hsqltmp"/>
<property name="hibernate.connection.username" value="sa"/>
<property name="hibernate.connection.password" value=""/>
<property name="hibernate.connection.pool_size" value="1"/>
<property name="hibernate.dialect" value="org.hibernate.dialect.HSQLDialect"/>

<property name="hibernate.current_session_context_class" value="thread"/>
<property name="hibernate.cache.provider_class" value="org.hibernate.cache.NoCacheProvider"/>
<property name="hibernate.hbm2ddl.auto" value="create"/>

 Again, it is mostly boilerplate. Never use this configuration directly in production, the following things that it assumes are almost certainly not what you will want in a real-world application:

  • Im using HSQLDB (a temporary in-memory database)
  • I am using Hibernate's default connection pool
  • I am creating the schema every time Hibernate starts up (hbm2ddl.auto=create)
  • I am not using any second-level cache

Otherwise everything is fine. The session-context strategy is set to thread meaning that the current thread (in our case, that associated with a request) will be used as the context for an EntityManager. So different requests will see different EntityManagers.

Now, we need to tell Warp that the persistence service is ready for use. This is done in the WarpModule (which you should remember from the Hello World example). This is roughly how your WarpModule should look:

public class BlogModule implements WarpModule {

public void configure(Warp warp) {
warp.install(PersistenceService
.usingJpa()
.across(UnitOfWork.TRANSACTION)
.buildModule()
);

warp.install(new AbstractModule() {
protected void configure() {
bindConstant().annotatedWith(JpaUnit.class)
.to("blogJpaUnit");
}
});

warp.addStartupListener(BlogStartupListener.class);
}
}

 And implement the startup listener to do some initialization work:

   public class BlogStartupListener implements StartupListener {
@Inject PersistenceService service;

public void onStartup() {
service.start();
}
}

 This tells Warp-persist to start up JPA and make the persistence layer ready when Warp starts up. Likewise, you must tell Warp when to shutdown too (services like connection pools require explicit shutdowns):

   public class BlogShutdownListener implements ShutdownListener {
@Inject EntityManagerFactory emf;

public void onShutdown() {
emf.close();
}
}

 Don't forget to add the BlogShutdownListener in your WarpModule configuration. You do not need to release any Warp resources explicitly (Warp acts on the Java Servlet destroy event).

OK, wow, we're done a lot of setup work. Let's actually convert our blog's pages to use all this neat configuration. First the ListBlogs page:

@URIMapping("/home")
public class ListBlogs {
@Finder(query = "from Blog", returnAs = ArrayList.class)
public Collection<Blog> getBlogList() {
return null;
}
}

 It looks completely different now! What happened? We've done a few things:

  • Gotten rid of the HashMap and @Singleton scoping
  • Converted our property getter for the blog list into a Dynamic Finder
  • Implemented a dummy stub for the Dynamic Finder

The finder will be intercepted by warp-persist and replaced with the query provided in the @Finder annotation. In this case, we're selecting every Blog available in the persistent store and returning them as an ArrayList.

Now lets do the Read page, this is similarly very simple. We will use a Dynamic Finder encapsulated in our @PreRender event handler method:

@URIMapping("/blog/{id}")
public class ViewBlog {
private Blog blog;

@Finder(query = "from Blog where id = :id")
Blog fetchBlog(@Named("id") String id) {
return null;
}

@OnEvent @PreRender @Transactional
public void onView(String id) {
this.blog = fetchBlog(Long.valueOf(id)); //dangerous, validate id first!
}
}

Neat, because the page objects are managed by guice, dynamic dispatch will ensure that calling fetchBlog() invokes the intercepted Dynamic Finder instead of our dummy stub which returns null. It is important to make the @PreRender event handler transactional because we are using a session-per-transaction strategy. Let's do the same with ListBlogs too:

@URIMapping("/home")
public class ListBlogs {
@Transactional
@Finder(query = "from Blog", returnAs = ArrayList.class)
public Collection<Blog> getBlogList() {
return null;
}
}

Typically we would prefer to encapsulate data access into an Accessor (Dao) style object. and run transactions on much coarser grained boundaries. But since our use case is very simple, I will leave it to your imagination to separate data access into its own semantic module/layer.

Lastly for the good part: storing new blogs!

@URIMapping("/blogs/compose")
public class ComposeBlog {
@Inject @Page private ListBlogs listBlogsPage;

@Inject private Provider<EntityManager> em;

private Blog newBlog = new Blog("", ""); //an empty blog

@OnEvent @Transactional
public ListBlogs save() {
//save the new blog!
em.get().persist(newBlog);

//return to the list page
return listBlogsPage;
}

public Blog getNewBlog() {
return newBlog;
}
}

This is fairly straightforward and should not need much explanation. We associate the new blog entity with the EntityManager by calling persist() and ensure that the button event handler is transactional.

There is almost no change to the templates themselves, except that now we have switched the identity of our data object Blog to use a surrogate key. Let's use that key to reference blogs in our ListBlogs.html template:

<?xml version="1.0" encoding="UTF-8"?>
<html xmlns:w="http://www.wideplay.com/warp/schema/warp_core.xsd"
xmlns="http://www.w3.org/1999/xhtml"
xml:lang="en" lang="en">

<head w:component="meta">
<title>Warp :: List Blogs</title>
</head>

<body w:component="frame">
<h1>A list of blog entries</h1>

<table w:component="table" w:items="${blogList}">
<td w:component="column" w:property="subject">
<a w:component="hyperlink" w:target="blog" w:topic="${id}">
${subject}
</a>
</td>
</table>

<a href="blogs/compose">compose new entry</a>

</body>
</html>
And that's it!

posted on 2009-02-17 16:31 草原上的骆驼 阅读(241) 评论(0)  编辑  收藏


只有注册用户登录后才能发表评论。


网站导航: