SUN TECHNOLOGIES: JAVA WEB &

MOBILE

Loc ROBERT, SUN

Created by XMLmind

XSL-FO Converter.

SUN TECHNOLOGIES: JAVA WEB & MOBILE

by Loc ROBERT
DESCRIPTION_CURSUS
Discovery

of

Web

with

Java

http://www.supinfo.com

and

Java

Micro

Edition

Chapter 1. Introduction JEE

Discover Java EE

What is it used for ?

How to use it ?

Current mistakes

Design an application

Know and choose libraries

1. Course
1.1. Introduction
1.1.1. Why should I use an application platform ?
An application platform is a generic application which provides useful functionalities for your own applications.
Its functionalities are shared with the applications that are running on it. There are different kinds of platforms,
from the most generics to the most complex (developed for specific needs, for example). A lot of companies
have created their own platform, like IBM with WebSphere, BEA with WebLogic, etc.

Database connections

Object Containers

Configurations

....

1.1.2. What is JEE?

JEE (Java Enterprise Edition) is an agreement created by Sun MicroSystem. Many companies work with them in
order to develop their own implementation of the components. Most of the times, we talk about JEE plateforms to define the API and the execution structure. JEE is composed of:

The Application Server, which is the execution structure

The services, through APIs

Sun provides a minimal implementation, through APIs called JEE SDK (JEE Software Development Kit). JEE
works only with Java, so it is specific to its advantages and drawbacks.
3
http://www.supinfo.com

For the advantages, we can say that it is portable and easy to maintain.
Furthermore, the JEE architecture uses components, which can be changed together and shared, which means
that:

It is simple extend an application

An application developed with JEE can use high availability, in order to guarantee a good quality of service

The maintenance of the applications is facilitates

1.1.3. The actors of a JEE application

The realization of an application based on architecture JEE calls various types of competences which are
difficult to find at the same person because that goes from the design until the supervision of the application
while passing by the development and the deployment. In order to be able to control this process JEE adopts the
approach of the shared responsibilities. More specifically for the EJB, this approach defines several levels of
responsibility:

The EJB supplier: is the actor who provides reusable business components either by the purchase to a supplier
of components, or by internal development

The assembler of applications: is the actor who is able to build an application starting from existing EJB

The installer : is the actor who deploy the application on an application server

The administrator: is the actor who controls the operation of the application server and ensures the
supervision of the applications

The supplier of container: is the editor who markets a Web container or an EJB container, this editor often
markets an application server which includes these containers

The server supplier: s the editor who markets the application server (BEA, IBM, etc)

1.2. More information about JEE

1.2.1. JEE Developpement Kit
JEE describes the architecture of a standard for application server. We will more precisely study the
components, the dependences and the protocols.

4
http://www.supinfo.com

The various rectangles define the containers (JEE execution environment) which provide the services for the
various components (represented by the rectangles in the rectangles). The whole of the components will be
explained in the following chapter. The arrows represent the connections between the components of the
application server. For example, the customer application can be connected to the "Web Container" by the
intermediary of the JSP/Servlet, it can be also connected to the "EJB Container", etc.
Now, the diagram of connections between the platform JEE and the other application (various protocols used).

This diagram is very important in the field of connection between various applications. It provides information
concerning the protocols used for each possible remote connection.
For example, "Web Container" provides accesses via HTTP/SSL or SOAP; "EJB Container" provides accesses
HTTP/IIOP (RMI: Internet Inter-Orb Protocol)/SSL, etc.
We can then think that a customer using C #, for example, can connect on a EJB in HTTP mode or via the IIOP
protocol (more widespread).
5
http://www.supinfo.com

1.2.2. The low-level tools

We have just seen the architecture of the JEE platform. We will now introduce you various tools.
There are 3 great types of tools:

Components

Infrastructure services

Communication services
1.2.2.1. Components
There is 2 great categories of components:

Web

Bussiness
It is the presentation layer. The client receives only HTML text, but it is only the visible part of the application.
Behind the scene, different technologies are working to allow the code to be powerful and easy to maintain.
WEB

JSP
JSPs (Java Server Pages) are web pages which generate HTML to the client user interface. In this text file
both HTML and Java code (called Scriptlet), and even tags are linked to java code (tag-lib). This technology is
useful to generate HTML pages, but not very much to convenient to treat a request from the user. This part is
called the View because it is related to the visible part of your application for your client.

Servlet
A servlet is a java class which allows to treat a request from the client. Then, once your treatment will be
performed, you will have to redirect the client to another view. This part is the Controller by it controls the
actions of your application. This Java class should never generate HTML code (or else as less as possible).
Business - EJB (Entreprise JavaBean)
It is specific components which are supposed to treat the data of your application (called business methods) and
this connection with the data base. This part is the View Layer of the model MVC because it is related to the
visible part of your client application.
1.2.2.2. Connection services
JDBC - Java Data base Connectivity

JDBC - Java Data base Connectivity : It is a data base connection API. It is provided too with the standard
version of java, but it has been improved by the application server by creating database connections pools. It
reduces the number of lines to write, and it optimizes the performance

6
http://www.supinfo.com

JNDI : With this API, you will be able to retrieve remote objects, and to access to directories like DNS, NIS,
LDAP, etc

JTA / JTS - Java Transaction Api / Java Transaction Services : This API defines the transaction manager

JCA (JEE Connector Architecture) : It is an API which allows you to connect different informations systems
together in you organization. It work most of the time with Legacy systems, like ERP

JMX (Java Management eXtension) : This API provides you a way to monitor your applications
1.2.2.3. Communications services

JAAS (Java Authentification and Authorization Service) : This is an authentification and permission manager
API.

RMI (Remote Method Invocation) : This API allows you to communication through remote objects.

Web services : The Web services allow you to share treatments. Web services are executed remotely. This
technology uses XML, it allows you to share your application with many different platforms and languages.
1.2.2.4. JMS (Java Message Service)
This API provides a way to create an asynchronous communication.(called MOM for Middleware Object
Message) between applications.

JavaMail : It is an API which allows to send web mail..

1.2.3. JEE Implementation: the application servers

It is necessary to define what a server application is. Indeed, most of the people do not understand the difference
between an application server and object server.
1.2.3.1. What is an application server
The application servers are execution environments for your server-side applications. It handles automatically
most of the configurations, and connects your application with clients:

User session Manager: define a specific session for each user of your application, even if they are using the
same objects. Most of the application servers create an user id for each new client. The client retrieves this
value each time it communicates with the server. In order to retrieve this value, you can use hidden fields, a
cookie, or a parameter in your request url.

Load balancing and recovery of the accidents: In order to manage more users, the application server needs to
be deployed on different computers. Furthermore, it is possible to recover your data if there is a problem.
Although, most of the time, the load balancing is only working on the network side with a DNS round-robin,
a reverse proxy, etc).

Manage database connections: The application server can manage your data connection to different data
source. This container is called the databases pool connections.

7
http://www.supinfo.com

...
The application server is necessary if you want to avoid re-writing these functionalities (like for GGI). By this
definition, the JSP/Servlets containers, Microsoft ASP, Cold Fusion, and PHP are applications server (even if
they are integrated with PHP/ASP web servers).
1.2.3.2. What is an object server
If you want to understand what is an object server, you have to understand that there is two ways to create your
applications:

The first one consists to access the data sources directly.

The second one consists in developing business components (clients, providers, etc) in order to hide the
difficulties of the data access. For example, if you have got an account in a bank, your account will have two
methods, one to add money, and another one to check out the money. Thus, the hidden part modify the values
automatically for you, and the data source can be either an ERP (Entreprise Resource Planning),a
CRM(Customer Relation Ship Managment) or a database.

1.
The client sends a request
2.
The web server transmit the request to the application server
3.
The application server treats your request by using the object server
4.
The object server treats your data through the database
5.
The object server returns the objects to the application server
6.
The application server returns the result to the web server
7.
The web server returns a response to the client
In order to manager these objects, an environment is necessary : the object server. This server provide different
services of the application server.

Object persistence manager

8
http://www.supinfo.com

Business object transaction manager

Load balancing: here the server does not work like an application server
Thus, as we have seen previously, the concepts are different. Most of the application servers today are EJB
server (Entreprise Java Beans). It is a necessary part only if you have important needs and you want to use the
whole business functionalities.
It is important to understand what is a difference between an object servers and an application server, or else
you will not buy the good one, like 80% of the customers. In fact, most of the organizations only need a
servlet/jsp container, and an EJB container is too difficult and useless in most of the cases.
On the market, the application servers are widely presents, more than the object servers.
During the past years, the application servers market, the standardization has decreased the number of offers.
From dozen of solutions, there is only three left today: Java, Microsoft .NET and PHP. We advise you to avoid
other solutions.
Here is a brief description of these offers:

PHP 4: The scripting applications are really easy to write, and able to perform a good job. Nevertheless, they
are difficult to maintain because the code has not a good structure

Java, Microsoft .NET, PHP 5: The object oriented solutions allow you to write a better code, easy to
maintain, without having the difficulties of the business objects

Java EJB: The business objects are more difficult and more expensive to program. You will need an
application server with an EJB Container. For example, Tomcat has a web container, but he has no EJB
container
Including this three families of solutions, Open-Source products are available for the development and the
production (Tomcat, NetBeans, JBoss, Jonas ).
The corner stone of your application server choice are:

The treatment capacity like load balancing

Ease of use

Compatibility with your own applications

You computer resources (IBMs application server is more heavy than Tomcat)
.

1.3. JEE in practice

1.3.1. JEE v.s. its competitors
Each development platform has its advantages and drawbacks. The first advantage of this platform is that it was

9
http://www.supinfo.com

adopted by the greatest groups in the whole world(IBM, Oracle, BEA..).

It is stable and reliable-it exists and moves constantly since 1998 . Even if it is Sun Microsystems which
organizes the specifications of this platform, it has learnt to listen to the advise of its developers since the
beginning . Sun set up a system of specifications for which the developers can indicate their needs, solutions or
dissatisfaction.
Moreover, the platform in the basis of Java, makes it possible to have applications completely independent of
the used platform system (Windows, Linux or Mac OS).
Its principal "competitor" is the Net platform (developed by Microsoft). This platform, although easier to use,
lack of maturity. Even if it attract some companies, it is used rather for the less important projects.
Moreover, interworking JEE and Net is possible.

1.3.2. Standard components against framework

We introduced you JEE with its whole components. However it is sometimes heavy to use a component
designed for great architecture whereas our application is still simple.
The OpenSource community has been developping frameworks.
Often more limited than the standard componant, the framework are simpier of use and more powerful in certain
cases.
1.3.2.1. EJB vs Hibernate & Spring
The EJBs can sometimes be the nightmares of developers JEE. Indeed, they are hard to set up and are often
heavy to use. They are generally integrated in the projects of great scales.
For the more current and smaller projects, the Hibernate and Spring frameworks can replace these EJB.
PRESENTATION
The EJBs manage the access and the data processing (persistent or not). To make the same thing with the
external use of framework, it is necessary to use two of them:

Hibernate: it is a framework which allows to "map" a relational data base into object (POJO: Plain Old Java
Object). Thus, It makes it possible to completely abstract the access to the data base and thus proposes an
access completely Object vision of the data

Spring: it is a framework which makes it possible "to replace" the heaviness of the application servers.
Indeed, one speaks about "light container". It deals with the creation and the connection of objects by the
intermediary of a configuration file describing the whole of these relations. One of the principal advantages is
that it does not force to inherit or to implement an interface or a class, contrary to the EJB
HOW TO USE
Using these two frameworks in the same application is recommended. Indeed Hibernate enables you to reach the
data whereas Spring is useful to you for "automated factory"! Moreover these both frameworks are very easily
to use and requires only one engine of servlet., contrary to the EJB which require an application server (much
heavier!).
1.3.2.2. Servlet & JSP vs Struts
With the arrival of the JSP, after the Servlets, the programmer could start to separate the presentation layer and
the application layer/treatment in a remarkable way . However the maintenance of the code and the heaviness of
use of the servlets showed their limits.
Model MVC (Model View Controller) that each developer thought on its side each time was too much limited
and not very evolutionary.
10
http://www.supinfo.com

The arrival of Struts gave a solid base answering a good MVC architecture.
1.3.2.2.1. Presentation
Struts is a framework which is designed to build Web applications. It is based on Servlet technology/JSP by
adding the an MVC2 Model (Model View Controller). It provides the frame of a Web application and avoids the
developers having to manage in a tiresome Modle View Controller separation.
1.3.2.2.2. How to Use
The use of this framework is often use in the development of Web applications. However Struts is not the only
one technology helping with the development of the application layer/Web presentation. Indeed, you can also
use JSF (Java Server Faces) or Cocoon...
1.3.2.2.2.1. Application Server: utilization and limits
The Application Servers have been developed since the creation of JEE. We can notice 2 mains categories of
servers:

Open Source : evolve thank to communities

Commercial : evolve according to the editor

Each kind has its own advantages and drawbacks. We will describe the most important server, in order to have a
wide vision of the marke
1.3.2.3. Open Source
1.3.2.3.1. Tomcat : Apache
Tomcat is a servlet container which implements the official references for the Servlet and JSPs. This server is
widely used for web applications.
Sources: http://jakarta.apache.org/tomcat/index.html
Related technologies:

JSP

Servlet

JDBC

JNDI
1.3.2.3.2. Jonas : ObjectWeb
Jonas is an application server which implements the official references for the EJBs. It embed tomcat in order to
create web applications.
Sources : http://jonas.objectweb.org/
Related technologies:

JSP
11
http://www.supinfo.com

Servlet

EJB

JCA

JDBC

JTA

JMS

JMX

JNDI

JAAS

JavaMail

1.3.2.3.3. JBoss : JBoss

JBoss has the same functionalities as Jonas. Nevertheless, its architecture is quiet different and works around a
BUS .
Many differents project are under developpement behind the server, like eclipse plugins, AOP modules (Aspect
Oriented Programming), Hibernate
JBoss is one of the most famous open-source application servers. It is more and more used by professionals.
Sources: http://www.jboss.org/products/jbossas

12
http://www.supinfo.com

1.3.2.3.4. GlassFish
In computing, GlassFish is an application server project by Sun Microsystems for the Java Enterprise Edition
(Java EE) platform. The commercial version is called Sun Java System Application Server 9.x.
GlassFish is free software, dual-licensed under two free software licences: the Common Development and
Distribution License (CDDL) and the GNU General Public License (GPL) with the classpath exception.
GlassFish is the name for the open source development project for building a Java EE 5 application server. It is
based on the source code for Sun Java System Application Server PE 9 donated by Sun Microsystems and
TopLink persistence code donated by Oracle. This project provides a structured process for developing a high
quality application server that makes new features available faster than ever before. It is the response to Java
developers who want access to the source code and the ability to contribute to the development of Sun's next
generation application server which is based on GlassFish. This project is designed to encourage communication
between Sun and Oracle engineers and the community and will enable all developers to participate in the
application server development process.
Sources : http://glassfish.dev.java.net
1.3.2.3.5. Commercials
The commercials application server are famous because they supply useful tools. They help you to configure
your applications.
Although, you have to pay to use them
The most important commercial application servers:

WebSphere : IBM

WebLogic : BEA

WebObject : Apple

Oracle Application Server : Oracle

1.3.3. IDE (Integrated Development Environment)

13
http://www.supinfo.com

To develop important applications, you are to use an IDE (Integrated Environnement Development).
Some often them are open-source, and freely usable.
Here is a short presentation of the most important ones.
1.3.3.1. Open Source
1.3.3.1.1. Eclipse : IBM

A very good tool because it is simple to use, and it allows to the beginners to learn java. Furthermore, a lot of
plugins are available, and they will help you to go farther in your learning path.
Link: http://www.eclipse.org
1.3.3.1.2. Sun: NetBeans

14
http://www.supinfo.com

A tool specialized for the professionals because it allows very powerful functionalities. The generation of code
and the use of modelling tools are very good. However it is permeable to the beginners, for whom Eclipse is
adapted better and it remains very expensive. It is based upon Eclipse.
Link: http://www-306.ibm.com/software/sw-bycategory/subcategory/SW710.html
1.3.3.1.3. XCode & WebObject : Apple

15
http://www.supinfo.com

Apple has a very good Application Server and IDE, mostly based on WYSIWYG, while keeping a clean and
effective code. Although,, it only works on apple platforms.
1.3.3.1.4. JDev: Oracle

Oracle JDeveloper is a free yet performant EDI allowing to modelize and develop java application for JSE, JEE,
or even JME. It is well integrated with Oracle Database, with in particular a PL/SQL debugger.

1.4. The future of JEE

JEE is a platform in constant evolution. We are currently with the version 1.4 and the version 1.5 is under
development. This version will correct specifications sometimes too complex. It will also add new components
to the platform in order to facilitate always more the development of company's application.
Here is the new architecture planned for this future version :

16
http://www.supinfo.com

The new functionalities are integrated by the JDK 1.5. This is why the version JEE 5 will use hem in order to
have a platform much more accessible and more optimized than the previous one.

17
http://www.supinfo.com

Chapter 2. Servlet JSP

By completing this course, you will :

Understand what is a servlet

Learn how to create basic servlets and more advanced ones

Discover Java Server Pages

Use JavaBeans and Taglibs

Final objective : develop a Java EE Web application

1. Course
1.1. Presentation
Nowadays, web sites are not static anymore, identical for all users. Its now essential to create web site with
personalized content for users which change to meet their needs, allowing them to access information and
interact with it and even with other users.
Servlet and JSP technologies are the solution created by Sun to allow you to create dynamic web content. Not
like scripting language where business tier, presentation tier and data access tier are all mixed together,
Servlet/JSP allow you to create a clear architecture, which motivate developers to separate layers. To faster
development, a web application have to be based on a MVC (Model View Controller) architecture.

Presentation tier are the web page for a client (web browser, applet, rich client application).
Web pages can be static or dynamic. In this last case, they are the result of a code executed on a server.
Development OT this tier can be realized by a J2EE application server with:

Servlets Java

18
Created by XMLmind

XSL-FO Converter.

Java Server Pages

JavaBeans

Tag Libraries

1.2. J2EE Application servers

1.2.1. Presentation
There are many different application server. Each one has its benefits and its drawbacks, and then they are
adapted to different use. Some of them only support Servlet/JSP technologies (Tomcat, Resin, Jetty, WebSphere
Express et Weblogic Express) and some others provide full support of all J2EE specifications (JBoss, Jonas,
Oracle 9iAS, WebSphere Application Server, Weblogic Platform). When you choose a server you take its price
into consideration. Even if most of them are not free, they are many open source server: Tomcat, Jetty, JBoss,
Jonas, Orion. Robust and stable, Tomcat is the standard web server for Servlet/JSP development. Then well use
it for all examples in this lesson.

1.2.2. How it works

1.2.2.1. Requests
1.2.2.1.1. Basic HTTP request
A HTTP request is sent by a web browser (client) to a web server in order to retrieve a web page.
Different steps of the process are:

Browser connect to the server

Browser send its request to the server

The server locates desired web page (the expected response)

The server begins to send web page content (the response) or a error page

Client browser display response

When the transfer is done, the browser close the connection

Remark: When the request is simple (web page for example), no servlet engine is implied in the response
generation.
1.2.2.1.2. Request on a Java server
A web server must contains a module which manages servlet. This module is called a servlet engine or
servlet container . This module can be directly integrated on a server or be independent. Differences between
the two technologies will be explained later. In this example, the servlet engine is independent. In the following
example, well take the case where our servlet engine is independent (in order to understand the task distribution
between the two components).
Different steps of the process are:
19
Created by XMLmind

XSL-FO Converter.

Web browser connect to the server and send its request

The server detect that the request is destinated to a servlet and transfer the request to the servlet engine

Servlet engine checks if the servlet is instanced

If its the first call, servlet engine create the servlet and call its method init()

Then the method service() is called. This method takes two parameters:

The first one represents the client request

The second one represents the response from the server

Servlet engine returns the result to the web server which return it to the client

Web browser display the page and close the connection

For next calls, no servlet instances will be created because the servlet engine always use the same instance.
When there are some synchronous call to the same servlet (a generic servlet), servlet engine create a thread for
each call. Each thread will generate a result for a client.
1.2.2.1.2.1. Servlet engine
1.2.2.1.2.1.1. Autonomous mode
An autonomous servlet engine is an engine which is totally independent, which means that it also contains a
Web server. All requests transit through the servlet engine, whether they are destinated to a static page or a
servlet. This mode is very useful when you only use servlet or when you want to test servlets (local development
for example).
1.2.2.1.2.1.2. Bound to a Web server mode
A servlet engine is usually bound to a Web server. A servlet is actually always associated with JSP or other type
of pages accessible via this Web server. A separated servlet engine provides an optimization of time reply,
because the servlet engine is only call for servlet processing (and not for all requests).
Although all servlet engines can be used in autonomous mode or unbound mode.
1.2.2.1.3. Configuration
1.2.2.1.3.1. Apache Tomcat
To indicates which folders are shared on a server, you have to create a Context Path in the server
configuration file: server.xml .
This file can be used to configure a lot of server parameters, however well only learn the basics to use servlet
and JSP.
It allows you to configure different host name in order to create specific parameters for all of them.
Heres the content of a host tag to configure localhost (default host).

20
Created by XMLmind

XSL-FO Converter.

[CODE:xml]<Host name="localhost" debug="0" appBase="webapps"

unpackWARs="true" autoDeploy="true"
xmlValidation="false" xmlNamespaceAware="false">
<Logger className="org.apache.catalina.logger.FileLogger"
directory="logs" prefix="localhost_log." suffix=".txt"
timestamp="true"/>
<Context path="/localhost" reloadable="true"
docBase="C:\www_root\localhost\"
workDir=" C:\www_root\localhost\work" />
<Context path="/localhost2" reloadable="true"
docBase="C:\www_root\localhost2\"
workDir="C:\www_root\localhost2\work" />
</Host>

Most important tag here is Context . This tag allows you to create different Context Path and then to
share folders of your hard drive.
In previous example we create two contexts which are mapped on local folder.
The first one is: localhost which is mapped to C:\www_root\localhost\ .
The second one is: localhost2 which is mapped to C:\www_root\localhost2\ .
You can access at this two context from a web browser via the following addresses :

http://localhost:port/localhost/ and http://localhost:port/localhost2/.

Port is defined on the server configuration file (by default its 8080).
1.2.2.1.3.2. Sun Application Server
When installed, the server creates a first virtual server. A parameter folder is created with the name http[HostName.User.com]. This structure can change based on parameters that you choose during installation.
Well use supinfo.labosun.com for all our examples. Now well see whats inside configuration folder.
In this folder you have a file : defaultweb.xml which is the default file used for Web application of this virtual
server. There is also a server.xml file which is used to configure an equivalent of context path .
To indicates a folder mapping, you have to add the following tag into the markup <VS ></VS> :
<WEBAPP uri="/pathMap" path="[Lecteur]:[Chemin]"enabled="true"/>

For our example, the file looks like :

<VSCLASS id="vsclass1" objectfile="obj.conf" rootobject="default"
acceptlanguage="off">
<PROPERTY name="docroot" value="d:/Sun/WebServer6.1/docs"/>
<VS id="httpssupinfo.labosun.com" connections="ls1" mime="mime1"
aclids="acl1" urlhosts="supinfo.labosun.com" state="on">
<PROPERTY name="docroot" value="d:/Sun/WebServer6.1/docs"/>
<USERDB id="default"/>
<SEARCH>
<WEBAPP uri="/search"
path="d:/Sun/WebServer6.1/bin/https/webapps/search" enabled="true"/>
</SEARCH>
<WEBAPP uri="/localhost" path="C:\www_root\localhost" enabled="true"/>
<WEBAPP uri="/localhost2" path="C:\www_root\localhost2" enabled="true"/>
</VS>

21
Created by XMLmind

XSL-FO Converter.

</VSCLASS>

Like in the previous example, weve mapped localhost and locahost2 to the two respective folders :
C:\www_root\localhost et C:\www_root\localhost2 .
1.2.2.1.4. Servlet compilation and tests
Well now use the simple servlet ever. It display Hello World when a client call it.
Heres the code :
[CODE:java]import java.io.IOException;
import java.io.PrintWriter;
import javax.servlet.GenericServlet;
import javax.servlet.ServletRequest;
import javax.servlet.ServletResponse;
public final class HelloServlet extends GenericServlet {
public void service (ServletRequest req, ServletResponse res)
throws IOException {
res.setContentType("text/html");
PrintWriter pageWriter = res.getWriter();
pageWriter.println("<html>");
pageWriter.println("<body>");
pageWriter.println("Hello World");
pageWriter.println("</body>");
pageWriter.println("</html>");
}
}

1.2.2.1.4.1. Compilation
If you compile from the command line and you use javac from the JDK, you have to check if your environment
variable: CLASSPATH includes a path to the servlet JAR libraries.
To compile from command line, type:
javac HelloServlet.java

The compiler create a class file which to be placed in specific folder depending on which server you are using.
1.2.2.1.4.2. Deployment
To use our servlet, we now have to deploy it.
1.2.2.1.4.2.1. Apache Tomcat and Sun Application Server
Well now deploy our basic servlet. You must have at least one context path on your server (well use the
context: localhost). Heres the tree structure which you have to use in order to use servlet:
/ => Context root
/.classpath => Define the classpath to access libraries
/WEBINF/classes => Contains packages and servlets
/WEBINF/lib => Can contains external libraries
/WEBINF/web.xml => Application configuration file
/work/ => Working folder for the servlet engine
The file: HelloServlet.class goes in the folder : /WEBINF/classes/.
We now have to indicates to the servlet engine that the context localhost is a web application . This can be
done by editing the file web.xml.
22
Created by XMLmind

XSL-FO Converter.

This file must contains the following code:

[CODE:xml]<?xml version="1.0" encoding="ISO88591"?>
<!DOCTYPE webapp PUBLIC
"//Sun Microsystems, Inc.//DTD Web Application 2.3//EN"
"http://java.sun.com/dtd/webapp_2_3.dtd">
<webapp>
<displayname>Web.xml de base</displayname>
<description>
Xml file used to define a basic Webapp
</description>
<!Servlet definition >
<servlet>
<servletname>HelloServlet</servletname>
<servletclass>HelloServlet</servletclass>
</servlet>
<!Name mapping for servlets >
<servletmapping>
<servletname>HelloServlet</servletname>
<urlpattern>/servlet/Hello</urlpattern>
</servletmapping>
</webapp>

The markup <webapp></ webapp> specify that its a webapp (web application).
Then we indicates, that this webapp contains a servlet named HelloServlet which is HelloServlet(.class).
Finally, well map the path servlet/Hello to this servlet. This allow us to access our servlet via the address:

http://localhost:8080/localhost/servlet/Hello
Remark : You have to be careful at the order of markups. All <servletmapping> markups must be after
<servlet> markups. Which means you cant have a <servlet> markup followed by a <servletmapping> markup
and then another <servlet> markup. Some servers can have problems to parse this file.

1.3. Servlets
1.3.1. Presentation
Servlets are the Java alternative to CG (Common Gateway Interface) programming. A servlet is a Java class,
which is dynamically loaded, allows to extend capacity of a Web server and to reply dynamically to request.
They are executed on a Java Virtual Machine (JVM) on the server which provide a total portability and better
security. Instead of applets, servlets dont force a client to have a JVM because they are executed on a server.
1.3.1.1. Advantages
Servlets offers a lot of advantages instead of CGI. All servlets can be managed by different thread inside the
same process or by different thread inside many processes on multiple servers. The other advantages are that
they are portable from one operating system to another but also from a Web server to another.
Heres a non exhaustive list of all advantages:

Efficacy

Semi compiled

Residents
23
Created by XMLmind

XSL-FO Converter.

Multithread

Cache management

Persistent connections (for database access for example)

Practical

Cookies managements

Session tracing

Simplification of HTTP protocol manipulation

Portable (on all operating system supporting Java)

A lot of free servlet engine

Power

Bidirectional communication on the Web server

Data exchange via URI

Data sharing between servlets

Servlet chaining (servlet included in another one, )
1.3.1.2. Drawbacks
The only drawback is that a servlet is not handy to render graphical user interface. However its possible to
associate a servlet with technologies like Flash, Applet, in order to make them interactive.

1.3.2. Basic implementation

1.3.2.1. Fundamental structure
Weve just seen how to install and configure a Web and a servlet engine. Well now see how to create servlets.
1.3.2.1.1. Servlet interface
To create a servlet, a class must implements the Server (from package javax.servlet) interface directly or
indirectly. This interface defines the following methods:

public void init(ServletConfig cfg)

24
Created by XMLmind

XSL-FO Converter.

public void service(ServletRequest req, ServletResponse res)

public void destroy()

public ServletConfig getServletConfig()

public String getServletInfo()

These methods will be explained later.
Remark : Three first methods represents life cycle of a servlet.
Its not common to use the Servlet interface. Different abstract classes have been created in order to simplify the
task of servlet creation.
1.3.2.1.2. GenericServlet class
This simple class from the package javax.servlet allows you to create a Servlet just by implementing the
method: public void service(ServletRequest req, ServletResponse res).
Heres an example with a Servlet which extends this class:
[CODE:java]import java.io.IOException;
import java.io.PrintWriter;
import javax.servlet.GenericServlet;
import javax.servlet.ServletRequest;
import javax.servlet.ServletResponse;
public final class HelloServlet extends GenericServlet {
public void service(ServletRequest req, ServletResponse res)
throws IOException {
res.setContentType ("text/html");
PrintWriter pageWriter = res.getWriter();
pageWriter.println("<html>");
pageWriter.println("<body>");
pageWriter.println("Coucou");
pageWriter.println("</body>");
pageWriter.println("</html>");
}
}

1.3.2.1.3. HttpServlet class

This class is a bit more evolved and more useful for Web development. It is however included in the package:
javax.servlet.http. Just as for the preceding class, we can simply implement the method: public void
service(ServletRequest req, ServletResponse res) however an alternative opens with us. As protocol HTTP
makes it possible to transmit data of various manners (GET or POST), this class has the methods referring itself
to it. That makes it possible independently to implement the methods of the methods which one wishes to use.

method public doGet(HttpServletRequest req, HttpServletResponse res) is called for a GET request.

method public doPost(HttpServletRequest req, HttpServletResponse res) is called for POST request.
There are other methods for request like Put, Delete, Options. These methods have a structure like public
doXxx(HttpServletRequest req, HttpServletResponse res) where Xxx = (Put, Delete, Options )
Remark : If you dont need to know which type of request it is, you can just use GenericServlet

25
Created by XMLmind

XSL-FO Converter.

You may have noticed

ServletResponse/ServletRequest.

that

we

use

HTTP[ServletResponse/ServletRequest]

instead

of

1.3.2.2. Life cycle

Weve seen that servlet engine only use one instance of a servlet , and a different thread is used for each request
of a client. Threads call doGet(), doPost() or service() (depending on the case). Well now see how servlets are
created, initialized, used and destroyed. This steps are called the life cycle of a servlet.
1.3.2.2.1. init() method
init() method is only called when a servlet is called the first time (either via an client call either when the server
start, depending on your server configuration). Therefore, this method is used to initialize and configure a
servlet.
init() has two prototypes :
the first one takes no argument, its used when the servlet doesnt need to read parameters specific to a server.
Definition of this method is:
[CODE:java]public void init() throws ServletException {
// Initialization Code
}

the second version is used when a servlet need to read configuration parameters. This parameters are used to
adapt a servlet to a specific server. Parameters can be used to retrieve database connection, a password file or
other things still.
Definition of this method is:
[CODE:java]public void init(ServletConfig cfg) throws ServletException {
super.init(cfg);
// Initialisation Code
}

The only argument of type ServletConfig provides access to different configuration parameters thanks to the
method getInitParameter(). You have to specify the parameter name as argument and the method returns its
value.
You must not forget to call super.init() if you override this method. super.init() from superclass store the object
ServletConfig at a specific place in the servlet which is used thereafter.
1.3.2.2.2. service method
This method is called for each request. If you dont use HttpServlet you have to override this method to
implement your servlet. But if you use a HttpServlet, this method checks the type of request and then call the
good method (doGet(), doPost(), doTrace()). If you want a servlet which process GET and POST with the same
manner, you may want to override service() method but its not the good solution! It can actually lead to conflict
if your servlet evolves. You better have to create a new method which will be called by doPost() via doGet().
Thanks to this method, you can override doPut(), doTrace(), doOptions() directly or in subclasses.
Definition of this method :
public void service(ServletRequest req, ServletResponse res) throws IOException

1.3.2.2.3. destroy method

A server can ask for a servlet destruction, either by explicit call from administrator (server reboot for example),
either because a servlet is inactive for a too long time. Server call the method destroy() in which all destruction
operations take place. You can put here database connection closing, threads interruption, file closing,
26
Created by XMLmind

XSL-FO Converter.

Remark : a Web server can be shutdown because of a bug or another technical problem (someone can plug out
the server!), thats why you cant be sure that this method will be called. So make sure to save your data
periodically.
Definition of this method :
[CODE:java]public void destroy();

1.3.2.3. Read request

You can read data from the HttpServletRequest object. This object is available in methods: service(), doGet(),
doPost()
From the request, you can retrieve all parameters passed to the server by the client browser.
There are four types of methods:

getParameter(String name) method: returns a parameter value from its name

getParameterValues(String name) method: returns a list of values of one parameter from its name (its used
of example for a multi choice list)

getParameterMap() method: returns a Map which contains all parameters name and their associated value

getParameterNames() method: returns a list of all parameters name

This object can also be used to get information about client:

getRemoteAddr() method: returns client IP address

getRemoteHost() method: returns complete host name of client

You can also get information about server :

getServerName() method: returns the server name

getServerPort() method: returns port number used by server

You can access to server environment variable (like CGI script).

Table 2.1. Heres a board which summarize all variables :

CGI environment variable
SERVER_NAME

HTTP Servlet methods

getServerName()

SERVER_PROTOCOLE

getProtocole()

SERVER_PORT

getServerPort()

REQUEST_METHOD

getMethod()

PATH_INFO

getPathInfo()

PATH_TRANSLATED
SCRIPT_NAME
DOCUMENT_ROOT

getPathTranslated()
getServletPath()
getServletContext().getRealPath("/")
27

Created by XMLmind

XSL-FO Converter.

QUERY_STRING

getQueryString()

REMOTE_HOST

getRemoteHost()

REMOTE_ADDR

getRemoteAddr()

AUTH_TYPE

getAuthType()

REMOTE_USER

getRemoteUser()

CONTENT_TYPE

getContentType()

HTTP_ACCEPT

getHeader("Accept")

HTTP_USER_AGENT

getHeader("UserAgent")

HTTP_REFERER

getHeader("Referer")

1.3.2.4. Write the response

You can write response thanks to HttpServletResponse object. This object is available in methods: service(),
doGet(), doPost()
(We will note that the instance of this object will be : res ).
1.3.2.4.1. Content type
The first thing is to specify the type of data your servlet will returns. A servlet can actually returns HTML,
picture, file,
You can do that with the method setContentType(). You specify the type of data in argument. Lets see some
examples of use:
[CODE:java]res.setContentType("text/html"); // page HTML
res.setContentType("text/xml"); // page XML
res.setContentType("image/jpeg"); // image JPEG

Its possible to returns all type of content by specifying the good MIME type.
1.3.2.4.2. Writing
The data writing of the text type in the answer is done by the object : PrintWriter which we can recover by the
method : getWriter().
[CODE:java]PrintWriter out = res.getWriter();
out.println("<HTML>");
out.println("<BODY>");
out.println("Bonjour le monde");
out.println("</BODY>");
out.println("</HTML>");
out.close();

If you want to returns binary data, its preferable to use the ServletOutputStream object returned by method
getOutputStream().
Heres an example to return a picture from a local picture :
[CODE:java]res.setContentType("image/gif");
ServletOutputStream out = res.getOutputStream();
FileInputStream fis = new FileInputStream (new File ("./logo.gif"));
while (fis.available() > 0) {
fis.read (buf, 0, buf.length);
out.write(buf);
out.flush();
}
fis.close();
out.close();

1.3.3. Advanced implementation

28
Created by XMLmind

XSL-FO Converter.

Weve seen in previous part, the basics for servlet development. Now well learn advanced concepts.
1.3.3.1. Cookies
1.3.3.1.1. Principle
Cookies are text information send by server, and send back by client without any modifications, at each later
visit. This technique is very useful to keep trace of visitors (to know if they already come to the site or not).
1.3.3.1.2. Advantages
Cookies allow to keep a link between the client and its session on a server thanks to an session login for
example.
Cookies are very simple to use and allow to manage a lot of interesting functionalities for client.
1.3.3.1.3. Drawbacks
Cookies are sent at each request, this can increase network traffic if you make an expensive use of them.
Usage of cookies is not recommended for security. Value of each parameters actually transit with no encrypting
and can be retrieved by a hacker.
Cookies are not always accepted by clients, this can be a problem when no other s means of session tracking
exist.
1.3.3.1.4. Cookies creation
Servlets use their own cookies API. Cookies are represented by the class Cookie. Heres a description of this
class:

Constructor Cookie(String name, String value) : constructor which takes in parameter the name of the cookie
and his assigned value.
Once a cookie is instanced, you can specify certain properties (with setters ) or retrieve information from a
cookie (with getters ) :

Method String getComment() : returns a comment

Method setComment(String) : set a comment

Method String getDomain() : returns the domain

Method setDomain(String domainPattern) : set a domain

Method int getMaxAge() : returns the duration in second before the cookies is out of date.

Method setMaxAge(int lifetime) : set the duration in second before the cookies is out of date. Value 0
indicates that the cookies must be destroy.

Method String getName() : returns the name of the cookie

Method setName(String name) : set the name of the cookie

29
Created by XMLmind

XSL-FO Converter.

Method String getPath() : return the path to the cookie

Method setPath(String path) : set the path to the cookie (allow to specify that the cookie is only used for
certain part of the web site)

Method boolean getSecure() : returns true if cookie is transmit only for secure connection

Method setSecure(boolean secureFlag) : set if the cookie must be transmit only for secure connection

Method String getValue() : returns the cookie value

Method setValue(String cookieValue) : set or modify the cookie value

Once you have set cookie parameters, you have to add it in the http header: SetCookie. To do that, you call the
method addCookie() in response object of your servlet.
Heres a cookie creation example :
[CODE:java]Cookie cookieAlreadyCome = new Cookie("AlreadyCome", "1");
cookieDejaVenu.setMaxAge(60 * 60 * 24 * 365); // 1 year
res.addCookie(cookieAlreadyCome);

1.3.3.1.5. Cookies reading

Received cookies are stored in an array which can be retrieved from the HttpServletResquest object via the
method getCookies(). This method returns an array of which the length represents the number of cookie sent (if
the length is equal to null, no cookie was sent). To obtain a particular cookie from its name you must traverse
the array by calling the method getName() on each element till you arrive to your wanted cookie. Once you have
it, you just have to call method getValue() to have its content.
Heres a cookie creation example :
[CODE:java]public void doGet(HttpServletRequest req, HttpServletResponse res)
throws ServletException, IOException {
// Initialisation
PrintWriter out = res.getWriter();
res.setContentType ("text/html");
Cookie[] tabCookies = req.getCookies();
Cookie cookie;
// Loop to display all cookies
for (int i = 0; i < tabCookies.length; i++) {
cookie = tabCookies [i];
out.println("Cookie : " + i + "<br>" +
" Name : " + cookie.getName() + "<br>" +
" Value : " + cookie.getValue() + "<br>");
}
out.flush();
out.close();
}

If you want to remove a cookie, you just have to set an age equals to 0 :
[CODE:java]c1.setMaxAge(0);

1.3.3.2. Sessions

30
Created by XMLmind

XSL-FO Converter.

1.3.3.2.1. Presentation
Weve seen how to keep information about a clients with cookies. However this mechanism is very limited,
thats why session system has been created.
When you talk about session, activity is on the server and not on the client (instead of cookies). Session are vert
limited in web application and more specifically in ebusiness web site Sessions allow to store data for each
client on the server.
1.3.3.2.2. Advantages
Sessions are executed on the server which is more secure than cookies. Only an identifier is used to link client
and server because data are stored on the server.
Sessions are simple to use thanks to the servlet session API.
1.3.3.2.3. Drawbacks
A link must be established between the client and the server (session login) either with a cookie, either inside
the URL.
1.3.3.2.4. session API
Servlet provides an API to manage sessions. Usage of sessions inside servlets is simple, you just have to retrieve
the session object bound to the current request. If no session object exists, you can create it automatically.
Class which represents session is : HttpSession.
Heres a description of its methods :

Method Object getValue() or Objet getAttribute(String name) : returns value of an existing session attribute
Remark : method getValue() is replaced by getAttribute() since servlets API version 2.2

Method putValue(String name, Object value) or setAttribute(String name, Object value) : bind an object to a
name
Remark : method putValue() is replaced by setAttribute() since servlets API version 2.2

Method removeValue(String name) or removeAttribute(String name) : remove an attribute from its name
Remark : method removeValue() is replaced by removeAttribute() since servlets API version 2.2

Method String[] getValueNames() or Enumeration getAttributeNames() : returns all attributes session name
Remark : method getValueNames() is replaced by getAttributeNames() since servlets API version 2.2

Method String getId() : returns a string containing the unique identifier assigned to this session. (this can be
used has a key)

Method boolean isNew() : returns true if the session has just been created or false if it already exists

Method invalidate() : invalidate the whole session and remove all objects bound to the session
1.3.3.2.5. Session management
31
Created by XMLmind

XSL-FO Converter.

First you must retrieve the session object with the method getSession() from the object HttpServletRequest.
[CODE:java]HttpSession session = req.getSession(true);

To automatically create the session (if it doesnt exists), you pass true to the method getSession().
Remark : If you pass false as parameter, the method getSession() will returns null.
With the session object, you can now add object (method setAttribute()), remove (method removeAttribute()) or
read (method getAttribute()).
1.3.3.3. SingleThreadModel interface
Weve seen that only a servlet instance is used by multiple thread (on thread by client). This can be a problem
when you want for example, that a connection to a database doesnt be shared between threads, a file
modification
The first solution is to add synchronized to service() method of servlet.
However there is a better way to do that. SingleThreadModel interface indicates to the servlet engine that you
want that a servlet can only be called by a Thread at the same time.
This interface doesnt defines any methods, its only used as a flag for the servlet engine.
Finally, this interface is useful because it can be used to secure concurrent access, however its not always the
best solution. In the case of database connection, the best things to do is to create a multi thread connection
pool. Indeed, the database can only accept a restricted number of connections (which can be lower than the
number of simultaneous clients).
1.3.3.4. ServletContext usage
A servlet associated to a servlet (ServletContext) is the environment in which the servlet is executed. Object
ServletContext is created by the servlet container and allow to access information about the environment.
The object is shared between all servlets container. This can be used to share information between servlets.
1.3.3.4.1. ServletContext description
Heres a description of all main methods of ServletContext :

Object getAttribute(String name) : returns attribute associated to the name

void setAttribute(String name, Object o) : bind an object to this attribute

removeAttribute(String name) : remove an attribute from the context

Enumeration getAttributeNames() : returns the list of all context attributes

int getMajorVersion() : returns the major version of the Java Servlet API that this servlet container supports.

int getMinorVersion() : returns the minor version of the Servlet API that this servlet container supports
1.3.3.4.2. Usage of object ServletContext
Method getServletContext() is used to retrieve the ServletContext object.

32
Created by XMLmind

XSL-FO Converter.

1.4. Java Server Pages

1.4.1. Presentation
Weve seen how to use servlets, theirs advantages and theirs drawbacks.
Servlets, Java classes which extends javax.servlet.http.HttpServlet, are mainly focused on Java code. JSP page
looks like HTML code.
Indeed, a JSP page looks like HTML code in which you add Java code inside markup (<% %>). However JSP
pages use functionalities from Servlet like the request (HttpSerlvetRequest) and object
javax.servlet.ServletOutputStream.
Instead of servlets, JSP provides a clear separation between presentation (HTML) and data processing (Java
code).
Its very easy to develop with JSP, indeed its not necessary to know the underlying API.
And instead of servlets, webmasters can modify HTML code without touching Java code.
Like weve just seen, a JSP page can contains Java Code and HTML.
HTML is used to create the static structure of the page, and JSP code compose dynamic elements of the page.
This elements can be of three types:

Scripting elements

Directives

Action elements

1.4.2. Scripting base elements

There are four types of scripting elements :

Scriplets

Comments

Declarations

Expressions
Scriplets, <% my java code%>, are Java code inside a HTML page. This characters allow to specify which type
of language to use to process encapsulated code. Variables declared inside this block can only be accessed from
the current block.
Example :
[CODE:java]<%String s = "Hello!"; %>

Comments, <% my comments %>, like the name said are used to create comments inside your code. You can
also use HTML comments code, but JSP comments cant be seen by clients.
Declarations, <%! %>, are used to declare methods and instance variable which can be used in all the current
33
Created by XMLmind

XSL-FO Converter.

page. They can be used anywhere in the code.

Expressions, <%= %>, are used to directly write on the output flow (it used the implicit out object). No
semicolon is necessary at the end of the expression.
Example:
Message :<%= s %>

Displays:
Hello!

1.4.3. Implicits objects in JSP

Java code inside scriplets have access to a list of elements called implicit objects. This objects allow to interact
with the environment of the servlet execution.

Request element references javax.servlet.http.HttpServletRequest associated to the request. It can be used to

access POST and GET parameters, HTTP header or cookies.
Response element references javax.servlet.http.HttpServletResponse send with the response to the client. Its
not often used in preference of out object.

out element references javax.servlet.jsp.JspWriter is the output flow send to the client

config element references javax.servlet.ServletConfig object of the page.

pageContext element references of the page context, represented by the object javax.servlet.jsp.PageContext.
It provides access to different attributes of the page.
application element references javax.servlet.ServletContext. Its used to store global data. Difference with
instance field are that these data will be shared between all the servlets of the servlet engine.

page element references an instance of the page itself, like the keyword this inside a Java class.

session element references javax.servlet.http.HttpSession.

exception element references java.lang.Throwable. Its used to catch exceptions.

1.4.4. Example of use with a form

We want to create a simple form which allow to enter data inside text field then to display them on another
page.

34
Created by XMLmind

XSL-FO Converter.

Heres the code of the two JSP page Formulaire.html et ResultForm.jsp :

1.4.4.1. FirstJSP.html :
This page only contains HTML code:
<html>
<head>
<title>My Form</title>
</head>
<body>
<h1>My Form</h1>
<form action="ResultForm.jsp" method="post" />
<table>
<tbody>
<tr>
<td>Your last name <input type="text" name="lastName" /></td>
</tr>
<tr>
<td>Your first name<input type="text" name="firstName" /></td>
</tr>
<tr>
<td><input type="submit" name="submit" value="Validate " />
<input type="reset" name="reset" value="Reset" /></td>
</tr>
</tbody>
</table>
</form>
</body>
</html>

1.4.4.2. Result.jsp :
Heres interesting part, we can see how to retrieve data sent by a form.
Hence, thanks to method getParameter(String s) from request variable, we can retrieve variables values passed
as parameters.

35
Created by XMLmind

XSL-FO Converter.

[CODE:html]<html>
<head>
<title>Result of my form </title>
</head>
<body>
<h3>My information</h3>
<%
if (request.getParameter("submit") != null) {
if (request.getParameter("submit").equals("Validate")) {
%>
<table>
<tbody>
<tr>
<td><b>Last name: </b>
<% out.println(request.getParameter("lastName"));%></td>
</tr>
<tr>
<td><b>First name: </b>
<% out.println(request.getParameter("firstName"));%></td>
</tr>
<tr>
<td><form action="Formulaire.html">
<input type="submit" value="Back"></form></td>
</tr>
</tbody>
</table>
<%
} else {
%>
<jsp:forward page="Formulaire.html"/>
<%
}
}
%>
</body>
</html>

1.4.5. Actions elements : Scripts inclusions and jsp: markup

An action elements is a set of markup which can be integrated inside a JSP page.
Heres a description of standard JSP actions.

jsp:useBean
useBean markup is used to separate presentation tier from business login tier. Indeed, you will Java classes
which will allow you to create JavaBeans. These JavaBeans will retrieve data from the JSP page, process it and
return it to the JSP page in order to display it.
A JavaBean is an object represented by a Java class which must conforms to the following rules:

The class must have a public constructor which takes no argument

No attributes can be public

Each attribute (property) must have a public getter and setter

For boolean attributes, you have to use isXXX() instead of getXXX().

useBean markup is used to instantiate a JavaBean class which can be used inside the JSP page.
36
Created by XMLmind

XSL-FO Converter.

MyBean.java :
[CODE:java]package monpackage;
public class MyBean {
private int nb;
private boolean empty;
public MyBean() {
}
public int increm(int i) {
return nb + 10;
}
public int getNb() {
return nb;
}
public void setNb(int nb) {
this.nb = nb;
}
public boolean isNb() {
return empty;
}
}

UTestBean.jsp :
<jsp:useBean id="testB" class="mypackage.MyBean"/>

Here an object of type mypackage.MyBean has been created and affected to testB variable.
Remark : You have to place your classes inside packages.
1.4.5.1. Markups jsp:setProperty et getProperty
There markups are used to get or set a property of a JavaBean instance.
There are 3 ways to do :
A star ( * ) for property allow to set automatically all attributes of a JavaBean based on the values retrieved
from a form.
<jsp:setProperty name="testB" property="*"/>

You can also specify the name of the property you want to set. The value of the attribute will be the value of the
parameter specify with param of the form.
<jsp:setProperty name="testB" property="propName" param="paramName"/>

You can directly indicates the name of the property and its value.
<jsp:setProperty name="testB" property="propName" value="Jean"/>

Markup <jsp :setProperty> is used to retrieve assigned values.

<jsp:getProperty name="testB" property="propName"/>

You can now use the JavaBean instance inside scriplets markup and use its method. Hence, instead of using
<jsp:getProperty> markup, you can use this syntax :
<%= testB.getNomProp() %>

1.4.5.2. Markups jsp:include and param

Markup include is used to integrate static or dynamic resources inside JSP page. It allows to include file at the
37
Created by XMLmind

XSL-FO Converter.

request time.

Its also possible to include file at the page translation time, just before the compilation (4.4.2).
If you want to pass parameters to the resource, you have to add the following lines :
[CODE:java]<jsp:include page="Relative path to the file " flush="true">
<jsp:param name="PARAM1" value="VALUE1"/>
<jsp:param name="PARAM2" value="VALUE2"/>(...)
</jsp:include>

1.4.5.2.1. jsp:forward
This markup is used to pass the request control to another dynamic or static request. If you want to pass
parameters to the resource, you use the same syntax that before:
Example :
[CODE:java]<jsp:forward page="Redirect.jsp">
<jsp:param name="redir" value="ca marche"/>
</jsp:forward>

Hence, inside the page Redirect.jsp you can get the value of redir thanks to the variable request
(request.getParameter("redir")).
1.4.5.2.2. Markups jsp:plugin et jsp:fallback
The plugin markup is used to generate HTML markups <object> or <embed> which indicates to the client
browser the loading of the corresponding markup.
The fallback markup is used to display a message if the client browser doesnt have support for applet.

1.4.6. Directives
Directives are messages send to the JSP container to indicates how to translate a JSP into a servlet. These
directives can be easily distinguished thanks to characters <%@ %>.
There are 3 types of directives, pages directives, inclusions directives and personalized directives.
1.4.6.1. Pages directives

Table 2.2. These directives have a form like : <%@page... %>. They can take the
following parameters :
Parameter

Parameter role

Values

language

Indicates the used language

java

38
Created by XMLmind

XSL-FO Converter.

info

Text inside this parameter contains Depend on the JSP container

information about this page and can
be retrieve thanks to the method
servlet.getServletInfo()

contentType

Set
the
MIME(Multipurpose text/html;charset=ISO8859-1
Internet Mail Extension) type of the
page and the character encoding

extends

Set the parent class of the generated class

servlet. Use this parameter with
precaution.

import

Set the package or classes None

inclusions. Its like a Java import.
You can import many classes
separated with comma.
<%@page import="java.util.*"%>
<%@page
import="java.util.*,java.io.*"%>

buffer

By default, JSP content is 8192b

bufferized in order to improve
performance.

autoFlush

Indicates if the buffer must be true / false

empty when its full.

session

Indicates if current page can access true / false

data store inside the session.

isThreadSafe

Indicates if the SingleThreadModel true / false

(see 3.3.3) model is used or not.

errorPage

All non catched exceptions will be url

redirected to the following page.

isErrorPage

Indicates to the JSP container that false

the generated servlet is an error
page.

* bold values are default values

1.4.6.2. Inclusions directives
These directives are used to include the content of a file inside a JSP page. To make an inclusion, you use the
syntax: <%@ include file="RelativeUrl" %>. This type of inclusion is realized before the JSP compilation into a
servlet.

1.4.6.3. Tag libs directives

These directives define the pat hand the namespace prefix of personalized markups (tags) which can be used
inside the page.
<%@ taglib uri="tagLibraryURL" prefix="tagPrefix" %>

They allow to create new tags like :

39
Created by XMLmind

XSL-FO Converter.

<tagPrefix:TagName attribute1="value" ... %> ... </tagPrefix:TagName>

These directives will be explained later in the next chapter ( 4.7).

1.4.6.4. Personalized markups (Taglib)
1.4.6.4.1. Presentation
Weve just seen that a JSP page is a page which contains Java and HTML code. Even if Java is a powerful
language, there can be time when you write the same code again and again inside the same JSP page (loop, test,
session object retrieving).
In order to minimize this waisting of time, a mechanism to execute repetitive tasks with the help of markup has
been created. This mechanism is called taglib.
Taglib allow to put Java code inside a Java file instead of a JSP page. These taglib aim to simplify the creation
Web site with JSP. This is very important when developing and maintaining code to isolate the work of designer
from developers.
With taglib you will have a TLD (Tag Library Descriptor) file which will describe the attributes of the markup
and the associated Java file. The java file describe all the markup actions.
1.4.6.4.2. JSP Custom Tag API
The personalized markup must be link to a Tag handler which will define its actions. A tag handler is a Java
class called each time that the JSP container want to translate the markup.
This class must implement an interface inside the package javax.servlet.jsp.tagext or extends a class of this
package.
The 3 most important interfaces are Tag, IterationTag et BodyTag.
1.4.6.4.2.1. Tag interface
[CODE:java]public interface Tag {
int doEndTag();
int doStartTag();
Tag getParent();
void release();
void setPageContext(PageContext pc);
void setParent(Tag t);
}

Tag interface is the interface to implement to create a tag handler. Different methods of this tag will allow
initialization, evaluation and closing of the markup. Different call of these methods obey to the following life
cycle:
1.4.6.4.2.2. Tag handler life cycle
Inside the same JSP page there is only one instance of the same tag. This means that is you call the same tags
several time, the same instance will be used.
Life cycle of a tag handler is controlled by a JSP container :

First, the JSP container get an tag handler instance from a instance pool, or create a new one.

It then call the setPageContext() method with the argument PageContext which represents the JSP which
contains the tag.

The container call the method setParent() with the parent tag (closest enclosing tag handler) as parameter or
null.
40
Created by XMLmind

XSL-FO Converter.

JSP container then define all attributes of the custom tag. Attributes are processed like JavaBeans properties
(getters, setters).

The method doStartTag() is called. This method can returns the following values:

Tag.SKIP_BODY : There is no evaluation of the body of the tag. It means that the content between the
open markup and the closing markup wont be displayed.

Tag.EVAL_BODY_INCLUDE : the container process the body of the tag like JSP and the body will be
displayed inside the page (the body can contains JSP, HTML and even others tags).

Whatever the returns value is, the container call the method doEndTag(). This method can returns :

Tag.SKIP_PAGE : No processing for the rest of the JSP page. The content of the page after the tag will not
be processing.

Tag.EVAL_PAGE : The rest of the page is processed.

The last method called dis release(). Its inside this method that you can write memory liberation code
(database connection).
This method is not called after each call of the method doEndTag() ! Indeed, the container only called this
method when the references of the tag handler will be deleted.

Finally, the JSP container returns an instance of the tag handler inside the pool.

For now methods of our Tag Handler are very limited. We can for example, evaluate the content of our tag (its
41
Created by XMLmind

XSL-FO Converter.

content) only one time, and if we cant display data, we cant retrieve the content of the tag and modify it. We
can only retrieve data passed as parameters of the tag.
Example of tag which implements the Tag interface :
Well create a Tag handler which represents a markup colorMessage. This tag will display Hello World Tag !
With a color specified as parameter.
[CODE:java]public class MyTag implements Tag{
private PageContext pageContext;
private Tag parent;
private String color;
public void setColor(String color) {
this.color = color;
}
public void setPageContext(PageContext arg0) {
pageContext = arg0;
}
public void setParent(Tag arg0) {
parent = arg0;
}
public Tag getParent() {
return parent;
}
public int doStartTag() throws JspException
{
JspWriter out = pageContext.getOut();
try {
out.println("<font color="+color+">Hello World Tag ! </font>");
}catch(IOException e){
}
return SKIP_BODY;
}
public int doEndTag() throws JspException{
return SKIP_PAGE;
}
public void release() {
}
}

1.4.6.4.2.3. IterationTag interface

IterationTag interface which extends Tag interface, allow to evaluate the content of tag many times.
[CODE:java]public interface IterationTag extends Tag {
int doAfterBody();
}

Method doAfterBody() is called after doStartTag() method and can returns Tag.SKIP_BODY or
Tag.EVAL_BODYAGAIN.
This allow to reevaluate the body of the tag many times. If doAfterBody() returns Tag.SKIP_BODY the method
doEndTag() is directly called, if Tag.EVAL_BODYAGAIN is returned the method doAfterBody() is called.

42
Created by XMLmind

XSL-FO Converter.

1.4.6.4.2.4. BodyTag interface

BodyTag, interface which extends IterationTag, interface allow to retrieve the content of the body.
[CODE:java]public interface BodyTag extends IterationTag {
void doInitBody();
void setBodyContent(BodyContent b);
}

This interface define two extra methods :

Method setBodyContent(BodyContent b), is called after doStartTag() and initialize the body content, and then to
retrieve the content of the tag and to modify it.
Method doInitBody() will check if the body content object has been correctly initialized and will throw a
JspTagException if the check fails.

43
Created by XMLmind

XSL-FO Converter.

1.4.6.4.2.5. Creation of a custom tag

Custom tags can be separated in two categories:
1.4.6.4.2.5.1. Tag without body content evaluation
Its tags like line break (markup <Br /> in HTML) which dont need content modification.
To create a tag which dont need body content modification, we use a class which extends TagSupport. This
class implements IterationTag :
[CODE:java]public class TagSupport implements IterationTag {
int doEndTag();
int doStartTag();
static Tag findAncestorWithClass(Tag from, Class classe);
String getId();
Tag getParent();
Object getValue(String k);
Enumeration getValues();
void removeValue(String k);
void setId(String id);
void setPageContext(PageContext pageContext);
void setParent(Tag t);
void setValue(String k, Object o);
}

Advantages of extending TagSupport instead of implementing IterationTag is to only override used methods.
Implementation of the class TagSupport is :

Method doStartTag() returns SKIP_BODY

44
Created by XMLmind

XSL-FO Converter.

Method doAfterBody() returns SKIP_BODY

Method doEndTag() returns EVAL_PAGE
Example of a tag without body content evaluation :
In this example, well see how to create a tag which display the message "Tag without body" with a defined
color font. To do this, well use a tag handler (TagWithoutBody) which extends TagSupport.
TagWithoutBody.java
[CODE:java]public class TagWithoutBody extends TagSupport {
private String color;
public void setColor(String color) {
this. color = color;
}
public int doEndTag() throws JspException{
JspWriter out = pageContext.getOut();
try {
out.println("<font color="+couleur+">Tag without body</font>");
} catch (Exception e) {
System.out.println(e.getLocalizedMessage());
}
return EVAL_PAGE;
}
}

The class TagWithoutBody has a color attribute which defines the font color of the message.
1.4.6.4.2.5.2. Tags with body evaluation
Its tags like bold or color markups (markup <b> </b> or <font color=> </font> in HTML) which modify its
content.
For this type of tags, well create a class which extends BodyTagSupport. This class extends and implements
BodyTag :
[CODE:java]public class BodyTagSupport implements BodyTag {
int doAfterBody();
int doEndTag();
void doInitBody();
int doStartTag();
BodyContent getBodyContent();
JspWriter getPreviousOut();
void release();
void setBodyContent(BodyContent b);
}

Hence to retrieve the body of the tag, you will retrieve from the class BodyTagSupport
BodyContent class:
This class extends javax.servlet.jsp.JspWriter. It represents the content of a custom tag (when it exists).
BodyContent of a custom tag can be retrieved with the method getBodyContent() of the BodyTagSupport class.
String content = bodyContent.getString() ;
Finally, the writing of the JSP page is made with the help of a PrintWriter inside the BodyContent.
JspWriter out = bodyContent.getEnclosingWriter() ;
Implementation of the class BodyTagSupport is :

Method doStartTag() returns EVAL_BODY_BUFFERED

45
Created by XMLmind

XSL-FO Converter.

Method doAfterBody() returns SKIP_BODY

Method doEndTag() returns EVAL_PAGE

Example of tag with a body content :
The class UpperTag.java that weve used with the example of the TLD file correspond to a tag with a body
content. It actually need to have the body content to put it in uppercase.
The class UpperTag will then extends BodyTagSupport :
UpperTag.java
[CODE:java]public class UpperTag extends BodyTagSupport{
private int nb=0;
private String color="";
public int doAfterBody() throws JspException{
nb;
if(nb>=0){
BodyContent bc = getBodyContent();
String s = bc.getString().toUpperCase();
JspWriter out = getBodyContent().getEnclosingWriter();
try{
out.println("<br><font color="+ color +">Your content: " + s+"</font>");
bc.clearBody();
return EVAL_BODY_BUFFERED; // returns of
//EVAL_BODY_BUFFERED will cause the method doAfterBody() to be called
}catch(IOException io){
}
}
return SKIP_BODY;
}
/**
* @param color The color to set.
*/
public void setColor(String color){
this.color = color;
}
/**
* @param nb The nb to set.
*/
public void setNb(String nb){
this.nb = Integer.parseInt(nb);
}
}

1.4.6.4.2.6. Role and format of the TLD File (Tag Librairy Descriptor)
The TLD file is a XML file which describes personalized markups.

Table 2.3. Syntax of TLD file :

Markup
<taglib>(mandatory)

Markup role
Its the root element of the file

<tlibversion> (mandatory)

Tag library version

<jspversion> (mandatory)

JSP specification version whose the library depends

<shortname>

Short name of the library

<info>

Information for documentation

<uri>

link to a source which identify the library in a single

46
Created by XMLmind

XSL-FO Converter.

manner
<tag>

Each tag markup define a personalized markup

1.4.6.4.2.6.1. Tag element

Table 2.4. Tag markup will define characteristic of a personalized tag. This markup
have the following elements :
Element name

Element role

name (mandatory)

Tag indentifier

tagclass (mandatory)

Tag Java class

teiclass

Information Java class name about the TagClass

Type of content

bodycontent

Empty : no content for the tag

JSP : the content can contains JSP

tagdependent : the body must be processed

info

description, information

Attribute

name (obligatory) : identifier

required : (true,false)

rtextprvalue:value of this attribute can be known at

the request time (true,false)
Attribute
1.4.6.4.2.6.2. Example of TLD file
In this example, weve the Java class UpperTag.java which allow to put a text into uppercase. We now want to
define a corresponding markup inside our tld file (Desc.tld).
File Desc.tld :
[CODE:xml]<?xml version="1.0" encoding="iso88591" standalone="yes"?>
<taglib>
<tlibversion>1.0</tlibversion>
<jspversion>1.2</jspversion>
<shortname>myTags</shortname>
<description>My first tag </description>
<tag>
<name>Uppercase</name>
<tagclass>UpperTag</tagclass>
<attribute>
<name>nb</name>
<rtexprvalue>true</rtexprvalue>
<required>false</required>
</attribute>

47
Created by XMLmind

XSL-FO Converter.

<attribute>
<name>color</name>
<rtexprvalue>true</rtexprvalue>
<required>true</required>
</attribute>
</tag>
</taglib>

1.4.6.4.2.7. Role of deployment descriptor

The deployment descriptor is a XML file which allow the JSP container to access a particular TLD file(by
specifying the file name and its location ). Its possible to avoid the usage o the deployment descriptor by
referencing directly a TLD file inside a JSP page. However this reduce the flexibility, because during changes,
you have to modify all JSP files using custom tags.
Example a web.xml file :
[CODE:xml]<webapp id="WebApp_ID">
<displayname>Td_Corrected</displayname>
<welcomefilelist>
<welcomefile>index.html</welcomefile>
<welcomefile>index.htm</welcomefile>
<welcomefile>index.jsp</welcomefile>
</welcomefilelist>
<taglib>
<tagliburi>/MyTags</tagliburi>
<tagliblocation>/WEBINF/tlds/Desc.tld</tagliblocation>
</taglib>
</webapp>

1.4.6.4.2.8. Syntax and usage of a custom tag

If you want to use a personalized markup inside your JSP, you first have to declare it, like this :
<%@ taglib uri="MyTags" prefix="mylibs" %>

The attribute uri specify the path to the TLD file. The attribute prefix defines a prefix which is used to prefix the
name of the tags (like a namespace).
To use a tag, you just use one of the following syntax :

<TagPrefix:TagName attribute1="value"/>

< TagPrefix: TagName >body</ TagPrefix: TagName >

Example :
<mylibs:Uppercase nb="12" color="red">Hello World</mylibs:Uppercase>

1.4.6.4.3. Richness of CustomTag

CustomTag have a great richness in JSP programming. The real interest if to reuse predefined taglibs, very
complete and perfectly reusable taglibs.
JSTL (JavaServer Pages Standard Tag Library) are references libraries. They contains a set of custom tags
which allow for example to iterate, to execute SQL request, to use XML document.
It also exists, many other powerful libraries:
Here are some links of the most known:

48
Created by XMLmind

XSL-FO Converter.

http://jakarta.apache.org/taglibs/
http://displaytag.sourceforge.net/install.html
http://struts.apache.org/userGuide/
1.4.6.4.3.1. JavaServer Pages Server TagLib
Well now see some functionalities of the JSTL.
To use JSTL, you must have a servlet engine which have support for servlet API 2.3 and JSP API 1.2
All the files necessary to their use are in the pack Java Web Services Developer Pack v1.4.
Once you have the jstl folder, you must place:

TLD (in /jstl/tld/) inside the folder /WEBINF/tlds/

Files jstl.jar and standard.jar (in /jstl/tld/) inside the folder /WEBINF/lib
JSTL are made up of 4 different libraries :

Core library represents basic functions and is associated to the TLD c.tld.

XML library represents XML processing and is associated to the TLD x.tld

I18n library represents internationalization functions and is associated to the TLD fmt.tld

Database library represents SQL processing and is associated to the TLD sql.tld
1.4.6.4.3.1.1. Core library
This library contains tags to create conditional structure, url management or iteration.

buoy of expression language : out, set, remove et catch.

out markup makes it possible to write to the output flow the content of value attribute.
There are also :

default value if value is null

escapeXml (boolean) set if special characters (<, >, &, ", ) must be translated into HTML equivalent (&lt;,
&gt;, &amp;, &quot;,) true by default
we will note that this beacon must contain at least the value attribute and does not have to contain body.
Example :
<coreTag:out value="Hello !" default="default value" escapeXml="true"/>

49
Created by XMLmind

XSL-FO Converter.

set markup makes it possible to store a variable and to define its scope.
It allow to instance a variable, thanks to the var attribute (define the variable name) and value attribute (value to
store). You can also modify the properties of a JavaBean with the target attribute (variable name which contains
the bean), property (property name) and value.
And finally the scope attribute will define the scope of the variable. It can take the following values: page,
request, session et application.
Example 1:
<coreTag:set value="Hello !" var="message" scope="page"/>

Example 2:
<coreTag:set value="Hello !" target="myBean" property="BeanName" scope="page"/>

To retrieve the value of a variable, you can use the out tag or use the following syntax :
${varname}
Its important to set in the page directive <%@ page %> the attribute isELIgnored at false. Indeed, this is
attribute is true by default, and it force the JSP to interpret the string "${varname}" like a common string and
not like a variable.
Example :
<coreTag:out value="${pageScope.message}" default="default value" escapeXml="true"/>

If you dont specify the scope of a variable, inside the attribute value (value="${message}"), the value will be
first searched in the page, then the request, then the session and finally the application.
remove markup makes it possible to remove a variable. You just have to specify its name in the var attribute and
its scope in the scope attribute.
catch markup is used to catch exceptions. Code which may throws exceptions must be encapsulated inside catch
markup. The var attribute contains information about the catched exception. Exception have a message property
which contains an error message.
Example :
<coreTag:catch var="Error">
<% code which may throw an exception %>
</coreTag:catch>
<coreTag:out value="${Error.message}"/>

Conditional and iteration tags: if, choose, forEach, forTokens.

url management tags: import, url, redirect.

1.5. Interactions between JSP and Servlet

1.5.1. Why make interactions between JSP and Servlet
Although they have enormous advantages (generation of image in real time, compression of pages...) when the
application requires much programming, servlets are not really adapted to the generation of code HTML, it is
also for that the JSP made their appearance.
Even if technologies like JavaBeans, TagLibs, the scriptlets made their appearance and showed their power with
regard to the integration of clean code java in the pages of result, there remains always a problem.
The JSP are provided to provide one type of presentation. If one wishes to create a complex application looking
50
Created by XMLmind

XSL-FO Converter.

to present information in a different manner according to the requests from the customer that becomes
increasingly complex.
At this point in time the transmissions of requests intervene.

1.5.2. From a Servlet

To use the transmission of request or the external inclusion of document it is necessary to use
RequestDispatcher. This object is obtained since ServletContext and the method getRequestDispatcher().
You must provide an absolute URL of the file towards which you want to transmit your request. If you want to
obtain a RequestDispatcher towards the address : http://host/folder/page.jsp it will be necessary for you to
make the following call:
RequestDispatcher dispat = getServletContext().getRequestDispatcher("/folder/page.jsp");

Once that you had your object, you only have to call the method forward() or include() of this one. You must
provide to this one the objects HttpServletRequest and HttpServletResponse.
If you use include, the resource is inserted in the active servlet.
If you use forward, the flow is completely redirected towards the new resource (caller cant write other things to
the client, this task is transmitted to the new resource only).
Inclusion or the forward towards a new resource (page or servlet) is done as follows:
[CODE:java]// Inclusion
void service(HttpServletRequest req, HttpServletResponse res) {
RequestDispatcher dispat =
getServletContext().getRequestDispatcher("/folder/page.jsp");
dispat.include(req, res);
}
[CODE:java]// Forward
void service(HttpServletRequest req, HttpServletResponse res) {
RequestDispatcher dispat =
getServletContext().getRequestDispatcher("/folder/page.jsp");
dispat.forward(req, res);
}

Remarks :

the inclusion of a servlet in another is synchronous, it means that it returns only at the end of the method
service() of the servlet called

the redirection is asynchronous

1.5.3. From a JSP script

The transmission of request in a JSP is almost the same one as that for a servlet. Indeed even if this concept is
much less used, it can be when the JSP discovers that a transmitted data is not correct for example.
The traditional method of redirection can be used with the JSP, however it is preferable to use jsp:forward for
reasons of conveniences.
Here the code which makes it possible to redirect towards another page:
// Forward
<jsp:forward page="/pageForward.jsp"/>

1.5.4. Transmission of the additional data

When you made an inclusion or a "forward" to pass the hand to a new script or a servlet. That do not means that
51
Created by XMLmind

XSL-FO Converter.

the data which you work with in the first script passed to other script.
The solution with that is to use the HttpServletRequest object and the method setAttribute(). The page of
destination will be able to then reach this information via the method getAttribute().
Since version 2.2 of the servlets a new manner of sending information has been created. You can directly put the
parameters in the URL (method GET) that you to pass at the time of the request of RequestDispatcher.
Here an example of use of this method :
[CODE:java]// Forward
void service(HttpServletRequest req, HttpServletResponse res) {
String url = "/folder/page.jsp?param1=value1&param2=value2";
RequestDispatcher dispat =
getServletContext().getRequestDispatcher(url);
dispat.forward(req, res);
}

1.6. Free host

http://developpeur.journaldunet.com/tutoriel/jav/011124jav_jspfreehost.shtml

52
Created by XMLmind

XSL-FO Converter.

Chapter 3. Struts
This chapter allow you :

to learn Struts

to understand how to create a web application with this framework

1. Course
1.1. Introduction
1.1.1. Java EE platform presentation
Java EE (Java Enterprise Edition) is composed of APIs used for enterpriseoriented application development.
Amongst those APIs, some of them already exist in the Java standard version (Java Standard Edition, Java
SE), whereas the others are specific to the enterpriseoriented platform. For example, APIs like EJBs (Enterprise
Java Beans) or JMS (Java Message Service) are only available in the Java EE JDK, but APIs like JDBC or RMI
are also included in Java SE. The Java EE platform offers an optimal solution to develop robust, secure and
evolutive applications. When you choose this technology, you have to follow a number of rules, in order to
separate your application in different layers. In this course, well see how to separate the different layers, which
model to use and with which tools.

1.1.2. Some recalls

Before starting this course, well see a little summary about JSP/Servlets, JavaBeans and Taglibs :
1.1.2.1. From Servlets to JSP
When developping a Web application with Java EE technology, Servlets are familiar. A Servlet is a Java class
loaded once by the Servlet engine (for example : Tomcat) at its first call. After that, its always the same
instance that will be used. Each time we develop a Servlet, well create a class implementing the Servlet
interface. It implies that some methods have to be implemented or redefined :

public void init(ServletConfig cfg) used at the first servlet call, when it is instanciated

public void service(ServletRequest req, ServletResponse res) called each time a client make a request on this
Servlet

public void destroy() called at the Servlet destruction

So its in the service() method that the main Java code will be written. The response is then sent to the browser,
which will display it. But with Servlets only, theres a problem : treatments and presentation code are mixed in
the same method. It doesnt respect Java EE recommendations, based on the development of separated layers as
we have seen before.
To separate treatments and presentation in Web applications, JSP are commonly used. A JSP page is like a
HTML one with additionnal JSP tags or/and Java code. Its used for displaying all treatment results, generated
in the background into Java classes. In this case, the JSP page will be converted to Servlet and will be processed
like a classical one.
1.1.2.2. JavaBeans and Taglibs

53
Created by XMLmind

XSL-FO Converter.

In our JSP page, different components are available to display the different responses generated in the lower
layers. JavaBeans are components used to instanciate and manipulate different Java objects (stored in page,
request, session or application scope) with JSP tags. The advantage : when the server encounters a tag, he will
automatically detect if the object has already been instanciated or not. If thats the case, he uses the existing one
else he creates a new instance. It allows the use of an object in several pages, and even during all applications
lifetime. Taglibs, like JavaBeans are composants used to instanciate and manipulate different Java objects
thanks to JSP tags. However, their visibility is limited to the current page. We cannot reuse the same object
instance on many pages. For each page, its a different instance. Taglibs, personnalized tags, let the possibility
to create our own tags from the simpliest to the most complex in order to display all results from different
treatments in a few lines. For more information about those components, refer to the Servlets/JSP course.
1.1.2.3. Definition and interests of a presentation framework
We have seen that JSP are an ideal solution to show treatment results in a page. They represent the applications
presentation part. Theres also the possibility to use JavaBeans or Taglibs to suppress Java code in our JSP
pages. A developer objectives is to reuse his code not to do the same things over and over again. For that, well
pick a framework which imposes a simple and efficient development method. It will produce an application
respecting the reutilisability aimed by objectoriented programmation. Concretely, a framework is the
applications skeleton. It will allow fast creation of actions, events for example forms (managing security,
validation). The presentation framework choosen is independent from the inferior layers. It helps us to
achieve our objective : layers independent from each other. Its then possible to switch a layer without having to
change the others.

1.2. MVC design pattern

1.2.1. MVC model (Model View Controller)

The ModelViewController pattern is one of the most known and famous models. It was designed first by Xerox
during the development of their first windowed system, in particular to manage user interaction with the
application. The problem solved by this MVC model is the simplification of the three major functions common
to many applications :

Data persistence in a backend storage or on a distant system

The presentation construction for the final user

The conditional logic managing which screens to show to the user, what happens when an error occurs, how
and when the distant systems are updated
Its possible to combine all those modules to obtain a functional system. However, its not easy to maintain this
code. For example with JSP, we have HTML designers for the look end feel and Java developers maintaining
the code and theyre generally different people. MVC deals with this problem by separating the code in three
54
Created by XMLmind

XSL-FO Converter.

distinct parts :

Model components, maintaining data in the storage

View components, composing the presentation displayed to the user

Controller components, managing the conditional logic. This module chooses which screens will be shownto
the user, how to handle errors and updates distant systems
MVC simplify the application maintenance by separating each logic type without mixing them up. It hides a
modules implementation details to the others and reduces dependencies between them. MVC creates a
frontier between web designers, dedicated to HTML and presentation layer, and Java developers. A
considerable advantage during a Web application development (JSP, Servlets) is that the exceptions are all
managed by the controller, so its easier to treat them. Of course its not the only advantage, here are some other
examples :

If different Views are necessary depending on the date obtained from a database or a distant system, the
Controller decides which screen to show

If the application changes depending on the date or hour, the Controller can manage that

If a data capture process requires several pages, with optional ones

However the MVC pattern is clearly defined, why are there MVC1 and MVC2?

1.2.2. MVC2 model

MVC2 model has nothing more than MVC1. This names origins are back from when Sun, releasing JSP
specifications, indicated in a section Page Access Model a model 1 and a model 2. The model 1 described
the JSP treatment and how it was commonly used at this time. The HTTP request was directly sent to a JSP
page. All treatments were done in the JSP, and the response came from the same file. The model 2 was quite
different. It indicated that it was the Servlets role to receive the HTTP request. It could then do all the necessary
tasks (like memorizing some data in a bean) and transmit the information to the JSP page used to generate the
HTTP response. No MVC name in those 2 models, but the model 2 is inspired directly by MVC.

1.3. What is Struts?

1.3.1. Presentation
Struts is a framework designed for Web applications. A framework is a structure designed to support or
contain something, in particular a skeleton used as a base for a construction . Struts is a set of Java programs
conceived to help you to build robust applications and to gain time. It provides a frame and some base elements.
Its up to you to arrange and decorate the pieces ! In addition to the fact that Struts can make you gain time, it
allows to see complex applications as a chain of base elements : Views, Actions, Models. It provides a solid
base and a set of good practices. Moreover, its useless to reinvent the wheel, Struts provides is ideal for
beginners ! However Struts is mainly designed to be used for Web applications and is perfectly conform to the
Servlet specification (version 2.4), part of the larger Java EE specification. Another advantage is the fact that
Struts is an OpenSource solution with many developers contributing to the project.

1.3.2. Architecture
1.3.2.1. Http request lifecycle

55
Created by XMLmind

XSL-FO Converter.

8.
The client sends a request to the application
9.
The Controller component receives the request. It starts taking decisions on how to continue, depending on
managing rules defined by the application
10.
Model components interact with persistent data storage
11.
Depending on what was returned by the Model, the Controller determines which View to show to and
transfers data to it
12.
The choosen View generates the response to the client
1.3.2.2. Overview of a Struts application
Struts is contitued of three main parts :

Struts librairies and dependencies

Java classes : actions, forms, exceptions
The configuration file strutsconfig. xml : application assembly (url linking, classes, forms, datasources )
1.3.2.3. Main packages
A little description of the applications main packages :

org.apache.struts : root package

org.apache.struts.action : core package with lowlevel classes for the Controller

org.apache.struts.config : package representing configuration elements (also contained in strutsconfig. xml)

org.apache.struts.taglib : root package for Struts taglibs

56
Created by XMLmind

XSL-FO Converter.

org.apache.struts.taglib.bean : package containing classes for the tag library Bean

org.apache.struts.taglib.html : package containing classes for the tag library Html

org.apache.struts.taglib.logic : package containing classes for the tag library Logic

org.apache.struts.taglib.nested : package containing classes for the tag library Nested

org.apache.struts.tiles : Tiles plugin package

org.apache.struts.upload : package simplifying upload with Struts

org.apache.struts.util : package for Struts utility classes

org.apache.struts.validator : Validator plugin package

1.3.2.4. Dependent librairies
Here are the librairies to include in your Web application to work with Struts :
struts.jar : jar containing Struts framework (Actions, forms )

commonsbeanutils. jar : tools for using Beans

commonsdigester. jar : xml mapping > Java object

commonsvalidator. jar : used for validation (Validator plugin)

commonslogging. jar : provides a logging sytem

commonsfileupload. jar : simplify upload

struts( bean / logic / nested / html).tld : Struts tag library descriptor files (optional)

1.3.3. Models : business modeling

A model component generally represents a business object or a physical backend support. For example, a
system allowing users to connect and to manage personal data may include a Model component representing an
user.
1.3.3.1. Business class
The business class provides methods to access information linked to an object (name, password ). There are
some base concepts :

57
Created by XMLmind

XSL-FO Converter.

Model classes serves as an access to database or distant systems to present them to the user
They must be conceived by hiding implementation details to data access (utilisation dinterface)
1.3.3.2. Model component example
If an application is built on a Servlet container (like Tomcat), the Model component can be limited to a simple
Java bean. It provides a business view to a JDBC logic, storing information in database. The application
interacts with the Model to readn and write data, only the Model interacts with database. In an application built
with a frontal system using JSP and a dorsal system accessed via WebServices, Model components provides a
business view of the dorsal system. The application access to the information thanks to the Model, the only one
interacting with the Webservices. Its the same with an EJB application (cf. Entreprise Java Bean). The Model
provides access to the EJB. It allows great flexibility for the application evolution but also power and flexibility
for developers.

1.3.4. Views
In MVC pattern, the View component concentrates on the presentation layer showing pages to the user. This
component must refrain from containing business logic and complex analysis. The View components are often
associated to the response to the HTTP request. In Struts case, JSP are commonly used. Those files generate the
HTML content that will be sent to the client. One of JSP advantages is that we can integrate HTML, JSP tags
and Java scriptlets in the same file. The inconvenient is that the web designers task is complicated for look and
feel maintenance! And a Java developer is easily tempted to abuse scriplets and to include business logic (or
conditional logic) directly in the JSP ! With the MVC pattern, this problem is solved because it isolates the
model, the treatments, and the views. And Struts provides a set of JSP tags helping for creating View
components. We can find various tags such as : <html:text>, <logic:iterate> giving Java developers and Web
designers the possibility to have functionalities without using Java scriplet.
1.3.4.1. Struts taglibs
Struts personalized taglibs are an important part of the framework. They greatly simplify JSP development
(View component) anf link them to other components. Those tags are useful during debugging because their
name and description are a considerable advantage over scriptlets. Struts taglibs are separated in four librairies:
Html : manage html tags (text fields, file )
Logic : manage view logic (iteration, condition )
Bean : manage bean (and its properties) access in the application (page, request, session )

Nested : contains html, logic, and bean and add an arborescence for bean access
All the descriptors of those librairies are included in the framework distribution under the name : struts<<
type>>.tld. For example : strutshtml. tld is the descriptor for the html library.
1.3.4.1.1. StrutsHtml
This library is mainly used for form handling in Struts. It covers the majority of HTML tags. A little table with
all available tags :

image: Affiche un champ image (formulaire)

img: Renders a <IMG> tag

58
Created by XMLmind

XSL-FO Converter.

javascript: Displays Javascript content used to validate a form on the client side (plugin Validator)

link Renders: a HTML anchor element

messages: Displays all messages generated during treatments

multibox Renders: a checkbox

option Renders: an option in a select tag

options Renders: a collection of options in select tag

optionsCollection Renders: a collection of options in select tag

password Renders: a password input field

radio Renders: a radio button

reset Renders: a reset button

rewrite Renders: an URI

select Renders: a select tag

submit Renders: a submit button

text Renders: a text input field

textarea Renders: a text input area

xhtml Renders: XHTML tag

1.3.4.1.2. StrutsLogic
This library contains tag used for conditional logic, iteration on collections and application stream management.

empty: Evaluates the tag content body if the variable is null or empty

equal: Evaluates the tag content body if the variable is equal to the specified value

59
Created by XMLmind

XSL-FO Converter.

forward: Transfers control to the page specified in the ActionForward entry

greaterEqual: Evaluates the tag content body if the variable is greater or equal than the specified value

greaterThan: Evaluates the tag content body if the variable is strictly greater than the specified value

iterate: Repeat the tag content body depending on the specified collection

lessEqual: Evaluates the tag content body if the variable is lesser or equal than the specified value

lessEqual: Evaluates the tag content body if the variable is lesser or equal than the specified value

lessThan: Evaluates the tag content body if the variable is strictly lesser than the specified value

match: Evaluates the tag content body if the specified value is a substring of a variable

messagesNotPresent: Generates tag content body if the specified message is not present in any scope

messagesPresent: Generates tag content body if the specified message is present in a scope

notEmpty: Evaluates the tag content body if the variable is not null/empty (works on collections too)

notEqual: Evaluates the tag content body if the variable is not equal to the specified value

notMatch: Evaluates the tag content body if the specified value is not a substring of a variable

notPresent: Generates tag content body if the specified value is not present in any scope

present: Generates tag content body if the specified value is present in a scope

redirect: Writes a HTTP header to the response

1.3.4.1.3. StrutsBean
This library is used to read and write data contained into beans, or bean themselves, or to declare new beans.

cookie: Defines a script variable based on the specified cookie value

define: Defines a script variable based on the specified bean value

header: Defines a script variable based on the specified header value

60
Created by XMLmind

XSL-FO Converter.

include: Loads the response from a dynamic application and includes it in a (type String)

message: Displays the internationalized message content in the response

page: Exposes a specific page context object to a bean

parameter: Defines a script variable based on the specified parameter value

resource: Loads the corresponding application resource and includes it in a bean (type String or InputStream)

size: Defines a script variable based on the size of a Map or Collection

struts: Exposes the Struts configuration object to a bean

write: Displays the content of a bean propriety in the page (out)

1.3.4.1.4. StrutsNested
This library is a bit special because it regroups the 3 each others. It allows tag structuration following the beans
arborescence. 3 new tags are included in this library :

nest Defines a new level that son tags could use

writeNesting Defines a new script variable based on the current level

root Starts a new hierarchy root

1.3.5. The Controller

Those are the components that lead the action ! Each time an user submits a form, its the controller that treats it.
If the user requests a page, the controller decides what todisplay. Its him again that collect data from Model
components and transmit them to the View.
Summary of the Controllers action :

Validate data submitted by the client

Take decisions about which Model components to access or to update

Retrieve data for the View component

Take decisions about how to treat errors occuring during response treatment

Decide which View to show

61
Created by XMLmind

XSL-FO Converter.

The Controller is composed of 2 parts :

Java : action class

Struts : xml configuration

The action receives data typed by the user, coordinates access to the persistent data storage, apply the
application logic and decides which View to display to the user. An action class must :

Extend org.apache.struts.action.Action class or derived class

Overload public ActionForward execute() method in order to realize the treatment specific to the action

Not contain direct link to a JSP page (this link is made in the configuration file)
This action class can range from the simple to the more complex. For example you can create a link forwarding
to an action class displaying a form (user account creation in this case). The action class will look like this :
[CODE:java]
package com.labosun.lessons.struts.action;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import org.apache.struts.action.Action;
import org.apache.struts.action.ActionForm;
import org.apache.struts.action.ActionForward;
import org.apache.struts.action.ActionMapping;
public class NewAccountAction extends Action {
public ActionForward execute(ActionMapping mappings, ActionForm form,
HttpServletRequest request,HttpServletResponse response)throws Exception {
return mappings.findForward("newUserForm");
}
}

No reference to the JSP page. Its described in the configuration file.

<action path="/addUser"
type="com.labosun.struts.action.NewAccountAction"
input="/home.jsp"
scope="request" >
<forward name="home" path="/adduser.jsp"></forward>
</action>

However the majority of actions are more complex than this one. They realize data access, treaments
Heres a set of common methods in Action class :

addErrors(javax.servlet.http.HttpServletRequest request, ActionMessages errors) : adds an error to the

request. All errors can be displayed with the <html:errors/> tag

addMessages(javax.servlet.http.HttpServletRequest request, ActionMessages messages) : adds a message to

the request. All messages can be displayed with the <html:messages/> tag

execute(ActionMapping mapping, ActionForm form, javax.servlet.http.HttpServletRequest request,

javax.servlet.http.HttpServletResponse response) : method to overload to realize specific treatment for the
62
Created by XMLmind

XSL-FO Converter.

action in progress. Returns an ActionForward object indicating the view or the action to transmit the request
to

setLocale(javax.servlet.http.HttpServletRequest request, java.util.Locale locale) : retrieves the user Locale

(useful to know from where it comes from and its language)

1.3.6. Internationalization
Struts provides a complete set of tools to manage internationalization of your web pages simply, like you could
do with JSTL. Internationalization is not only reserved for text into JSPs but also for form errors and messages.
In Struts, the name message bundle is used to describe a set of dynamic messages. A bundle is represented by a
Properties file (wellknown in Java world).
1.3.6.1. Properties file
Whats a Properties file? In Java APIs, there is a class used to read a standardized file (properties file) very
simply : java.util.Properties.
This file generally has the .properties extension an is composed of key/values couple. Each dynamic value is
linked to an unique key indepently from the language to associate several values to the same key depending on
the language. Little example of a Properties file :
key1=value1
key2=value2
key3=value3

1.3.6.2. Struts tools

Struts uses a default bundle linked to the file : ApplicationRessources.properties. However its possible to
personalize the access to this file in Struts configuration file. Its also possible to create separate files each
corresponding to a bundle. The main usage will be done in JSP thanks to the <bean:message> tag :
<bean:message key=key0 />
Ou
<bean:message key=key0 bundle=bundleName />

You can also retrieve the value in an action :

[CODE:java]
getResources(request).getMessage("key");

1.3.7. Errors handling

Struts can centralize exceptions management linked to the Controller. Connections can be made between the
exception type and the JSP (to display a message). Its possible to create and add a handler class automatically
called when an exception is cast.
Heres a diagram showing the mechanism in its globality :

63
Created by XMLmind

XSL-FO Converter.

13.
An action class cast an exception Type Exception
14.
Struts detects a configuration for this exception in the configuration file
15.
Depending on the configuration, Struts call the Handler class for this exception
16.
Struts redirects the client to the configured jsp error page

1.3.8. Configuration file

The configuration file is the core of a Struts application. Indeed, it gathers all objects and their configuration. All
the application logic (navigation) is described in this file. Despite the complexity implied by xml usage in large
Web applications (depending of the number of functionalities), tools have been released to help the developer in
parameter settings and theyre stable and robust.
1.3.8.1. Application assembly
The configuration file (strutsconfig. xml) role is to indicate to Struts which forms and actions are available. It
also includes the possibility to add plugins to the Struts base package.
Well

learn

all

struts

base

tags

thanks

to

the

DTD

provided

by

http://struts.apache.org/dtds/strutsconfig_ 1_2.dtd (1.2 version DTD).

The first tag in the configuration file is strutsconfig . Heres the corresponding DTD:
<!ELEMENT strutsconfig
(displayname?,
description?, datasources?,
formbeans?,
globalexceptions?,
globalforwards?,
actionmappings?,
controller?,
messageresources*,
plugin*)>
<!ATTLIST strutsconfig
id ID #IMPLIED>

Its called the root tag. Theres only one in each file. It accepts in its body the following tags :

displayname : application name linked to the configuration file

description : configuration file description

datasources : datasource declaration (1 instance max.)

formbeans : all formbean (1 instance max.)

globalexceptions : all exception handlers (1 instance max.)

globalforwards : all global forwards (1 instance max.)

64
Created by XMLmind

XSL-FO Converter.

Apache

actionmappings : all action mappings (1 instance max.)

controller : specific controller setting (RequestController) (1 instance max.)

messageresources : bundle message declaration (multiple instances)

plugin : activation of plugins (multiple instances)

1.3.8.2. Datasources
Your models may need to establish a connection to a database. In this case, you can use Struts to initialiaze a
connection pool instead of doing this in the application server. Its not needed anymore to change the server
configuration when deploying the application. It makes deployment easier.
DTD part for datasources :
<!ELEMENT datasources
(datasource*)>
<!ATTLIST datasources
id ID #IMPLIED>
<!ELEMENT datasource
(setproperty*)>
<!ATTLIST datasource
id ID #IMPLIED>
<!ATTLIST datasource
className %ClassName; #IMPLIED>
<!ATTLIST datasource
key %AttributeName; #IMPLIED>
<!ATTLIST datasource
type %ClassName; #REQUIRED>

Each datasources tag contains all datasources for the application. It contains 0 or n datasource tags.
Each datasource tag definites precisely all configuration parameters to a datasource (and more precisely JDBC
2.0 DataSource objects). Here are all different available tags :

className : optional tag. It defines which class will be used to configure the datasource. It must contain all
necessary data to instanciate a DataSource according to its type. It has to manage all specific properties in
setproperty tag. If the tag is not present, Struts use the default class : org.apache.struts.config.Config to
manage properties

key : a key linked to a datasource. For example, for a stockdb database, the application can access it via
org.apache.struts.action.DATA_SOURCE/stockdb (if no prefix exists : prefix = "")

type : specify the DataSource class name that implements java.sql.DataSource. This class is used by Struts to
create a datasource instance, its configured with the class defined in the className attribute

setproperty : used for connection parameters for the datasource, maps a key to a value (login, password,
database parameters)
Corresponding DTD :
<!ELEMENT setproperty
EMPTY>
<!ATTLIST setproperty

65
Created by XMLmind

XSL-FO Converter.

id ID #IMPLIED>
<!ATTLIST setproperty
property %PropName; #REQUIRED>
<!ATTLIST setproperty
value CDATA #REQUIRED>

property : the key (property)

value : value bind to the key

1.3.8.2.1. Exemple
XML example of the datasources clause :
<datasource>
<setproperty
property="autoCommit" value="true"/>
<setproperty
property="decription" value="DataSource MysqlLocal"/>
<setproperty
property="driverClass" value="com.mysql.jdbc.Driver"/>
<setproperty
property="maxCount" value="4"/>
<setproperty
property="minCount" value="1"/>
<setproperty
property="user" value="root"/>
<setproperty
property="password" value=""/>
<setproperty
property="url" value="jdbc:mysql://localhost/contacts"/>
</datasource>

1.3.8.2.2. Application usage

You can acces to the DataSource object from an Action with the following code :
[CODE:java]
DataSource ds = servlet.getServletContext().getAttribute(org.apache.struts.action.DATA_SOURCE/stockdb);

Now, its possible to retrieve a connection to the datasource to realize operations on it.
1.3.8.3. FormBeans
formbeans and formbean indicates to Struts to bind an unique indentifier to an ActionForm class. Theyre useful
to initialize DynaForms too.
Corresponding DTD :
<!ELEMENT formbeans
(formbean*)>
<!ATTLIST formbeans
id ID #IMPLIED>
<!ATTLIST formbeans
type %ClassName; #IMPLIED>
<!ATTLIST formbean
id ID #IMPLIED>
<!ATTLIST formbean
className %ClassName; #IMPLIED>
<!ATTLIST formbean
dynamic %Boolean; #IMPLIED>
<!ATTLIST formbean
name %BeanName; #REQUIRED>
<!ATTLIST formbean

66
Created by XMLmind

XSL-FO Converter.

type %ClassName; #REQUIRED>

<!ELEMENT formproperty
(setproperty*)>
<!ATTLIST formproperty
className %ClassName; #IMPLIED>
<!ATTLIST formproperty
initial CDATA #IMPLIED>
<!ATTLIST formproperty
name %PropName; #REQUIRED>
<!ATTLIST formproperty
size %Integer; #IMPLIED>
<!ATTLIST formproperty
type %ClassName; #REQUIRED>

Little description of each tag :

formbeans : encapsulate the list of all defined form beans

formbean : defines a unique form :

className : used to define a configuration bean for this form (like for DataSource), not common

dynamic : deprecated

name : unique identifier used to access to the form to eventually associate it to an action. Important notice :
if the forms scope is session, you can access to the form with the getAttribute() method from the session
object

type : precises a valid ActionForm class, instanciated to create the form

If type extends org.apache.struts.action.DynaActionForm, formproperty tags are inspected by Struts. Each one
defines an initial value, a name and a type (for example java.lang.String) for each form property.
Example
<formbean name="contactForm"
type="com.society.contactbook.application.forms.ContactForm" />

1.3.8.4. Exceptions
The globalexceptions tag gathers all exceptions managed by the application during a request treatment. Its
possible to manage as many exceptions as desired. This configuration section can be assimilated to a try
catch block in Java language. You declare a catch block with an exception tag.
Attributes admitted by this tag :

bundle : specifies the bundle name that will be used for this exception

handler : specifies which class will be called when the exception is cast. It must extends the
org.apache.struts.action.ExceptionHandler class or derived class

key : indicates the key to search for and use for the generated error message in the selected messages bundle

67
Created by XMLmind

XSL-FO Converter.

path : defines the URL path to transfer the control to when the exception is cast

scope : tells to Struts if it needs to memorize ActionError created by the exception handler in session or
request scope

type : exception name intercepted by the handler

1.3.8.4.1. Example
<globalexceptions>
<exception key=database.error path=/errors/generalErrors.jsp scope=request type=java.sql.SQLException />
</globalexceptions>

If an exception is cast during a process, database.error model is sought in the default resources bundle
(ApplicationResources.properties), a new ActionError is created thanks to the model, and then placed in request
attributes. Control is finally transmitted to the standard error page.
1.3.8.5. Global Forwards
Global forwards are useful to declare pathes accessible by all defined actions in the configuration file. For
example, if many actions needs to transfer control to a login page, define the transfer in this section instead of
doing it separately in each action.
Corresponding DTD :
<!ELEMENT globalforwards
(forward*)>
<!ATTLIST globalforwards
id ID #IMPLIED>
<!ATTLIST globalforwards
type %ClassName; #IMPLIED>
<!ELEMENT forward (icon?, displayname?,
description?, setproperty*)>
<!ATTLIST forward id ID #IMPLIED>
<!ATTLIST forward className %ClassName; #IMPLIED>
<!ATTLIST forward contextRelative %Boolean; #IMPLIED>
<!ATTLIST forward module %RequestPath; #IMPLIED>
<!ATTLIST forward name CDATA #REQUIRED>
<!ATTLIST forward path %RequestPath; #REQUIRED>
<!ATTLIST forward redirect %Boolean; #IMPLIED>

Attributes description :

contextRelative : indicates if path is relative to an application module (if false) or for all the application (if
true)
name : represents the unique forward identifier. Its possible to get this forward in an action with the
findForward() method

redirect : if positioned at true, control is transmitted to the page like a redirection instead of a transfer. So a
new request is created
Generally, setproperty are not used for forwards. But hey can be useful if you want to create your own forward
classes (extending the base one). In this case, you have to precise the complete class name in the className
attribute.
1.3.8.5.1. Example
68
Created by XMLmind

XSL-FO Converter.

globalforwards usage example :

<globalforwards>
<forward name=home path=/pages/home.jsp />
</globalforwards>

1.3.8.6. Action mappings

This part is assurely the most complex. It links actions, forms and forwards.
Corresponding DTD :
<!ELEMENT actionmappings
(action*)>
<!ATTLIST actionmappings
id ID #IMPLIED>
<!ATTLIST actionmappings
type %ClassName; #IMPLIED>
<!ELEMENT action (icon?, displayname?,
description?, setproperty*,
exception*, forward*)>
<!ATTLIST action id ID #IMPLIED>
<!ATTLIST action attribute %BeanName; #IMPLIED>
<!ATTLIST action className %ClassName; #IMPLIED>
<!ATTLIST action forward %RequestPath; #IMPLIED>
<!ATTLIST action include %RequestPath; #IMPLIED>
<!ATTLIST action input %RequestPath; #IMPLIED>
<!ATTLIST action name %BeanName; #IMPLIED>
<!ATTLIST action parameter CDATA #IMPLIED>
<!ATTLIST action path %RequestPath; #REQUIRED>
<!ATTLIST action prefix CDATA #IMPLIED>
<!ATTLIST action roles CDATA #IMPLIED>
<!ATTLIST action scope %RequestScope; #IMPLIED>
<!ATTLIST action suffix CDATA #IMPLIED>
<!ATTLIST action type %ClassName; #IMPLIED>
<!ATTLIST action unknown %Boolean; #IMPLIED>

Description of each tag :

attribute : unique ID used tomemorize theActionForm in reqest or session scope, else name is used

forward / include : transfers control to a new path instead of processing the action

input : indicates the page or action which generated the input form processing data
parameter : transmits a parameter to the action, but its preferable to use the setproperty tag
path : links an action to a path it must be the path without suffix. For example if you specify
/admin/add.do in the action form then you have to put /admin/add in the path

prefix / suffixe : specify a prefix/suffix added to the ActionForm bean properties names before associating
them to request parameters

roles : restrict access to an action with a list of rols separated by , .

69
Created by XMLmind

XSL-FO Converter.

type : Action class.

validate : if the attribute is false,the the controller dooesnt execute the ActionForms validate() method
Example:
<action path="/Welcome"
scope="request"
validate=false
type="com.society.contactbook.application.actions.WelcomeAction">
<forward name="success" path="home" />
</action>

1.3.8.7. Controller
This tag is not used often. Only people with the need to modify Struts core will use it !
<!ELEMENT controller (setproperty*)>
<!ATTLIST controller id ID #IMPLIED>
<!ATTLIST controller bufferSize %Integer; #IMPLIED>
<!ATTLIST controller className %ClassName; #IMPLIED>
<!ATTLIST controller contentType CDATA #IMPLIED>
<!ATTLIST controller forwardPattern CDATA #IMPLIED>
<!ATTLIST controller inputForward %Boolean; #IMPLIED>
<!ATTLIST controller locale %Boolean; #IMPLIED>
<!ATTLIST controller maxFileSize CDATA #IMPLIED>
<!ATTLIST controller memFileSize CDATA #IMPLIED>
<!ATTLIST controller multipartClass %ClassName; #IMPLIED>
<!ATTLIST controller nocache %Boolean; #IMPLIED>
<!ATTLIST controller pagePattern CDATA #IMPLIED>
<!ATTLIST controller processorClass %ClassName; #IMPLIED>
<!ATTLIST controller tempDir CDATA #IMPLIED>

Attributes description :

bufferSize : set the buffer size for processing file loading

maxFileSize : set maximum file size to be loaded, could be followed by K (kilo), M (mega) or G (giga)

contentType : set content type, by default text/html

debug : activates debug if value > 0

forwardPattern : modify the way an application is linked to a context relative URL with forwardPattern.$A
value is expanded as application prefix (/applicationContext for example) and $P as required path. Default
value $A$P

inputForward : when set to true, input attributes parameters from action tags are processed like forwards, not
pathes. It means theyre compared to forward tags defined globally and locally instead of beig processed as
URI
locale : memorize a locale in user session if its not already done when this attiribute is set to true

70
Created by XMLmind

XSL-FO Converter.

nocache : sends a deactivation request of the content cache to the client browser for each HTTP response

pagePattern : tells Struts how to link pages to subjacent URL, similarly to forwardPattern

processorClass : replaces Struts request processor by a personalized one

Example :
<controller nocache=true contentType=image/jsp/>

1.3.8.8. Messageresources
This tag section allows to declare application resources bundles. Its not limited to one, several bundles can be
declared.
<!ELEMENT messageresources
(setproperty*)>
<!ATTLIST messageresources
id ID #IMPLIED>
<!ATTLIST messageresources
className %ClassName; #IMPLIED>
<!ATTLIST messageresources
factory %ClassName; #IMPLIED>
<!ATTLIST messageresources
key %AttributeName; #IMPLIED>
<!ATTLIST messageresources
null %Boolean; #IMPLIED>
<!ATTLIST messageresources
parameter CDATA #REQUIRED>

Attributes description :

factory : specify message resource data location. By default, properties files are configured

key : defines a key for a bundle.

null : sets application behaviour when a message is not found. If attribute is set to true, it returns null. When
set to false, returns an error message with the faulty key.

parameter : transmits a parameter to the factory when a bundle is created. For factories based on property
files, its the file path.
Example :
<messageresources
key=com.labosun.MAIN_MESSAGES
parameter=com.labosun.MainMessages />

1.3.8.9. Plugin
This last element is used to load plugins dynamically in Struts framework.
<!ELEMENT plugin
(setproperty*)>
<!ATTLIST plugin
id ID #IMPLIED>
<!ATTLIST plugin
className %ClassName; #REQUIRED>

71
Created by XMLmind

XSL-FO Converter.

Attributes description :

className : plugin lading class.

setproperty : set properties needed by the plugin

Example Validator plugin actvation :
<plugin
className=org.apache.struts.validator.ValidatorPlugIn>
<setproperty
property=pathnames
value=/WEBINF/
validatorrules.
xml,
/WEBINF/
validation.xml />
</plugin>

1.3.9. Web application integration

Weve seen how Struts operate independently from a Web application web built on a servlet engine.
Of course Struts could be integrated with a Web application, else it would be useless ! Well realize a mapping
to redirect an URL path to Struts framework servlet.
It is org.apache.struts.action.ActionServlet. There are properties to set the configuration file, debugging
Then we map the servlet to an url.
Descriptor file (web.xml) example for the Web application :
<!Standard
Action Servlet Configuration (with debugging) >
<servlet>
<servletname>actionstruts</servletname>
<servletclass>org.apache.struts.action.ActionServlet</servletclass>
<initparam>
<paramname>config</paramname>
<paramvalue>/WEBINF/strutsconfig.xml</paramvalue>
</initparam>
<initparam>
<paramname>debug</paramname>
<paramvalue>2</paramvalue>
</initparam>
<initparam>
<paramname>detail</paramname>
<paramvalue>2</paramvalue>
</initparam>
<loadonstartup>2</loadonstartup>
</servlet>
<!Standard Action Servlet Mapping >
<servletmapping>
<servletname>actionstruts</servletname>
<urlpattern>*.do</urlpattern>
</servletmapping>

1.4. Tiles
1.4.1. Presentation

72
Created by XMLmind

XSL-FO Converter.

During a Web application development, its noticeable that many pages possess their own central element with
more or less static ones around it (navigation bar, header & footer )
Tiles library is a Struts plugin, used to create templates for our application thaks to another XML configuration
file.
Its possible to use this library without Struts. Well how to associate it with Struts (to see how powerful a Struts
application can be with Tiles).

1.4.2. Architecture

This architecture preserves the separation between page structure and page content. Its possible to create as
many layouts as desired. In the configuration file, to apply a layout to a page, you have to write that this page
extends this layout. We can see that the application layer is not present into the JSP pages, instead they use use a
structuring
layer
(tiles)

73
Created by XMLmind

XSL-FO Converter.

Chapter 4. Web Service

With this course you will :

Define a Web Service

Respect Web Services Norms

Use existing Web services

Create Web Services

1. Course
1.1. Web Services Generalities
1.1.1. Definition
Device conversions for implementing in a Web application or for Widget available on certain operating systems,
you surely dealt with the Web services without too being aware of it.
A Web Service is a functionality implemented in many languages and system in a layer of standards derived
from the XML. It must be able to be required and called upon dynamically by other services.
Generally, the Web Services are basic functionalities of the companys available server. This technology,
initiated by IBM and Microsoft, then partly standardized by the W3C, is now accepted largely by other
information & computing industrys players without exception.
This is main point which makes Web Services a revolutionary technology and is very popular.
It makes it possible for the applications to dialogue through the network, independently of their platform of
execution (Windows, Linux) and of the type of the client (Mobiles, Office Applications ) or their
implementation language. They fall under the continuous initiatives such as CORBA (Common Object Request
Broker Structures, of l' OMG) in bringing however a simpler answer, being pressed on technologies and
recognized standards which characterize the concept even of Web service:

SOAP

WSDL

UDDI
There is a multiple use of Web Services like in B2C ,B2B, stock management domains, etc.

B2C (Business to Customer): Qualified application, a website addressing general public.

B2B (Business to Business): Qualified application, a website addressing professional to professional.

Which applications are intended for Web Services technologies? Why to use them? How to setup them? Web
74
Created by XMLmind

XSL-FO Converter.

Services technologies can be applied to all kinds of applications to which they offer considerable advantages in
comparison to the former API owners, with the specific platform traditional implementation restrictions that one
can meet. They can be useful in the majority of the scenarios when the communication can be established on a
bidirectional model (request/response). It is nevertheless farto be as restrictive, much of other models can have
recourse to the Web Services, without same as you in returned account. The companies which place them at the
disposal allow the developers interested by its functionalities to re-use them without having to re-code them.

1.1.2. How it works?

Basically, the pure technological aspects are not innovative. On the contrary, the architecture of the Web
Services is important (just like language XML) thanks to its simplicity, with its legibility and its standardization.

The concept of the Web Services is currently articulated around the three following acronyms:

SOAP (Simple Object Protocol Access) is inter-application exchange platform independent protocol based on
XML language. A call of SOAP service is an ASCII flow framed in XML and transported with HTTP
protocol.

WSDL (Web Services Language Description) gives the description of XML format of Web Services by
specifying the methods being able to be called upon, their signatures and the access point (URL, port, etc.). It
is, to some extent, the equivalent of the IDL distributed programming language CORBA.

UDDI (Universal Description, Discovery and Integration) standardizes a solution of distributed directory of
Web Services, allowing at the same time the publication and exploration (research) of Web Services. UDDI
behaves almost like a Web service of which the methods are called via protocol SOAP.
We will thus see in details these standards and their operation in the continuation of this course (cf.chp 2)

75
Created by XMLmind

XSL-FO Converter.

The XML (eXtended Markup Language) is at the base of all the described protocols here. The fact that the Web
Services use the XML brings the advantage of non-ownership and thus really makes it multi platform. It is thus
recommended to have a minimum of bases (XML, DTD, diagrams, XSL,) in order to be able to really set up
of the Web Services optimized.
Example of Web services installation in a company:

1.1.3. Architecture
76
Created by XMLmind

XSL-FO Converter.

The Web Services architecture rests on a mechanism of transport of service request between a customer and a
server, both connected to network (corporate network or Internet).

Current standardization around the Web Services is however a vast building site which is a simple method
invocation of a distant object. Various work thus started to define a true distributed infrastructure, able to satisfy
all needs for a distributed application, as well in terms of standardization of the exchanges as in terms of
services.
One can figure this organization of the standardization committees according to following cutting:

Standardization of the transverse services on three horizontal axes:

Transport Layer: Definition of the structure of the messages used by the applications to be discovered and
dialogue between them

Semantics Layer: Standardization of the data taking part in the exchanges according to trades criteria

Process management Layer: Standardization of the processes management exchanges which extend on
several applications available on Internet

Standardization of the transverse services on three vertical axes:

Directory Service: Standardization of the means of access to a service for a request relating to the contents
of service or on a supplier

Security service: Standardization of the means allowing to cover the problems of authentification and
management of the access rights.

Transaction Service

77
Created by XMLmind

XSL-FO Converter.

Standardization of the means guaranteeing integrity of the long transactions implying several Web Services
(orchestration of Web Services).
All the major actors (J2EE or .NET) propose a directed Web Services solution as well as the integration of the
associated standards. In parallel, the first uses of this technology are born, in particular in the field of the CRM
solution (Sevina or Onyx), human resources management (CCMX) or of the contents (System U, Digiplug).

1.1.4. Advantages - Disadvantages

Advantages:

Non-ownership

Multiplatform

No Firewall restrictions (Use of port 80)

Facilitates the development

Transparent for user

Layer Operations
The principal disadvantage which is not really is on the level of safety. In the use of the port 80 which makes it
possible to avoid any problem of firewall also allows outside penetrating the information systems without access
controls nor analyzes contents. But this problem remains foreseeable and with thus solutions during the
development. We thus will see more in detail safety of the Web services with under following chapter (Cf.
chap.1.5).

1.1.5. Security
The current problems of safety in Web service have varied causes. Such as:

Installations without safety: password by defect or stored in light, files accessible in reading or read/write

Use of the same certificate for session SSL and XML signature

Programs can infiltrate the company via the Web Services.

A non interest for the safety part in the projects Web service
But certain solutions are already applicable to split that. As to use combined SSL/TLS has certificates or
signatures XML. Following advises are not negligible to set up your clean Web services:

To use a relay HTTP/HTTPS/SOAP/XML

Syntactic and lexical validation before any transfer to a connector XML

78
Created by XMLmind

XSL-FO Converter.

Sending with the antivirus

To force an analysis safety in the analysis of the project

A filtering IP can also be a solution for the accesses to small scales

Example of a solution using the XKMS (XML Key Management Services) which makes it possible to manage
protected information (key, certificate, signature) of a Web service:

Another solution are addressed in chapter WSDL (cf.chap. 2.2) The Web service is expanding considerable; to
take into account safety in this field is thus not negligible.

1.2. Protocols and normes

1.2.1. SOAP
Web Services communicate with SOAP protocol.
79
Created by XMLmind

XSL-FO Converter.

SOAP or Simple Object Access Protocol, is a simple and light protocol to share informations. SOAP is different
from RMI, CORBA and COM, because it use an system of auto description of data. It define XML message
structure used by applications to communicate.
You can use it to:

call methods(SOAP RPC)

exchanging messages (SOAP messaging)

1.2.1.1. Message structure

SOAP define the principal structure of the message named "envelop" which have two parts :

header (optional)

body

It define also encodage for different data types which based on XML technology. Datas could be simple (string,
integer, float, ...) or advanced.
SOAP manage datas's format during sending or receiving. So, there are two types of trafic:

message

80
Created by XMLmind

XSL-FO Converter.

envelop
All messages SOAP in an XML document named envelop which are containers of a SOAP message. An
envelop is the first element in an XML document. Envelop is formated with SOAP specifications, the sender
and the consumer could be written on different platforms.
1.2.1.2. HTTP request
Les requtes HTTP consistent avoir une mthodologie de requte, une requte en URL qui spcifie les enttes
et corps du flux envoy. Les messages SOAP sont transmis via HTTP permettent une complte interoprabilit
entre les clients et les Web Services. Lors de lenvoi dune requte HTTP, le champ dentte doit avoir la valeur
SOAPAction : pour que le message soit assimil un message SOAP.
We use request methodology by adding header and body (stream) in the URL. The header field must have the
value "SOAPAction" in order to be know like a SOAP message.
GET return source
HEAD return the header
POST send value to the server
PUT save a resource for the request
DELETE delete the resource
OPTIONS return HTTP method provide by the server
HTTP 1.0 include only methods GET, HEAD and POST. Although J2EE servers want to have only HTTP 1.0,
some of servers use HTTP 1.1 version.
POST /LocalWeather HTTP/1.0
Host: www.mindstrm.com
ContentType:text/xml; charset="utf8"
ContentLength:328
SOAPAction: "WeatherStation"
<SOAPENV:Envelope
xmlns:SOAPENV="http://schemas.xmlsoap.org/soap/envelope/"
SOAPENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<SOAPENV:Body>
<m:GetCurrentTemperature xmlns:m="WeatherStation">
<m:scale>Celsius</m:scale>
</m:GetCurrentTemperature>
</SOAPENV:Body>
</SOAPENV:Envelope>

1.2.1.3. HTTP response

HTTP response use the same process. A HTTP response have:

the result code

the header

81
Created by XMLmind

XSL-FO Converter.

the body
Here are some errors with HTTP protocol:
404 resource is unavailable
401 need authentification
500 internal error
503 the server is overloaded
HTTP/1.0 200 OK
ContentType:text/xml; charset="utf8"
ContentLength:359
<SOAPENV:Envelope xmlns:SOAPENV="http://schemas.xmlsoap.org/soap/envelope/"
SOAPENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<SOAPENV:Body>
<m:GetCurrentTemperatureResponse xmlns:m="WeatherStation">
<m:temperature>26.6</m:temperature>
</m:GetCurrentTemperatureResponse>
</SOAPENV:Body>
</SOAPENV:Envelope>

IBM created a new standard named HTTPR or HTTP Reliable, HTTP Fiable, in order to make sending et
receiving SOAP message reliable.
1.2.1.4. Namespaces
It's important to understand XML namespaces to understand SOAP.
<m:GetCurrentTemperature xmlns:m="WeatherStation">
<m:scale>Celsius</m:scale>
</m:GetCurrentTemperature>

Let's go to analyse this XML code. XML allow you to create variables names like in a programmation language.
Here the namespace identity called "m" and this variable have "WeatherStation" as namespace. This principle of
variable is very used in data description files. The exemple "m:GetCurrentTemperature" allow to identify
namespace thank to "m".
XML use default namespaces, you can meet streams like that:
<GetCurrentTemperature xmlns="WeatherStation">
<scale>Celsius</scale>
</GetCurrentTemperature>

1.2.2. WSDL
1.2.3. UDDI directory
UDDI (Universal Description, Discovery and Integration) is a specification which defines how to publish and
discover Web Services through a search engine. Thus, in order to grant access to a new service, we must create a
file called Business Registry which describes the service using XML Language with UDDI specifications.
UDDI allows to classify and search Web Services. In fact, if a consumer cannot find a specific Web Service, it
means nothing to deploy it. Thats why the idea of a universal directory has been established. Unlike LDAP
(Lightweight Directory Access Protocol) directories, UDDIs doesnt deal with resources such as hardware,
software or peoples. In fact, they are private directories based on the Internet. They are usually used for
Business to Business (B2B) exchanges. UDDI directories provide both technical information (WSDL
documents) as usual information about the company which provides the service (the company Web page,
82
Created by XMLmind

XSL-FO Converter.

products and so on). The UDDI API is divided in two interfaces. The first registers Web Service in the
directory and the second is used to search inside of the directory. This API has two main libraries:

Publishing API

Request/Research API
1.2.3.1. Searching a Web Service
You can search Web Services through a search engine which is embedded in the website of the UDDI operator.
You can define a lot of search criteria in order to refine your request:
Companys name
Companys location

Web Service name

...
Thanks to the Request API, UDDI directories will locate Web Services which are deploy on all public servers in
the world. By searching a Web Service, you can find a lot of information about the company which provide this
service, even other services they can offer.
UDDI operators certify the security and the integrity of Web Services they have in their directories.
Calling operators websites grant access to information about a Web Service entry in the UDDI directory :

find_binding : get the link to the service

find_business : search the identity of the company which provide the Web Service

find_relatedbusiness : find a list of companies which are link (partnership, departments, subsidiary) to this
company

find_service : get the definition of the service

find_tmodel : get the data model of the service

get_bindingdetail : get individual fields with the link of find_binding

get_businessdetail : get attributes of the company with find_businessdetails

get_servicedetail : get the service thanks to find_service

83
Created by XMLmind

XSL-FO Converter.

get_tmodeldetail : get the model thanks to find_tmodel

1.2.3.2. Publishing a Web Service
Like WSDL, a UDD link regroups, for a given communication protocol, technical data to consume the Web
Service (IP address, domain name, how to use the service). In order to publish a Web Service, a company
must authenticate to the UDDI operator (or register first if it does not). When a company chooses an operator,
further changes or update must be done in this same operator. You can simultaneously register a group of
affiliate companies, relationship between other companies and so on. The publishing API has three major parts:

Manipulation (save and delete)

Authentication of commands by using tokens (get_authToken and discard_authToken)

Adding relations between companies (joint_ventures)

1.3. In practice
1.3.1. JWS (Java Web Services)
Java Web Services, also known as JWS, is a new standard for implementing Web Services. This standard was
first written by WebLogic Workshop. Also, this new standard is now supported by the Java Community
Process. The main aim for JWS is to provide an easily usable and understandable implementation for Web
Services. AXIS is based on this standard to provide features enabling the developer to deploy and use Web
Services.
1.3.1.1. Deploy a Web Service using JWS
In order to deploy a Web Service using JWS features, we just have to rename a Java source file describing the
Web Service, replacing the file extension to .jws instead of .java For example, a file named
MyHelloWS.java should be renamed into MyHelloWS.jws. Once our JWS file created, we have to copy
this file into the webapps/axis directory, inside Tomcat server directory. You can verify that your Web
Service was deployed, trying to access it with a Web browser at the following URL :
http://localhost:8080/axis/MyHelloWS.jws
You should have access to the following page :

You can view the WSDL corresponding to your Web Service by clicking on the previous link, or by adding
?wsdl to the Web Service URL.
1.3.1.2. Undeploy a Web Service using JWS
Undeploy a Web Service using JWS is a very easy task. You just have to delete the JWS file previously
deployed on the server. If you try to access the URL for an undeployed Web Service, you'll see the following
page :

1.3.2. Axis
Apache AXIS is a SOAP (Simple Object Access Protocol) implementation for JWS which provides some
utilities to help developers create and deploy Web Services. You can find more informations at
http://ws.apache.org/axis/ Also, you have to download Axis at the URL above, in order to install it in your
84
Created by XMLmind

XSL-FO Converter.

Tomcat server.
1.3.2.1. Installation
In our case, we'll use Axis inside a Tomcat Servlet container. To install Axis inside our Tomcat server, you just
have to copy the webapps directory contents from the Archive Axis to the webapps directory inside Tomcat.
You should new have a directory tree for your Tomcat looking like this :

Now, start your Tomcat server using the startup.bat script located in the bin/ repertory inside your Tomcat
folder, or via your favorite IDE.
Now, try to access the following URL: http://localhost:8080/axis You should see the following page:

You can try to click the Validation link, in order to check the Axis installation.

85
Created by XMLmind

XSL-FO Converter.

You can also click the List link on the home page in order to list all the deployed Web Services:

1.3.2.2. Features
Axis provides some features related to Web Services. You can easily deploy or undeploy Web Services inside
your Tomcat server, thanks to Axis. Also, you'll have access to the WSDL directly trough your Web Browser.
No .wsdl file are written, WSDL is dynamically generated. You'll also have access to your Web Services
methods via a web browser, specifying the Web Service name and the method you want to invoke. Finally, Axis
helps developers writing console, Swing and/or JSP clients thanks to some specific classes included into Axis,
which will simplify the client development. A traffic monitor is also available to check every calls on every
Web Services deployed on your server.
1.3.2.3. Invoke Web Services
Invoking a Web Service can be made via a Web browser, a console client or even a Web client. If you want to
invoke a method via a Web browser, you'll have to specify which method you want to invoke though the
method parameter.
For example, to invoke the sayHello method,
http://localhost:8080/axis/MyHelloWS.jws?method=sayHello

we'll

access

You'll now see the following page, corresponding to the returning SOAP message:

86
Created by XMLmind

XSL-FO Converter.

the

following

URL:

1.3.2.4. Modify your environment variables

During the following steps, we'll use an environment variable called axis_libs. Here are the steps to create this
environment variable on Windows:

axis_libs corresponds to a system variable containing the path of all Axis' JAR files. You have to create this
variable with the following value:
%CATALINA_HOME%\webapps\axis\WEBINF\lib\activation.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\axis.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\axisant.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\commonsdiscovery0.2.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\commonshttpclient3.0rc2.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\commonslogging1.0.4.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\jaxrpc.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\log4j1.2.8.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\mailapi_1_3_1.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\saaj.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\servlet.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\wsdl4j1.5.1.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\xercesImpl.jar;
%CATALINA_HOME%\webapps\axis\WEBINF\lib\xmlParserAPIs.jar

These paths are using the system variable CATALINA_HOME. This variable should contain the Tomcat
installation path, for example: d:\java\tomcat5.0.28 For Linux users, use the export command, as shown
below:
export axis_libs=${CATALINA_HOME}\webapps\axis\WEBINF\lib\activation.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\axisant.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\axis.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\commonsdiscovery0.2.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\commonshttpclient3.0rc2.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\commonslogging1.0.4.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\jaxrpc.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\log4j1.2.8.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\mailapi_1_3_1.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\saaj.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\servlet.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\wsdl4j1.5.1.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\xercesImpl.jar:
${CATALINA_HOME}\webapps\axis\WEBINF\lib\xmlParserAPIs.jar

Be careful, your environment variable can't contain spaces.

1.3.2.5. Deployment descriptors
In order to use advanced deployment features, you'll have to use deployment descriptors. This priciple,
implemented by Axis, allows you to specify the methods you want to publish, the port type, the encoding...
These elements are customizable only during when deploying using the JWS principle we exposed above. You'll
have to use WSDD files (WSDD standing for Web Service Deployment Descriptor). Create two of these
specific files : one for deployment and another for undeployment.
By convention, these files are named deploy.wsdd and undeploy.wsdd.
1.3.2.6. Deployment

87
Created by XMLmind

XSL-FO Converter.

Actually, the WSDD file is an XML file used at deployment time. The WSDD file should contain the used name
spaces in order to check it's validity.
In a second time, you have to use the service markup, specifying the following attributes:

The Web Service name, to identify it into your server

The Web Service mode: RPC (which stands for Remote Procedure Call)
Also, you have to specify at least two parameter tags:

One of them should define the name of the class associated to the Web Service. Using deployment
descriptors, you can specify different names for the Web Service and for its class.

The second one define the methods published into the Web Service.
Here is an example for a deploy.wsdd file:
<deployment xmlns="http://xml.apache.org/axis/wsdd/" xmlns:java="http://xml.apache.org/axis/wsdd/providers/java">
<service name="MyHelloWorldWebService" provider="java:RPC">
<parameter name="className" value="MyHelloWorldWS"/>
<parameter name="allowedMethods" value="*"/>
</service>
</deployment>

Once the file created, you'll have to use it in order to deploy your Web Service. To do this, you'll have to use an
utility provided by Axis and called AdminClient. This utility will also be used to undeploy a Web Service. The
syntax to use is the following: java cp %axis_libs% org.apache.axis.client.AdminClient deploy.wsdd axis_libs
is corresponding to the environment variable previously created and containing the path to all JAR files
available into the Axis' lib directory.
After executing the line above, you'll obtain the following response:

You can now access the Axis' home page and click on the List button. You'll see the following page, listing
all the deployed Web Services:

This feature, listing all the deployed services, wasn't available using the JWS deployment. Here, you can also
read the WSDL corresponding to each service.
You

can

access

the

following

URL

to

verify

your

Web

88
Created by XMLmind

XSL-FO Converter.

Service

deployment:

http://localhost:8080/axis/service/MyHelloWS You'll obtain the following page:

If you want to invoke a method on this Web Service, you can proceed the same way we saw above with a Web
Service deployed using the JWS principle. You just have to add the method parameter into your URL. Here is
the
URL
to
call
the
sayHello
method
on
the
MyHelloWS
Web
Service:
http://localhost:8080/axis/services/MyHelloWS?method=sayHello
You'll now have access to the following page, corresponding to the returned SOAP message:

1.3.2.7. Undeployment
Undeploying a Web Service previously deployed using a Deployment Descriptor is very easy. We have to create
an undeploy.wsdd file. It's an XML file containing a service markup, with a single name attribute. The
name value is the name of the Web Service you want to undeploy. Here is an example, undeploying the Web
Service named MyHelloWS:
<undeployment
xmlns="http://xml.apache.org/axis/wsdd/">
<service name="MyHelloWorldWebService"/>
</undeployment>

Once the file created, we'll use the AdminClient utility, as we have done during deployment.
The syntax is exactly the same, we just change the WSDD file name: java cp %axis_libs%
org.apache.axis.client.AdminClient undeploy.wsdd
Executing the above command produces the following output:

On the page listing all the deployed Web Services, we can see that MyHelloWS has been correctly undeployed:

89
Created by XMLmind

XSL-FO Converter.

Trying to access an undeployed Web Service will produce the following page:

You're now able to deploy and undeploy Web Services using WSDD files.

1.3.3. The WSDL2Java utility

1.3.3.1. Utility presentation
WSDL2Java is another utility provided by Axis. It allows you to produce some classes and interfaces from a
WSDL file, enabling you to access the Web Service in a Java application. The difference between the classes is
made when we call the utility. Until now, we've seen how to create a Web Service, deploy and undeploy it, but
in most of the cases, we'll reuse existing Web Services. Once you've located the WSDL file you're interested in,
save the corresponding WSDL file into a local file, with the .wsdl extension.
1.3.3.2. Client side class generation
In order to use a Web Service into a Java program, we have to use Java classes and interfaces. These class and
interfaces can be written by the developer, but they are heavy to write, and aproximatively the same for each
Web Service. That's why we use the WSDL2Java utility to write these classes for us.
Here is an example on how to use this utility: java cp %axis_libs% org.apache.axis.wsdl.WSDL2Java
NomDuFichierWSDL.wsdl
The WSDL2Java utility produces lot of Java classes and interfaces:

A Java class for each complex type defined by the WSDL

An interface, defining all the method exposed by the Web Service

A proxy client (stub) for each binding defined by the service. A binding is the implementation of a service
corresponding to a given exchange model (RPC or messages) using a given protocol (SOAP) trough a given
transport protocol (HTTP)

A service locator, comparable to a Factory, which will create implementations for the interface
1.3.3.3. Server side generation
The server side generation is similar to the client side generation. We use the same command, adding some
specific options. This principle will be used only if you want to deploy the service on your server. The main
90
Created by XMLmind

XSL-FO Converter.

advantage of this mode is the automatic WSDD generation.

Here is the syntax to use: java cp %axis_libs% org.apache.axis.wsdl.WSDL2Java serverside skeletonDeploy
true NomDuFichierWSDL.wsdl
The utility will generate the following files:

deploy.wsdd

undeploy.wsdd

localhost\axis\<NomDuWS>_jws\<NomDuWS>.java

localhost\axis\<NomDuWS>_jws\<NomDuWS>Service.java

localhost\axis\<NomDuWS>_jws\<NomDuWS>ServiceLocator.java

localhost\axis\<NomDuWS>_jws\<NomDuWS>SoapBindingImpl.java

localhost\axis\<NomDuWS>_jws\<NomDuWS>SoapBindingSkeleton.java

localhost\axis\<NomDuWS>_jws\<NomDuWS>SoapBindingStub.java

1.3.4. Java 6 new features

1.3.4.1. JAX-WS
JAX-WS (Java API for XML based Web Services) is the new name for JAX-RPC (Java API for XML based
Remote Procedure Calls), because the JAXB 2.0 was incompatible with JAX-RPC 1.1. It's also an
appropriateness for Sun to update their products using the new language features and others new technologies.
This API is based on JAXB 2.0 and uses the new Java language facilities. They relieve the developer from
adding informations hitherto necessary to transform a Java class into a Web Service.
1.3.4.2. JAXB (Java Architecture for XML Binding)
This new version allows bidirectional binding. Also, it is now possible to define bindings only on an XML
fragment, instead of a full XML document.
1.3.4.3. Web Services creation steps
Setting up a Web Service is very important into the enterprise's logic, and it is important to identify correctly the
treatments to do.
The analysis before developing the Web Service is very important. The objectives are:

Identify the treatments

Expose the services

Administrate
91
Created by XMLmind

XSL-FO Converter.

Coordinate
Web Services can take place into different layers inside the enterprise information system. JAX-RPC is used
during the development of Web Services.
This API supports the following data types (on top of the primitive types):

java.lang.Boolean

java.lang.Byte

java.lang.Double

java.lang.Float

java.lang.Integer

java.lang.Long

java.lang.Short

java.lang.String

java.math.BigDecimal

java.math.BigInteger

java.net.URI

java.util.Calendar

java.util.Date

1.3.5. Web Service integration

The integration of Web Services inside the Java EE platform is done in many ways. You can use generation and
integration tools, they will base them self on existing libraries and API you could use yourself. The element you
need to set up a Web Service are the following:

Java API for Web Services: JAX-WS, JAXP, JAXM, JAXR, JAX-RPC

Apache's libraries: Axis, XML-RPC, WSIF

92
Created by XMLmind

XSL-FO Converter.

Web Service build using Axis

Coordination tools: BPEL, BPWS

These notions are not mandatory to master. This course has only for aim to present you the different aspects of
Web Services. If you want more informations , you now have some starting points to do some researches.

1.3.6. Searching a Web Service

To search for an existing Web Service, you just have to lookup into an UDDI directory (Universal Description,
Discovery and Integration). The UDDI defines different directories allowing to expose the services from an
enterprise to another. An enterprise can publish a Web Service she developed into a directory in order to share
them with persons who may need it. Thanks to these directories, you can find working Web Services. Some are
free, some others are evaluation versions, and some others are professional editions and not free. You can search
for some existing Web Service.
You just have to use a search engine via an UDDI directory. A lot of web sites exists which are dedicated to this
feature:

https://uddi.ibm.com/beta/find
http://www.strikeiron.com
http://uddi.microsoft.com/
1.3.7. Invoking a Web Service
Once you find, thanks to UDDI directories, the Web Service you want to use, you have to get the corresponding
WSDL address, for example: http://www.xmethods.net/sd/2001/TemperatureService.wsdl The client invoking
the Web Service can be either a Web browser, or a client application (Swing, SWT, ...). The invocation can be
asked by the user or be automated (Web Service invoking another, automated task, ...).
The client can be another platform than the one hosting the Web Service (Main frame, server, desktop, laptop,
PDA, mobile phone, ...). Also, if the Web Service is written using C#, clients can be developed using Java, PHP,
C, ...
Actually, the client doesn't has to know the language used to develop the Web Service in order to use it.
93
Created by XMLmind

XSL-FO Converter.

Knowing that invoking a Web Service go trough always trough a lot of XML dialects, we talk about XML
Web Services. It's mandatory to master the XML syntax in order to understand how Web Services
communicate
between
them,
and
how
implementations
are
build.

94
Created by XMLmind

XSL-FO Converter.

Chapter 5. Java Micro Edition

By completing this course, you will:

Discover JavaME

Learn the bases

Practice

Enjoy it !

1. Course
1.1. Introduction
1.1.1. Presentation
Java Micro Edition (Java ME), is the most recent Java Plateform.
Incoming of new technologies are often gone with peculiarity, each constructor wanting to leave its claw. That
led to heterogeneity in the same domain however having the same goal. These bringing the nuisances that
everyone knows, the necessity for entirely reencode entirely a program which works on a machine X and not on
Y.
SUN already has a solution for this problem, remember in the World of personal computers with their principle
which consists on compile once, run anywhere.
The idea was to do the same with for the mobile devices, however towards available material, the use of
framework Java SE was impossible. An example, the power processor of a mobile phone is far from a PC.
So, Java ME has been thinking and conceives for mobile devices, including cell phones, PDA, TvBox and any
other device with limited resources.
However, dont forget Java SE, because Java Me approaches it on many aspects, and uses many of its concepts
(identical syntax, bytecode, package organisation, etc).
We will discover that Java ME is composed of 3 fundamental elements:

Configurations

Profiles

Optional APIs.

1.1.2. Three Fundamentl elements

Figure 5.1.

95
Created by XMLmind

XSL-FO Converter.

As we can see on this diagram, there are 2 configurations:

CLDC for Connected Limited Device Configuration

CDC for Connected Device Configuration.

We will particularly be interested in CLDC which is the most used in mobile devices.
With a Configuration, there is always a Profile. Like MIDP for Mobile Information Device Profil, or MHP for
Multimedia Home Plateform and lot of others.
By adding such profile with such configuration, we can adapt as well as possible to the material.
Then there is the optional API, which manufacturer provides in order to take benefits from specificities of their
product. Like a camera, or the bluetooth on the cell phone.

1.1.3. Software
In order to be able to develop for these devices, in particular cell phones, we can use the Wireless ToolKit that
we can download on SUNs website.
Although sufficient to develop, it is more simple to use an integrated development environment (IDE) than a
notepad, we will use Netbeans with the plugin Netbeans Mobility which provided us the necessary.
1.1.3.1. The Wireless ToolKit
To use the Wireless ToolKit, just install it.
In this kit, we will find a lot of tools, we will use one of them: the KToolBar.
It makes possible to compile, use, and emulate the project.
Here is a representation of it:

Figure 5.2.

96
Created by XMLmind

XSL-FO Converter.

On the shell, there is a compiled project, then emulated.

The comboBox lets us choose the Skin phone that we want to emulate.
If you have Projects to add to emulator on ToolKit, you will have to put them in a special folder .
1.1.3.2. EclipseMe
Refer to Slide or Annexes.
Its use will be shown by the SCT SUN.

1.2. Configuration profil and optional API

1.2.1. Configuration: CLDC
1.2.1.1. Definition
A configuration is composed of a virtual machine and a whole of basic API.
CLDC (Connected limited Device Configuration) defines common basis to small devices. Its role is to define an
adaptable platform on all the basic configurations of type 16 or 32 bits (16Mhz minimal), 128 Kb of memory,
and a connection of weak bandwidth. The virtual machine attached is named KVM for Kilobyte Virtual
Machine.
Besides of devices that CLDC is intented, it is obvious that it cannot implement too greedy specificities in
memory. So it was necessary to decide something, And in the first version(1.0), there is no floating point, no
RMI, no serialisation, and just few Errors are managed.
In next version 1.1 of CLDC, there was changes, for example, there is floating point, Why ? because devices
evoluates.
1.2.1.2. feature
Seeing javadoc, we found 5 packages, which someones are from Java SE:

java.io: Provides classes for input and output through data streams

java.lang: Provides classes that are fundamental to the Java programming language

java.lang.ref: Provides support for weak references

97
Created by XMLmind

XSL-FO Converter.

java.util: Contains the collection classes, and the date and time facilities

javax.microedition.io: Classes for the Generic Connection framework

For most of these packages, you must already know them. There is one that is specific to Java ME,
javax.microedition.io, it will be discussed further. Indeed, it is taken again (and completed) by the Profile.

1.2.2. Profile
A profile, as this word indicates, is usefull to specialize. In other words that means that profile will correspond
to a type of device or a whole of functionalities.
There are considerable profiles, like MIDP, Foundation Profil, Personal Profil. Each one of them is used with a
Configuration. Example: MIDP with CLDC, Foundation Profil with CDC
We will take interest in MIDP.
MIDP (Mobile Information Device Profil) provides necessary packages for graphic uses, data persistence and
for connectivity.
Applications developed with this Profile are named MIDlet, acronym of MIDP and Applet because they have
the same life cycle.

1.2.3. Optional API

We will not talk a lot of this, indeed, optional API is even specialized because (?) most of the time for only one
device. These API is provided by the manufacturer of the device to use pecificities from this one, for example a
camera.

1.3. Profile MIDP

1.3.1. Packages
MIDP provides following packages:

java.lang: Same as Java SE, Fundamental classes to the Java programming language. Ex: management of
Wrappers (String, Float, Boolean, etc.)

java.util: Same as Java SE, Contains the collection classes, and the date and time facilities

javax.microedition.io: Classes for the Generic Connection framework adapted to CLDC

javax.microedition.lcdui: Provides a set of features for implementation of user interfaces

javax.microedition.lcdui.game: Provides a series of classes that enable the development of rich gaming
content

javax.microedition.media: Provides classes to manage media. The MIDP 2.0 Media API is a directly
compatible building block of the Mobile Media API (JSR135) specification

javax.microedition.media.control: This package defines the specific Control types that can be used with a
Player

98
Created by XMLmind

XSL-FO Converter.

javax.microedition.midlet: This package define MIDP applications and the interactions between the
application and the environment in which the application runs

javax.microedition.pki: Provides necessary to use Certificates which are used to authentication on secure
Connections

javax.microedition.rms: Provides necessary to use data persistence

Now we have choosen a Configuration and a Profile, we are able to begin programming.

1.4. The midlet

1.4.1. Definition
A MIDlet is a Java program for embedded devices which runs on a java ME machine.
Like there is a lot of Profiles, this java ME machine must implement MIDP.

1.4.2. How To Write A Midlet

Like any application, making MIDlet applications requires following a path :
17.
Coding application

Ressources are limited, doing an application must be in code and treatments

ex: Your application would have not more than 5 classes and a size of 130ko
18.
Compilation et Prechecking

The process compilation is the same as for a Java SE application

La prechecking is a specific task to Java ME

With a Java SE application, the stage to check the classes, like calls of package are made by the virtual
machine. However this operation is extremely expensive in memory resource.
So for Java ME, it makes a previsualisation after the compil time. This stage adds information in
Bytecode, and some flags that will be used by the KVM.
This made KVM (JVM for Java ME) not to make again the checking, it just has to detect flags, and if they
are not good, it starts an exception.
19.
Test on Emulator
20.
Packaging
21.
Real Test

99
Created by XMLmind

XSL-FO Converter.

If you have cellphones, it is preferable to check on them that your application works. Indeed, although a
standard has been established, blurs remain, and results in some virtual machine are not very ended. In this
way, your application can work on a Motorola device and not on Nokia. Sometimes there are differences
on same devices because they dont have the same firmware.

Most of the time, Software publishers have a specific team to testing compatibility of their products in
many devices.

For the game development, sometimes, most part of the time ( ) is attribuate to test the application, in
order to test it on devices which are on the market.

1.4.3. Midlet's states

There are 3 states in a MIDlet: Active, Pause, Destroyed.
All these states corresponds to a method that you must redefine when you program your application.
These methods are represented in this diagram:

Figure 5.3.

As you can note it on the diagram, the 3 methods are: startApp(), pauseApp() et destroyApp().
You will find them in a class that extends MIDlet:
The method startApp() is equivalent of main() in a Java SE program. Indeed, in Java ME there is no main()
method.
The method startApp() is equivalent of main() in a Java SE program. Indeed, in Java ME there is no
main() method.
100
Created by XMLmind

XSL-FO Converter.

The method pauseApp() signals the MIDlet to enter the Paused state.

The methode destroyApp(Boolean unconditional) signals the MIDlet to terminate and enter the Destroyed
state.It takes a parameter, if it is true , MIDlet is destroyed, if it is false , we ask the MIDlet to be
destroyed, if MIDlet is not ready to be sestroyed, it throw a MIDletStateChangeException.
There is a simple MIDlet which post Welcome and allows you to shut down application while pressing on a key.
[CODE:java]import javax.microedition.midlet.*;
import javax.microedition.lcdui.*;
public class premierMIDlet extends MIDlet implements CommandListener {
private Command commande1;
private Command commande2;
private Command commande3;
private Display display;
/**
* Constructor
**/
public premierMIDlet() {
display = Display.getDisplay(this);
commande1 = new Command("Cmde 1", Command.EXIT, 1);
}
/**
* Start the MIDlet
**/
public void startApp() {
// Creating a textbox.
TextBox t = new TextBox ("First MIDlet", "Welcome", 256, 0);
// Associate the command commande1 to the listener.
t.addCommand (commande1);
// Set up the listener
t.setCommandListener (this);
display.setCurrent (t);
}
/**
* Pause the MIDlet
**/
public void pauseApp() { }
/**
* Clean what the garbage collector dont deal, and destroy the MIDlet
**/
public void destroyApp (boolean unconditional) {}
/**
* Response to a command request.
**/
public void commandAction (Command commandeActivee, Displayable d) {
if (commandeActivee == commande1) {
destroyApp (false);
notifyDestroyed();
}
}
}
Now, were going to describe more deeply some points of this code.
The result of the MIDlet compiled and emulated by the Sun Toolkit :

Figure 5.4.
101
Created by XMLmind

XSL-FO Converter.

Skin avec le DefaultGrayPhone

Figure 5.5.

102
Created by XMLmind

XSL-FO Converter.

Skin DefautColorPhone

Figure 5.6.

103
Created by XMLmind

XSL-FO Converter.

Skin MediaControlPhone

Figure 5.7.

104
Created by XMLmind

XSL-FO Converter.

Skin QWERTYDevice

1.5. User interface

1.5.1. Introduction
Compare your computers screen with your cellphones one. What difference do you notice? Add to that the
processors frequency, and you will understand why its not possible, at least for now, to have same user
interfaces.
So, we find similarities between Java SE and Java ME on object level, but no direct comparison is possible.
Indeed, Java ME does not use the same API like Java SE for user interface.
We will find two API, the low level API, and the high level. (we will discuss them).
The first allows handling directly the screen, but there is no access to user commands.
The second allows using graphic components for user interface. However, it can not be display directly on the
screen.

105
Created by XMLmind

XSL-FO Converter.

We will be able to use this APIs in the same application, but not at the same time, remember this 2 API use the
screen, but there is only one screen.
It is well to use simple interfaces, especially to propose lists of choice because informations input is boring for
a user.

1.5.2. Display
It will be necessary to make difference between several objects on which display rests, here are these objects:

Displayable: An object that has the capability of being placed on the display

Display: Display represents the manager of the display and input devices of the system

Screen: The common superclass of all highlevel user interface classes

Canvas: Its reference of low level API. The Canvas class is a base class for writing applications that need to
handle lowlevel events and to issue graphics calls for drawing to the display
Firstly, we will need a virtual display where we will made our modification and after show it on screen, this
virtual display is an object of type Display.
This object will be initialised by the method getDisplay(MIDlet m) and will be single on your MIDlet.
Now, you know you only have one virtual display, but you can have a lot of forms to put on this display.
To exhibit an element on the display, you have to call the method setCurrent() and give it a parameter: a
container (it will be discuss later). And to get the current container, we can use the method getCurrent().
When you get the object of type Display, you will be able to have informations about the screen like:

Is it a color screen ?
isColor() This method returns a boolean

If yes, how many colors?

numColors() This method returns an int representing the number of colors

1.5.3. High-Level Interface

Of course, the highlevel interface inherits package javax.microedition.lcdui.
This interface is extended by four classes, Alert, List, TextBox et Form.

Figure 5.8.

This interface is more independent of device than lowlevel interface, so it is turned towards the portability.
106
Created by XMLmind

XSL-FO Converter.

Take a look on each class that implements This interface.

1.5.3.1. Alert
This class is similar to a dialog box on Java SE.
It will be used to transmit to user particular information, like an error, a confirmation request
There ar two constructors to make an object of type Alert :
[CODE:java]Alert(String title)
[CODE:java]Alert(String title, String alertText, Image alertImage, AlertType alertType)

It will have been guessed, the first argument of these constructors is the title of dialog box, this being the only
one compulsory. In the second constructor, we can add a message which indicates with detail the error for
example, a possible image, and the alerts type. You should know there is five types of alerts:

ALARM: A hint to alert the user to an event for which the user has previously requested to be notified

CONFIRMATION: A hint to confirm user actions

ERROR: A hint to alert the user to an erroneous operation

INFO: provides nonthreatening information to the user

WARNING: A hint to warn the user of a potentially dangerous operation

Of course, there are some methods (getter and setter) which provide modification of this parameter thereafter:
getString(): return Alerts text
setString(String str): set the Alerts text to display

getType(): return the reference to the object AlertType which is used

setType(AlertType type): Set the AlertType to use

Finally , getTimeout() returns an int which indicates how much time the Alert will be visible.
And setTimeout (int time) will set up this time. (Note that time can be FOREVER).
So we will display this Alert, in following way: display.setCurrent(alert, display.getCurrent).
The Alert will be displayed int the foreground.
1.5.3.2. List
The constructor List can take some parameters:
[CODE:java]List(String title, int listType)

Create a list with a title and a type of list (one of IMPLICIT, EXCLUSIVE, or MULTIPLE ).

107
Created by XMLmind

XSL-FO Converter.

[CODE:java]List(String title, int listType, String[] stringElements, Image[] imageElements)

Create a list with a title, a type list, elements of the list, and images of this elements. The most used is the first
constructor.
To add some element to a list, we can use the method append(String stringPart, Image imagePart) setting
Image parameter to null if you dont want images. The parameter stringPart will show the text in the list.
To empty a List, use the method deleteAll() .This will delete all the elements in the list. If you want to delete a
specific element, use the method delete(int elementNum) which takes in parameter the id of the element.
To find which one is selected in the list, you can use the method getSelectedIndex(). This method returns the id
of the element.
So, you can select the default element. For this selection there are two methods available:
[CODE:java]setSelectedFlags (boolean[] selectedArray)
[CODE:java]setSelectedIndex (int elementNum, boolean selected)

The first takes in parameter a Boolean tables and the second allow you to modify select an element usin his id.
But you can unselect elements using the same method, just say false for parameter selected.
With the first method we can select all the elements at the same time, also for second method we have to call it
for each selected element.
Also, we can modify an element of the list using the method set(int elementNum, String stringPart, Image
imagePart), but if you just want to change the text you can use getString(int elementNum) which returns the text
of elementNum, such for image with getImage(int elementNum).
1.5.3.3. TextBox
The TextBox is similar to a TextArea in Java SE.
Like you see above, the constructor takes parameters, the title, the default text which will be displayed,
maximum size of the text and some constraints:
[CODE:java]TextBox(String title, String text, int maxSize, int constraints)

Constraints can be ANY, EMAILADDR, NUMERIC, PHONENUMBER, URL, PASSWORD.

ANY: The user is allowed to enter any text.

EMAILADDR: The user is allowed to enter an email address.

NUMERIC: The user is allowed to enter only an integer value

PHONENUMBER: The user is allowed to enter a phone number

URL: The user is allowed to enter a URL

PASSWORD: Confidential data that should be obscured whenever possible

see the MIDP specifications
The constraint PASSWORD could be associate with another on using character | .
108
Created by XMLmind

XSL-FO Converter.

So we will find setters to modify the values of this argument thereafter:

[CODE:java]setMaxSize (int maxSize), setString(String text) et setConstraints(int constraints)

Another method to just modify a part of the text :

[CODE:java]setChars (char[] data, int offset, int length)

We can also insert text with:

insert (char[] data, int offset, int length, int position)
[CODE:java]insert (String src, int position)

And delete text with:

[CODE:java]delete (int offset, int length)

Offset is the beginning of the text to erase, and length, number of characters to delete. With this method is easy
to delete all the text, just set offset to 1, and te length use size() to have the length of the text.
1.5.3.4. Form and Items
This object will be useful for Forms, its a container.
Form (String title)

As for other classes, there are two constructors, with parameter title in both cases, and for the second constructor
a list of items.
This list will be the component which form contains.
We can also detect how much items there is with the method size() which returns an int.
In same way, we will be able to select each items with get(int itemId). This method will return an object of type
item.
Thereafter we can add more items using the method append(Item item). which returns an int, the id of the item.
Of course, we can insert and delete items using insert(int itemNum, Item item) and delete(int itemNum). And to
erase all items at the same time use deleteAll().
1.5.3.4.1. Items
Here are different items that we can use:
ChoiceGroup, CustomItem, DateField, Gauge, ImageItem, Spacer, StringItem, TextField
All of them have some methods in common, in particular, those allowes modify its label: Text that generally is
display with the component.

getLabel (): return the label

setLabel(String Label): set the Label

Let us describe some:
1.5.3.4.2. StringItem
StringItem is a text which the user will not be able to edit it.
StringItem(String label, String text) is the constructor, it takes the label and the text to display as parameters.

109
Created by XMLmind

XSL-FO Converter.

We can modify this text, or get them thanks to:

getText(): return the text

setText (String text): set the text.

1.5.3.4.3. Gauge
A graphical display, such as a bar graph.
Gauge(String label, boolean interactive, int maxValue, int initialValue)

The label allows to annotate the gauge, the Boolean allows user to modify or not the gauge, and next ther is the
maximum value and the initial value..
And there is again getter and setter to modify parameters when the application is running.

getMaxValue(): return the maximum value

getValue(): return the actual value

isInteractive(): Return a Boolean, to know if it can be modify by user

setMaxValue(int maxValue): Set the maximum value

setValue (int value): Set the actual value of the gauge

1.5.3.4.4. TextField
Fields TextField correspond to the same object as on JavaSE.
TextField(String label, String text, int maxSize, int constraints)

Its constructor is similar to TextBoxs one.

So, constraints are the same as TextBox.

ANY

EMAILADDR

NUMERIC

PHONENUMBER

URL

PASSWORD

110
Created by XMLmind

XSL-FO Converter.

We can add constraint to PASSWORD adding | with other one.

We can modify parameters with these methods:

setMaxSize (int maxSize), setString(String text) et setConstraints(int constraints) .

setChars (char[] data, int offset, int length)

And insert text with:

insert (char[] data, int offset, int length, int position)

insert (String src, int position)

And delete text with:

delete (int offset, int length)

1.5.3.4.5. Tickers
Its a piece of text that runs continuously across the display.
We will see it more in detail as an independent component. But it can be integrated into a column by the method
setTickers(Tickers myTickers).

1.5.4. Commands
The first approach is simple.
To manage events, your MIDlet has to implements the interface CommandListener.
So, your MIDlet will have to redefine the methode
commandAction (Command c, Displayable d)

Just implement it once, but you know that you can have a lot of commands. So to know what command has
been pressed, you will use the object Command, it contains the name of the object you used.
The Displayable object contains informations of display.
When you create a command, you just have to add it to a form, there is an example:
import java.io.IOException;
import javax.microedition.midlet.*;
import javax.microedition.lcdui.*;
/**
*
* @author Michel
* @version
*/
public class Midlet extends MIDlet implements CommandListener{
// a command
Command c;
public void startApp() {
Display d = Display.getDisplay(this);
d.setCurrent(seeTextBox());
}
public void pauseApp() { }
public void destroyApp(boolean unconditional) {}

111
Created by XMLmind

XSL-FO Converter.

public TextBox seeTextBox(){

TextBox tbox = new TextBox("Text box"," Be Aware !",
10,TextField.UNEDITABLE);
// create a new command.
c = new Command("OK",Command.OK,1);
tbox.addCommand(c);
return tbox;
}
//redefine the method commanAction
// testing the command which is received.
public void commandAction(Command command, Displayable displayable) {
if (command.getCommandType() == Command.OK){
System.out.println("Hellllllloooo");
}
}
}

1.5.5. Low-Level Interface

The lowlevel interface allows us to uses all functionalities of the device.
The class Canvas and Graphics inherit from this lowlevel API Canvas.

Figure 5.9.

1.5.6. Canvas
There is the Canvas constructor, its very simple :
Canvas()

This class allows us to manipulate the display, because it has access to the events due:

At keys says Key Events

on Game Actions, usually by the joystick

To stylus

To commands that we can add ourselves. (cf 5.4)

1.5.6.1. Keys
Every key has an associated code:

Static int KEY_NUM0 for 0

Static int KEY_NUM1 for 1

112
Created by XMLmind

XSL-FO Converter.

Static int KEY_POUND for #

Static int KEY_START for *
The events management relating to each key is managed by the following methods :
protected void keyPressed(int Keycode)

This method is called when a key is pressed..

protected void keyReleased(int Keycode)

this one when a key is released.

protected void keyRepeated(int Keycode)

This one when key is repeated, i.e. that one remains to support above.
It should be noted that this event could not be support by the device, so its necessary to check it using this
method:
boolean hasRepeatEvents()

1.5.6.2. Game actions

The profil MIDP preset game actions.
UP, DOWN, LEFT, RIGHT, FIRE, GAME_A, GAME_B, GAME_C, et GAME_D.
Each action have da define value:
Ex : public static final int LEFT a une valeur de 2
In order to extend an application, a game, we can use some methods.
int getGameAction(int keyCode)

This method allows to transform a key code into a game action

int getKeycode(int gameAction)

This method makes the opposite.

1.5.6.3. The stylus
Some devices are equipped with a touch screen, so we can use a stylus to operate on the display.
To manage that, some methods are provided, there are three.
protected void pointerPressed(int x, int y)

This method is called when the pointer touches the screen

protected void pointerDragged(int x, int y)

This one is called when the pointer is dragged on the screen

protected void pointerReleased(int x, int y)

this one is called when the pointer is released.

It should be noted that this events could not be support by the device, so its necessary to check it using this
method:
public boolean hasPointerEvents()

1.5.6.4. Events
113
Created by XMLmind

XSL-FO Converter.

Management of events has already been discussed before.

Your MIDlet has to implement the interface CommandListener.
See 5.4 for more information.

1.5.7. Graphics
This has some similarities with the one which you can find on the JDK Java SE. The difference is thaht there are
less methods in Java ME. Thi class allows to make simple 2D graphics like squares, lines, words, lle permet
aussi de crer des graphiques simples 2 dimensions. Tel que des rectangles, des lignes, des mots.
Some methods like:
drawImage (Image img, int x, int y, int anchor)
drawString (String str, int x, int y, int anchor)

They take in parameter an in anchor. Anchor is the point for positioning the image, the string .. Example: if
we want to draw an image, on which one we define the anchor point on his center, so the his center will be
located on x,y.
The definition of the anchor point must be one of the horizontal constants (LEFT, HCENTER, RIGHT)
combined with one of the vertical constants (TOP, BASELINE, BOTTOM).
Example: LEFT | TOP .
Another important method,
setClip (int x, int y, int width, int height)

Rendering operations has no effect outside of the clipping area.

The following example illustrates these methods.
import javax.microedition.lcdui.Canvas;
import javax.microedition.lcdui.Display;
import javax.microedition.lcdui.Graphics;
import javax.microedition.lcdui.Image;
import javax.microedition.midlet.MIDlet;
import javax.microedition.midlet.MIDletStateChangeException;
public class TestGraphics extends MIDlet {
private Display d;
private Dessin1 visual1;
private Dessin2 visual2;
public TestGraphics() {
super();
d = Display.getDisplay(this);
visual1 = new Dessin1();
visual2 = new Dessin2();
}
protected void startApp() throws MIDletStateChangeException {
d.setCurrent(visual1);
visual1.repaint();
try {
Thread.sleep(2000);
} catch (InterruptedException e) {e.printStackTrace();}
d.setCurrent(visual2);
visual2.repaint();
}
protected void pauseApp() {}
protected void destroyApp(boolean arg0) throws
MIDletStateChangeException {}
}
class Dessin2 extends Canvas{

114
Created by XMLmind

XSL-FO Converter.

Image myImage;
public Dessin2(){
try{
myImage = Image.createImage("/1.png");
}catch(Exception e){
e.printStackTrace();
}
}
protected void paint(Graphics g) {
g.drawImage( myImage, this.getWidth()/2, this.getHeight()/2,
Graphics.VCENTER | Graphics.HCENTER );
}
}
class Dessin1 extends Canvas{
public Dessin1(){}
protected void paint(Graphics g) {
g.setColor(0,0,255); //On choisit la couleur Bleu
g.drawRect(5,5,25,25); //On dessine un carr bleu
g.fillRect(35,5,25,25); // On dessine un carr plein bleu
g.setColor(0,255,0); //On choisit la couleur verte
g.drawString("Hello tout le monde",5,50,Graphics.LEFT |Graphics.TOP);
// On dessine une chaine de caractere, la font peut etre modifier par
// setFont(Font _font)
g.setColor(255,0,0) // On choisit la couleur rouge
g.drawLine(1,90,110,130); //On dessine une ligne rouge
g.drawRect(40,100,50,50); //On dessine un carr rouge
g.setColor(0,0,255);
g.setClip(40,100,50,50); //On ne modifiera plus que la zonedefinie
//ici, remarqu quil sagit du carr predefini juste au dessus
g.drawLine(1,90,110,130); //On redessine la mme ligne en bleu
}
}

sortie:

Figure 5.10.

1.6. Multimedia
1.6.1. Presentation
The MIDP 2.0 Media API is a directly compatible building block of the Mobile Media API (JSR135)
specification. It opens us the doors of mutlimedia, but contrary to the MMA, the MIDP 2.0 Media API is limited
to the audio.
Media just define one class, Manager, but also there are 4 interfaces Control, Controllable, Player,
PlayerListener and 1 exception class MediaException..

115
Created by XMLmind

XSL-FO Converter.

1.6.2. Concept
The concept of this api rests on 3 blocks:

Manager

Player

Control

Figure 5.11.

A Manager can have more than one Player, each one for a specific data type.
Ex: a player for data type Wav, another one for MP3
Each player can have several controls, we have to define them.

1.6.3. Methods
static Player createPlayer(InputStream stream, String type)

116
Created by XMLmind

XSL-FO Converter.

Allows the creation of a Player using a stream, and indicating its type of stream. Ex of type : "audio/Xwav"
static Player createPlayer(String locator)

Create a player starting from an URL ex : createPlayer("http://www.MonSite/song.wav")

static String[] getSupportedContentTypes(String protocol)

Return the list of supported content types for the given protocol
static String[] getSupportedProtocols(String content_type)

Return the list of supported protocols given the content type

static void playTone(int note, int duration, int volume)

Play back a tone as specified by a note, its duration and its volume ranging between 0 and 100.
import java.io.IOException;
import java.io.InputStream;
import javax.microedition.media.Manager;
import javax.microedition.media.MediaException;
import javax.microedition.media.Player;
import javax.microedition.media.control.ToneControl;
import javax.microedition.midlet.MIDlet;
import javax.microedition.midlet.MIDletStateChangeException;
public class TestMedia extends MIDlet {
public TestMedia() {
super();
// TODO Autogenerated
constructor stub
}
protected void startApp() throws MIDletStateChangeException {
String[] type = Manager.getSupportedProtocols("audio/xwav");
int i = type.length;
for (int j=0;j<=i;j++){
System.out.println(type[j]);
}
try {
InputStream is = getClass().getResourceAsStream("BONAL.WAV");
Player p = Manager.createPlayer(is, "audio/Xwav");
p.start();
} catch (IOException ioe) {
} catch (MediaException me) { }
try {
Manager.playTone(ToneControl.C4, 80000 /* ms */, 100 /* max
vol */);
} catch (MediaException e) { e.printStackTrace();}
}
protected void pauseApp() { }
protected void destroyApp(boolean arg0) throws
MIDletStateChangeException { }
}

This code returns the list of supported protocols by the type wav. The list can be shown in the shell.
This code shows us how to play a song named here Bonal.wav, this one is on the folder project.
And we can show how to diffuse the song.

1.7. THE NETWORK

1.7.1. Introduction
MIDP includes the network support based onto the Generic Connection framework from CLDC.
All classes bound to the communication through the network take place in the javax.microedition.io package.
117
Created by XMLmind

XSL-FO Converter.

This package provides all classes and interfaces requisited to the client/server communication mainly throughout
HTTP 1.1 protocole.
Besides http is the only protocole that a MIDP implementation must contain, it means that all the other
protocoles (TCP, UDP) are not required because a lot of devices dont get an IP stack.

1.7.2. First Example

This is the example of a very simple http connexion. From an URL, we get the information retrieved from a
server. In our example, we call the research page of Google and we observe the result.
HttpConnection hc = null;
InputStream in = null;
String url = http://www.google.fr;
try {
hc = (HttpConnection)Connector.open(url);
in = hc.openInputStream();
int length = 256;
byte[] raw = new byte[length];
int readLength = in.read(raw);
String message = new String(raw, 0, readLength);
printForm.append(message);
} catch (Exception e) {
printForm.append(
new StringItem("Exception: ", e.toString()));
} finally {
if (in != null) {
try { in.close(); } catch (IOException ioe) {}
}
if (hc != null) {
try { hc.close(); } catch (IOException ioe) {}
}
}

As you can notice, this example get us the HTML code of the Google research welcome page. If the URL was
wrong, a error message would appear.
Lets check the steps to respect to get a connexion.

1.7.3. The Connexion steps

1.7.3.1. Choose the connexion type
First of all, you have to determine the type of connexion you need. These are the main types of connexion
represented by the following interfaces:

Connection: the most basic type. Can only be opened and closed

InputConnection: represents the connexion to a device from which you can read data

OutputConnection: represents the connexion to a device to which you can write data

StreamConnection: this interface gathers the Input and Output types of connexion

ContentConnection: this interface is a StreamConnection subinterface. It provides methods to write basic

metadata useful for the http connexions

StreamConnectionNotified: this interface waits for a connexion to be established

118
Created by XMLmind

XSL-FO Converter.

DatagramConnection: represents a Datagram end point

Usually, when well get a connexion we wont directly use these interfaces except particular cases. We prefer to
use other interfaces that use (or implement) those interfaces.
1.7.3.2. Retrieve the connexion
To retrieve the connexion, you will have to use the Connection class that provides the static method open() that
permits to open a stream with the device or the server to which we want to connect. We pass in the parameters
of the open() method the string corresponding to the URL of the demanded resource.
The open() method returns a Connection. We have to cast this connection to get the exact type of the connection
wished. These are the main types of connection you are to use:

HttpConnection: this interface provides the methods and the constants appropriated for a http connection. http
is a request/response protocole and we just have to provide parameters to the request. The response is
formated and corresponds to a content type (html, simple texte, images, xml, wap)

HttpsConnection: this interface extends the HttpConnection interface and proveds the methods and constants
appropriated to a secure connection: http by TLS, SSL v3, WTLS, or WAP TLS

SecureConnection: this interface defines a secure connection to a socket by the same protocoles that
HttpsConnection except WTLS

DatagramConnection: this interface permits us to to send Datagram packets through the network
1.7.3.3. Retrieve the information
Once the connection is retrieved, you can read and write data into the data streams according to the use you
want to make. First of all, you will have to open the input ans output streams thanks to the openInputStream()
ans openOutputStream() methods.
This is an example:
InputStream is = null;
OutputStream os = null;
HttpConnection c = null;
String url = "http://www.myserver.com";
try {
c = (HttpConnection)Connector.open(url);
os = c.openOutputStream();
is = c.openInputStream();
} catch (Exception e) {
throw new IllegalArgumentException("Not an HTTP URL");
} finally {
...
}

1.7.4. The specificities of an HTTP connection

HTTP is a protocole that can be both implemented using the IP (TCP/IP, UDP/IP) protocoles and the nonIP
protocoles such as WAP. So we will use a gateway.
An HTTP connection is reprensented by the HttpConnection interface. These are the main methods to best use
this interface:

getResponseCode(): the response code to the request (e.g.: 200 for OK)
119
Created by XMLmind

XSL-FO Converter.

setRequestMethod(String method): defines the method to pass parameters to the request (HEAD, GET or
POST)

getType(): returns the content type retrieved by the server

openInputStream(): retrives the input stream of the data

openOutputStream(): retrives the output stream of the data

setRequestProperty(String key, String value): defines the properties of the request

Consult the MIDP documentation for more information.

1.7.5. Conclusion
Despite the reduced capabilities of the mobile devices for the network connection, the MIDP profile provides
classes and interfaces sufficent to connect and establish robust and secure applications for theses devices.
Concerning the bluetooth, an optional API is available but it wont be handled in this course.

120
Created by XMLmind

XSL-FO Converter.

Chapter 6. RMI
By completing this course you will :

Understand the RMI architecture

Know the java.rmi package

Build distributed applications with RMI

Make safe its application

1. Course
1.1. Introduction
Remote Method Invocation(RMI) allows an easy creation and use of distributed java applications.

1.1.1. Distributed Application

A distributed application is an application with classes shared on several machines. In such applications, we can
invoke remote methods. So you can use one object method which is not on the local machine.
With the C programming language, it was already possible to invoke remotely by using RPC (Remote procedure
Calls). RPC is data structure oriented, and isnt oriented object.
With RMI (Remote Method Invocation) you can use objects on a remote JVM (Java Virtual Machine). RMI is
going further than RPC because it allows us to send data of one object, but also its methods, thanks to object
serialization (see Input/Output lecture). Its also possible to load the classes byte-code dynamically.
A technology not related to java, call CORBA (Common Object Request Broker Architecture) exists and is a
shared object standard. CORBA has been made by OMG (Object Management Group), a consortium gathering
about 700 hundred companies including SUN. The biggest CORBA advantage is that it works with several
languages, but it is harder to set up.
RMI is a distributed object system easier than CORBA. We wont speak about CORBA in this lesson. But you
have to know that IIOP (Internet InterObject Protocol) gives the interoperability between RMI and CORBA.

1.1.2. RMI
RMIs goal is to create a distributed object integrating plainly the java language and local object model. RMI
extends the java security and robustness to the distributed application world. RMI allows the developer to create
distributed application easily because the syntax and semantic remains the same compared to the usual
applications.
RMI came up with the JDK version 1.1 and has been extended during the next versions.
With RMI, the methods of some objects (call remote objects) can be called from different JVM (distinct
memory location) of the one where the objects are. RMI manages the communication between the server and the
client with TCP/IP and this in a transparent way for the developer.
RMI uses mechanisms and protocols defined and standardized, as sockets and RMP (Remote Method Protocol).
So the developer doesnt have to care of lower layers issues specific for network technologies.

121
Created by XMLmind

XSL-FO Converter.

The RMI architecture defines the way the object have to act, how and when the exceptions can happen, how to
manage the memory and how the called methods give and get parameters.
RMI keeps the security inherent in the java environment thanks to the RMISecurityManager class and the DGC
(Distributed Garbage Collector).

1.2. RMI Architecture

1.2.1. Interfaces
The interface is the RMI central piece. The RMI architecture is based on an important rule: the definition of the
behavior and its execution are separated concepts.
RMI allows the code that defines the behavior and the code that implements this behavior to be separated and
run on separate JVM. The definition of a remote service is coded by using a java interface. The implementation
of this remote service is coded in a class.
Consequently, the understanding of RMI is in the fact that the interfaces define the behavior and the classes
define the implementation.
RMI support two types of class which implements the same interface. The first is the implementation of the
behavior and run on the server. The second acts like a proxy for the remote service and run on the client.
A client program creates some methods' calls and gives it to the proxy, RMI sends the query to the remote JVM
and transfers it to the implementation. All the return values given by the implementation are given back to the
proxy and then to the client program.

1.2.2. The RMI layers

RMI is based on the abstraction of 3 layers.
1.2.2.1. Stubs
The first layer is the one of the Stubs. This layer intercepts the methods' calls made by the client to the reference
interface and redirects these calls to a remote RMI service. This structure avoids the developer communicating
explicitly with the remote object.

JDK version 1.1 : Stub on the client, Skeleton on the server.

Version 1.2 : The stubs have the same role as the skeletons and the skeletons are no more required.

Version 1.5 : The use of stubs is transparent for the developer.

1.2.2.2. Remote Reference Layer
The next layer is the remote reference layer. With this layer you can interpret and manage the references made
from the client to the remote service. This layer allows to get a reference of the remote object from the local
reference (the Stub).
In the JDK 1.1, this layer connects the clients to the remote services, run it and exports it on a server. It is a
unicast connection(end to end).
In the JDK 2, this layer has been improved on activating the service objects in stand-by with the
RemoteObjectActivation class. It ensures a permanent reference to the remote objects (eventually with a
reconnection).
This service is given by a registry program which can be run dynamically or from the command rmiregistry.
1.2.2.3. Transport Layer
122
Created by XMLmind

XSL-FO Converter.

The transport layer is based on TCP/IP connections between machines. It gives the basic connectivity between
JVM.
Furthermore, this layer gives strategy to get through firewalls. The layer builds a table of all the remote objects
available and also listens and answers to the invocations.
This layer uses the Socket and SocketServer classes. It also uses a proprietary protocol RMP. (Remote Method
Protocol).
By using a layered architecture, each layer can be replaced without affecting the rest of the system. This is a
schema showing the information transfer between each layer:

1.2.3. RMI needs

We will see the different steps needed to set up properly a RMI service.
1.2.3.1. The interface
The first step is to create a remote interface describing the methods, which can be called by the client. Only
these methods will be remotely available.
To allow the client to have access to these methods, this interface must inherit from an interface named Remote,
in order to be identified as being callable from a remote JVM. Each remote object must implement this interface
directly or not and must be on both machines (server and client).
All the method remotly usable must be able to throw RemoteException exception specific for remote calls. So
you have to declare it in the methods throws declaration.
RemoteException is the commun upper class with a lot of exceptions, related with the communication, which
might happen during the call of a remote method.
1.2.3.2. The implementation
Now you have to implement this remote interface in a class. By convention, the name of this class should have
Impl as suffix.
In this class; we will define the body of remote methods usable by our clients. Of course, it is possible to add
other methods but the client wont be able to access them. The implementation class can implement several
remotes interfaces and extend another remote implementation class.
Our class must extend the java.rmi.server.RemoteObject class or one of its subclass. The easiest way is to use
java.rmi.server.UnicastRemoteObject.
If it doesnt, the class must assume the responsibility of the hashCode(), equals() and toString() methods. Its
easier to use the UnicastRemoteObject class because you dont have to define these methods yourself.
123
Created by XMLmind

XSL-FO Converter.

The UnicastRemoteObject class defines a not duplicated object, with valid references only if the server process
is active. It brings a support for the references of active objects using TCP/IP stream.
1.2.3.3. Stubs and Skeletons
When our client will call a remote method, this call will be transferred to the stub. Its the local representation of
the remote object. It marshalls the method argument of the remote objects and sends it in the data stream. It also
unmarshalls the return value given by the remote method.
The stubs are the intermediate between the client and the server, and manage the data transfers.

1.2.3.4. The RMI registry

The clients can get a service by using a directory service on the host with a port number. RMI can use several
directory services, Java Naming and Directory Interface (JNDI) included.
The registry is executed on each machine hosting remote objects (the servers) and accepting the request for
these services, on port 1099 by default.
A server creates a remote service by creating a local object implementing the service we want to create. Then, it
exports this object to RMI. When the object is exported, RMI creates a listening service waiting for a client
connection sending a request. After the exportation, the server adds the object in the RMI registry with a public
name in order to get a reference to this object.
Then the client can consult the registry to get a reference to reference to the remote object.
With the Registry class, you can manage the registry. It gives you methods to store and get objects on it, with a
name defined for each object, in order to have a unique name for each one.
The methods bind(), unbind() and rebind() are used to manage the references in the registry and are often used
on the server. The methods list() and lookup() are used to get references to the object stored in the registry, and
are used on the client.
124
Created by XMLmind

XSL-FO Converter.

The LocateRegistry class is used to get a reference on the registry itself or to create it on a host with a specified
port.
On the server, you have to call the static method createRegistry() which creates a registry, with the port number
you want to use. This method creates and returns a Registry reference allowing us to handle the objects on the
registry.
In the other hand, on the client, you have to get a Registry reference, thanks to the getRegistry() method. This
method takes the address of the server as parameters and optionally its port number (1099 by default)
The function details given by the Registry class is written below in the server and client subsection. All the
methods are static and can throw RemoteException exceptions in case of trouble with the communication.
1.2.3.5. The server
Our server has to set in the RMI registry the local object with the methods remotely callable. Here are some
methods often used on the server.
bind()
public static void bind(String name, Remote obj) throws AlreadyBoundException, MalformedURLException,
RemoteException

With this method you can add an object in the registry by giving a name. It takes as parameters a String which is
the URL of the object and an object of the Remote type. If there is already a link to an object with this URL,
an exception AlreadyBoundException is thrown.
rebind()
public static void rebind(String name, Remote obj) throws RemoteException, MalformedURLException

The way to use this method is the same as the previous one, but with this one, you can link the name with an
object and if this name was already bound to one object, itll be replaced. We would prefer this method instead
of the bind() method to avoid the AlreadyBoundException to be thrown.
unbind()
public static void unbind(String name) throws RemoteException, NotBoundException, MalformedURLException

This method removes the link of the object referenced with the name given as parameters in the RMI registry. If
the link doesnt exist, an exception NotBoundException will be thrown.
1.2.3.6. The client
The client can get a reference to the remote object by using the Registry class with the static method lookup() .
Then you can invoke the remote methods on this object.
The lookup() method is used by the client to call the registry and get a remote object. It takes as parameters a
URL specifying the server name and the name of the service required . It returns a reference of the remote
object with a Remote type. So you have to cast this object with the interface implemented by the remote object.
In fact, the client works with the remote interface and not with the remote object.
Then there are two useful Registry methods on the client.
lookup()
public static Remote lookup(String name) throws NotBoundException, MalformedURLException, RemoteException

The lookup() method is used by the client to get a reference of a remote object. The objects return is of type
Remote. This method takes as parameter the remote object URL with a String type. If the link in the registry
doesnt exist, an exception NotBoundException will be thrown.
list()
public static String[] list() throws RemoteException, MalformedURLException

125
Created by XMLmind

XSL-FO Converter.

This method returns a String array with every names previously stored in the registry.
1.2.3.7. Class distribution
With the JDK version 1.5, client and server have only their own classes and remote interfaces byte-code used. If
a class is required, it will be automatically downloaded (Part 4 : Dynamic Stub class loadind).
1.2.3.8. Summary schema

1.3. Example: Date server

Now, we will see an example by following the steps described in the previous chapter. The application goal is to
use the remote invocation with a date server. The server will give its date and time to any client.

1.3.1. DateServer interface

The only one method required in our application is a method returning an object with a java.util.Date type. So
we build our interface as :
package com.supinfo.labosun.common;
import java.rmi.Remote;
import java.rmi.RemoteException;
import java.util.Date;
public interface DateServer extends Remote {
Date getDate() throws RemoteException;
}

You have to remember that our interface extends the Remote interface in order to give the remote access for the
class implementing these methods. Furthermore, our getDate() method can throw RemoteException exception. If
we had declare some other methods, they would also have to be included this clause to be accessible by the
clients. All the returned objects by a remote method must implement the interface java.io.Serializable in order
to be sent on the network.
126
Created by XMLmind

XSL-FO Converter.

A remote method can return three kind of value actually:

A primitive type (boolean, int, float, )

An object implementing the Serializable interface

A remote object
Actually, the way the parameters are given to a remote object depends of its type:

The primitive types are given by value.

The basics objects are also given by value. That means that when an object is given to a remote method, RMI
sends the object itself, not its references. In the same way, when a remote method returns an object, a copy of
the object itself is sent.
RMI uses a technology call Object Serialization to transform an object in the linear format which can be sent by
the network. The serialize objects can be unmarshalled in the remote JVM memory and used by a java program.
The objects given as parameters must be serializable (implementing the java.io.Serializable interface)
For the remote objects, a reference is given by the remote Skeleton (java.rmi.server.RemoteRef). The user's code
gets back a reference to the stub class implementing the initial interface.

1.3.2. DateServerImpl implementation

The DateServerImpl class implements our DateServer interface previously declared. It extends the
UnicastRemoteObject class, which is a sub-class of RemoteObject. This class makes the link with the RMI
system. You have to write a constructor which throws RemoteException exception.
package com.supinfo.labosun.server;
import java.rmi.RemoteException;
import java.rmi.server.UnicastRemoteObject;
import java.util.Date;
public class DateServerImpl extends UnicastRemoteObject implements DateServer {
public DateServerImpl() throws RemoteException{}
public Date getDate() throws RemoteException {
return new Date();
}
}

Its in this class that we have to declare the getDate() method which is returning an object with the
java.util.Date type. You can return this object because the Date Object is serializable

1.3.3. MyServer
The Server class will create an object of DateServerImpl type. Then this object will be added in the registry with
the name: MyServer, by using the rebind() method of the Registry class.
package com.supinfo.labosun.server;
import java.rmi.Registry;
import java.rmi.RemoteException;

public class MyServer {

public static void main(String[] args) {

127
Created by XMLmind

XSL-FO Converter.

try {
DateServerImpl ds = new DateServerImpl();
Registry myRegistry = LocateRegistry.CreateRegistry(); myRegistry.rebind("myServerObject", ds);
} catch (RemoteException re) {
re.printStackTrace();
} catch (Exception e) {
e.printStackTrace();
}
}
}

You can also use the bind() method, but the rebind() method is better because it can add the object in the
registry again, even if the name already exists in the registry.

1.3.4. MyClient
The client will display on the console the date sent by the server. So we have to create an object with the
DateServer (our interface) type by calling the object with the name MyServer in the registry. Then we call the
remote method getDate().
package com.supinfo.labosun.client;
import java.rmi.Registry;
import java.rmi.RemoteException;

public class MyClient {

public static void main(String[] args) throws RemoteException {
new MyClient(args[0]);
}
public MyClient(String host) {
try {
Registry reg = LocateRegistry.LocateRegistry("localhost");
DateServer serverObjet= (DateServer) reg.lookup("myServerObject");
System.out.println("Server date: " + serverObjet.getDate());
}
catch (Exception e) {
e.printStackTrace();
}
}
}

1.3.5. Starting
Now you can start the server, and then your clients, to check if everything is working properly.

1.4. Dynamic Stub class loading

The RMI framework includes the automatic distribution of stubs byte-code. In this case, the application (on the
server and on the client) needs only its own classes in each interface used. The others classes required will be
automatically downloaded from a web server (FTP or HTTP). This operation is executed by the
RMIClassLoader class. If a client or a server executes a RMI system and needs to load a remote class, it will
automatically call the RMIClassLoader to do the work. The load of the class by RMI is controlled with the
parameters java.rmi.server.codebase of the Java interpreter.
Here we have the syntax used to define the parameters at the execution of the interpreter :
java [ -D<PropertyName>=<PropertyValue> ] <ClassFile>

The propertie java.rmi.server.codebase sets the URL where we can find the files to download. If a programm
executed in a JVM send an object reference to another JVM (as return value of a method), the other JVM will
need to load the class file of this object. When RMI sends the object with the serialization, it adds the URL with
it. So RMI dont send the class files with the serialized objects.

128
Created by XMLmind

XSL-FO Converter.

If the remote JVM needs to load a .class file for an object, it looks for a URL and contact the server chosen for
the file. When the properties java.rmi.server.useCodebaseOnly is active, the JVM loads the classes in the place
specified in the CLASSPATH, or at the URL specified by this property.
By using several combinations of available properties, you can create several RMI configuration systems.
In the Server date example, we can imagine that you have place the DateServerImpl_Stub.class file (the stub) on
a web server. If your programms (client or server) don't have the files in their CLASSPATH, they will look for
them at the URL specified. To specify that, you have to type the followings commands :
java Djava.rmi.server.codebase=http://172.16.X.X/path_des_classes/ MyServer

If the stubs are on a ftp server.

java Djava.rmi.server.codebase=ftp://172.16.X.X/path/ MyServer

If the stubs are on a http server listening on the port 2001.

java Djava.rmi.server.codebase=http://172.16.X.X:2001/ MyServer

In these examples, we execute the server but the syntax is the same to execute the client.
You must write the last URL slash (/) in the URL.
The advantage of using a web server is to modify your remote objects without changing anything on the clients.
Usually the web service is started on the server with the RMI server.
The following code is an other way to define the java.rmi.server.codebase properties,but this time in your
application :
System.setProperty("java.rmi.server.codebase",
" http://172.16.X.X:2001/ MyServer");

Example : http://java.sun.com/developer/JDCTechTips/2001/tt0227.html#dynamic

1.5. Security strategy

In java, you can set properly a security to your application. We wont explain all the details on the subject
because itll be done in an another lecture. Here, we will study only the security to set up if you want to use
RMI.
On some systems, the dynamic loading of classes can be refused for security reasons. An
AccessControlException exception can be thrown, when you use the methods of the Registry class.
To avoid this trouble, you can set up a security strategy, authorizing that. Its possible with the security manager
specific to RMI by the RMISecurityManager object. It checks the classes definition and authorizes only the
transfer of remote methods parameters and return values.
To use the security manager in our code, you have to type:
if (System.getSecurityManager() == null){
System.setProperty("java.security.policy", "file.policy");
System.setSecurityManager(new RMISecurityManager());
}

First we check that there are no security manager already existing (in order to not overwrite it), and then we use
the RMISecurityManager object.
You have to use this code before calling the registry using the Registry class.
You can use this security manager on both side (server and client). It depends if you need to load dynamically
your classes.
Then you just have to define the permissions. It consists to set the permissions in a specific text file and this file
will be read when youll create the RMISecurityManager. By convention, this file have the .policy extension.
129
Created by XMLmind

XSL-FO Converter.

Format:
grant signedBy "signer_names", codeBase "URL",
{
permission permission_class_name "target_name", "action",
signedBy "signer_names";
permission permission_class_name "target_name", "action",
signedBy "signer_names";
...
};

The signedBy and codebase values are optional, and the order of these fields does not matter. A signedBy value
indicates the alias for a certificate, it's used to verify the digital signature on the code. An example is
"Duke,James,Oliver". A codeBase value indicates the code source location, an empty codeBase entry signifies
"any code"; it doesn't matter where the code comes from. For exemple : "file:/C:/somepath/api/"
Here we have examples of this file:
grant {
permission java.security.Allpermission
};

In this example, everything is allowed.

Its the easiest permission but also the most dangerous.
grant {
permission java.net.SocketPermission "172.16.23.32:1099", "connect";
};

We accept all the connection on the host with the IP 172.16.23.32 on the port number 1099.
grant codebase "http://www.labo-sun.com" signedBy "Duke",{
permission java.io.FilePermission "${PROJECT_DIR}\\*", "read, write";
};

This allows code downloaded from "www.games.com", signed by "Duke", permission to read and write into the
project directory..
Permissions list : http://java.sun.com/j2se/1.3/docs/guide/security/permissions.html#PropertyPermission

1.6. Distributed Garbage Collector

RMI gives a distributed garbage collector. Its advantages are the same as the local garbage collector.
It deletes remotes objects no more referenced by the clients and is activated when no more client have stubs of
remotes objects.
Its based on a references counter of the stubs and the interaction of each Garbage collector (server and client).
When a garbage collector deletes a stub, the server garbage collector counter is decremented. If the counter is
equals to 0, the garbage collector deletes the object.
Its possible to inform the object that they are no more references.
In order to do that, they need to implement the Unreferenced interface and must implement the unreferenced()
method. This one will be call before the object's destruction. Its the same principle as the finalize() method for a
standard object.
import java.rmi.*;
import java rmi.server.*;
public class RemoteObjectImpl extends UnicastRemoteObjet implements ObjetDistant, Unreferenced {
public RemoteObjectImpl() throws RemoteException {
}

130
Created by XMLmind

XSL-FO Converter.

public void unreferenced() {

System.out.println("The object is no more referenced.");
}
public void remoteMethod () {
}
}

131
Created by XMLmind

XSL-FO Converter.