Calling C functions in java using JNI API

JNI Example:
The ability to write just one set of code in Java and have it run on every system with a Java run-time is one of Java's primary strengths. But this platform independence has one key drawback: What do we do with the vast amount of existing code? The trick is to use the so-called native method interface.

Writing native methods involves importing C code into your Java application. In this tip I'll walk you through the basic recipe for creating native methods and using them in a Java application.

Seven steps to native method nirvana The steps to creating native methods are as follows:

• Write Java code
• Compile Java code
• Create C header (.h file)
• Create C stubs file
• Write C code
• Create shared code library (or DLL)
• Run application
  

Our exercise is to write some text to the console from inside the native method. The specifics of this example will be geared toward a Unix-like system, specifically Linux. I'll point out the couple of spots where the details differ for other platforms.

Write Java code

Write your Java code as you normally would. To use native methods in your Java code, you must do two things. First, write a native method declaration for each native method that you want to use. This is just like writing the declaration of a normal Java method interface, but you must specify the native keyword, as follows:

public native void printText ();

The second hoop to jump through is you must explicitly load the native code library. (We will create this later.) We do this by loading the library in a class static block:

static

    {

    System.loadLibrary ("happy");

    }

To put these pieces together for our example, create a file called Happy.java with the following contents:

class Happy

    {

    public native void printText ();

    static

   {

   System.loadLibrary ("happy");   /* Note lowercase of classname! */

   }

    public static void main (String[] args)

   {

   Happy happy = new Happy ();

   happy.printText ();

   }

 }

Compile Java code

Compile the Happy.java file:

% javac Happy.java

Create a C header file

There are various magic incantations that must be made available so that our C code can be used as a native method. The javah functionality of the Java compiler will generate the necessary declarations and such from our Happy class. This will create a Happy.h file for us to include in our C code:

% javah Happy

Create a C stubs file

In a manner reminiscent of the mangling that C++ translators do to the names of C++ methods, the Java compiler has a similar madness. To ease the pain of having to write a lot of tedious code so that our C code can be invoked from the Java run-time system, the Java compiler can generate the necessary trampoline code automatically for us:

% javah -stubs Happy

Write C code

Now, let's write the actual code to print out our greeting. By convention we put this code in a file named after our Java class with the string "Imp" appended to it. This results in HappyImp.c. Place the following into HappyImp.c:

#include <StubPreamble.h>      /* Standard native method stuff. */

#include "Happy.h"         /* Generated earlier. */

#include <stdio.h>         /* Standard C IO stuff. */

void Happy_printText (struct HHappy *this)

    {

    puts ("Happy New Year!!!");

    } 

In interfacing your C code with Java, many other aspects are involved -- such as how to pass and return the myriad types. For more information, see the Java tutorial or the Hermetica Native Methods Paper (see the Resources section for URLs).

Create a shared library

This section is the most system-dependent. It seems like every platform and each compiler/linker combination has a different method of creating and using shared libraries. For folks using any of the various Microsoft Windows platforms, check the documentation for your C compiler for the nitty-gritty details.

For you Linux folks, here's how to create a shared library using GCC. First, compile the C source files that we have already created. You have to tell the compiler where to find the Java native method support files, but the main trick here is that you have to explicitly tell the compiler to produce Position Independent Code:

% gcc -I/usr/local/java/include -I/usr/local/java/include/genunix -fPIC -c Happy.c HappyImp.c

Now, create a shared library out of the resulting object (.o) files with the following magical incantation:

% gcc -shared -Wl,-soname,libhappy.so.1 -o libhappy.so.1.0 Happy.o HappyImp.o

Copy the shared library file to the standard short name:

% cp libhappy.so.1.0 libhappy.so

Finally, you may need to tell your dynamic linker where to find this new shared library file. Using the bash shell:

% export LD_LIBRARY_PATH=`pwd`:$LD_LIBRARY_PATH

Execute the application

Run the Java application as usual:

% java Happy

Configuring Datasource in Jboss using MySQL and Oracle

MySQL Datasource Configuration in Jboss App Server:

MySQL is an open source database used by many open source projects and small organizations. To use JBoss 4.0 with MySQL, we first need to put the MySQL driver classes into the CLASSPATH. Copy the .jar file mysql-connector-java-3.0.9-stable-bin.jar to the /server/default/lib directory.
To use the MySQL data source, copy /docs/examples/jca/mysql-ds.xml to the /server/default/deploy directory. Modify the mysql-ds.xml configuration file by setting to com.mysql.jdbc.Driver and to jdbc:mysql:///, where is the MySQL host server and is the MySQL database.
Next, we need to set the and elements in the standardjaws.xml or jaws.xml file:

java:/MySqlDS
mySQL

We also need to set the and elements in the standardjbosscmp-jdbc.xml or jbosscmp-jdbc.xml file:

java:/MySqlDS
mySQL

Finally, we modify login-config.xml with MySQL database settings. Add the following element to login-config.xml:

sa
sa

jboss.jca:service=LocalTxCM,name=MySqlDS

By modifying the mysql-ds.xml, standardjaws.xml, standardjbosscmp-jdbc.xml, and login-config.xml files, the JBoss 4.0 server is configured to be used with a MySQL database.

In Client:

Hashtable ht=new Hashtable();
ht.put(InitialContext.INITIAL_CONTEXT_FACTORY,"org.jnp.interfaces.NamingContextFactory");
ht.put(InitialContext.PROVIDER_URL,"jnp://localhost:1099");
ht.put(InitialContext.URL_PKG_PREFIXES,"org.jboss.naming:org.jnp.interfaces");
initialContext = new InitialContext(ht);
javax.sql.DataSource ds = (javax.sql.DataSource) initialContext.lookup("java:MySqlDS");
java.sql.Connection conn = ds.getConnection();
// do the necessary database operations on connection.



Troubleshooting:
---------------------

javax.naming.NamingException: MySql not bound.

Then make sure that you have following entry in mySql-ds.xml:

false


Similarly for configuring oracle datasource in jboss:
----------------------------------------------------------------

Oracle Configuration

Oracle is a very popular enterprise database used for its performance and reliability. To configure JBoss 4.0 with Oracle, we first need to put Oracle's driver classes in the CLASSPATH. Copy Oracle's JDBC driver .zip file /jdbc/lib/classes12.zip to the server/default/lib directory.
To use Oracle's transactional (XA) data source, copy /docs/examples/jca/oracle-xa-ds.xml to the /server/default/deploy directory. To configure with the non-XA data source, copy /docs/examples/jca/oracle-ds.xml instead, to /server/default/deploy dir.
Next, we need to modify the oracle-ds.xml configuration file. The and settings for Oracle are as follows:

Oracle OCI Type 2 Driver
Class: oracle.jdbc.driver.OracleDriver
URL: jdbc:oracle:oci8:@
Oracle OCI Thin Type 4 Driver
Class: oracle.jdbc.driver.OracleDriver
URL: jdbc:oracle:thin:@::
Oracle OCI XA Type 2 Driver
Class: oracle.jdbc.xa.client.OracleXADataSource
URL: jdbc:oracle:thin:@::
Oracle OCI Type 2 Driver
Class: oracle.jdbc.driver.OracleDriver
URL: jdbc:oracle:oci8:@

In the Connection URL setting, is the HOST value specified in the /network/ADMIN/tnsnames.ora file, and is the PORT value specified in the tnsnames.ora file, and is the database name.
Next, we modify the standardjaws.xml or jaws.xml configuration file. Set the and elements as follows:


java:/OracleDS
Oracle8


Next, we modify the standardjbosscmp-jdbc.xml or jbosscmp-jdbc.xml configuration file, setting the and elements to use Oracle:



java:/OracleDS
Oracle8



Finally, we need to modify login-config.xml to use Oracle. Add the following element to login-config.xml:




sa
sa


jboss.jca:service=LocalTxCM,name=OracleDS





By modifying the oracle-ds.xml, standardjaws.xml, standardjbosscmp-jdbc.xml, and login-config.xml files, the JBoss 4.0 server is configured to be used with a Oracle database.

Integrating Struts and Spring framework

Integrating Struts and Spring:

Like Struts, Spring can also function as an MVC implementation. Both frameworks have their merits and drawbacks, although most would agree that Struts is still king when it comes to MVC. Many development teams have learned to rely on Struts as the foundation for building quality software under strict deadlines. With so much momentum behind Struts, even development teams that would like to integrate features of the Spring framework don't want to switch to Spring MVC. The good news is that you don't have to. The Spring architecture allows you to connect Struts as your Web framework to Spring-based business and persistence layers. The end result is that you can have your cake and eat it too!
In the recipes that follow, you'll learn three ways to integrate Struts MVC into the Spring framework. I'll expose the cons of each recipe as well as its comparative advantages. Once you've seen all three in action, I'll show you an exciting application of the approach I like best.


Three little recipes
Each of the following integration techniques (or recipes) has its merits, as well as its own particular quirks. I'm partial to only one of them, but knowing them all will deepen your understanding of both Struts and Spring. It will also provide you with a broad range of options for dealing with various scenarios. The recipes are as follows:
1. Use Spring's ActionSupport class to integrate Struts
2. Override the Struts RequestProcessor with Spring's DelegatingRequestProcessor
3. Delegate Struts Action management to the Spring framework

Loading the application context:
No matter which technique you use, you will need to use the Spring ContextLoaderPlugin to load the Spring application context for the Struts ActionServlet. Simply add the plug-in to your struts-config.xml file as you would any other plug-in, as shown here:

As previously mentioned, you'll find the complete source for the three fully functional example applications in the Download section. Each example presents a different approach to combining Struts and Spring for a book-search application. You can follow the basics of the examples here, but download the applications to see all the nitty-gritty details!

Recipe 1. Use Spring's ActionSupport:

Creating a Spring context manually is the most intuitive way to integrate Struts with Spring. To make it even easier, Spring offers a little help. The org.springframework.web.struts.ActionSupport class provides a getWebApplicationContext() method to easily obtain a Spring context. All you need to do is extend your action from Spring's ActionSupport instead of the Struts Action class, as shown in Listing 1:Listing 1. Using ActionSupport to integrate Struts

package ca.nexcel.books.actions;
import java.io.IOException;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import org.apache.struts.action.ActionError;
import org.apache.struts.action.ActionErrors;
import org.apache.struts.action.ActionForm;
import org.apache.struts.action.ActionForward;
import org.apache.struts.action.ActionMapping;
import org.apache.struts.action.DynaActionForm;
import org.springframework.context.ApplicationContext;
import org.springframework.web.struts.ActionSupport;
import ca.nexcel.books.beans.Book;
import ca.nexcel.books.business.BookService;
public class SearchSubmit extends ActionSupport { (1)
public ActionForward execute(
ActionMapping mapping,
ActionForm form,
HttpServletRequest request,
HttpServletResponse response)
throws IOException, ServletException {
DynaActionForm searchForm = (DynaActionForm) form;
String isbn = (String) searchForm.get("isbn");

//the old fashion way
//BookService bookService = new BookServiceImpl();

ApplicationContext ctx =
getWebApplicationContext(); (2)
BookService bookService =
(BookService) ctx.getBean("bookService"); (3)

Book book = bookService.read(isbn.trim());
if (null == book) {
ActionErrors errors = new ActionErrors();
errors.add(ActionErrors.GLOBAL_ERROR,new ActionError
("message.notfound"));
saveErrors(request, errors);
return mapping.findForward("failure") ;
}
request.setAttribute("book", book);
return mapping.findForward("success");
}
}
Let's quickly consider what's happening here. At (1), I create an Action by extending from the Spring ActionSupport class rather than the Struts Action class. At (2), I use the getWebApplicationContext() method to obtain an ApplicationContext. To obtain the business service, I use the context obtained at (2) to look up a Spring bean at (3).
This technique is simple and easy to understand. Unfortunately, it couples the Struts action to the Spring framework. If you ever decide to replace Spring, you would have to rewrite the code. Moreover, because the Struts action isn't under Spring's control, it can't reap the benefits of Spring AOP. This technique may be useful when using multiple independent Spring contexts, but for the most part it's not as desirable a solution as the other two choices.

Recipe 2. Override the RequestProcessor:

Decoupling Spring from the Struts action is a much smarter design choice. One way to do this is to override the Struts RequestProcessor processor with the org.springframework.web.struts.DelegatingRequestProcessor class, as shown in Listing 2:Listing 2. Integration via Spring's DelegatingRequestProcessor

(1)

Here, I've used the tag to override the default Struts RequestProcessor with the DelegatingRequestProcessor. My next step is to register the action in my Spring config file, as shown in Listing 3:Listing 3. Registering an action in the Spring config file

(1)

Note that at (1), I've registered a bean using the name attribute to match the struts-config action mapping name. The SearchSubmit action exposes a JavaBean property, allowing Spring to populate the property at run time, as shown in Listing 4:Listing 4. A Struts action with a JavaBean property
package ca.nexcel.books.actions;
import java.io.IOException;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import org.apache.struts.action.Action;
import org.apache.struts.action.ActionError;
import org.apache.struts.action.ActionErrors;
import org.apache.struts.action.ActionForm;
import org.apache.struts.action.ActionForward;
import org.apache.struts.action.ActionMapping;
import org.apache.struts.action.DynaActionForm;
import ca.nexcel.books.beans.Book;
import ca.nexcel.books.business.BookService;
public class SearchSubmit extends Action {

private BookService bookService;
public BookService getBookService() {
return bookService;
}
public void setBookService(BookService bookService) { (1)
this.bookService = bookService;
}
public ActionForward execute(
ActionMapping mapping,
ActionForm form,
HttpServletRequest request,
HttpServletResponse response)
throws IOException, ServletException {
DynaActionForm searchForm = (DynaActionForm) form;
String isbn = (String) searchForm.get("isbn");

Book book = getBookService().read(isbn.trim()); (2)
if (null == book) {
ActionErrors errors = new ActionErrors();
errors.add(ActionErrors.GLOBAL_ERROR,new ActionError("message.notfound"));
saveErrors(request, errors);
return mapping.findForward("failure") ;
}
request.setAttribute("book", book);
return mapping.findForward("success");
}
}
In Listing 4, you can see how to build the Struts action. At (1), I create a JavaBean property. This property is automatically populated by the DelegatingRequestProcessor. This design protects the Struts action from knowing it's being managed by Spring while giving you all the benefits of Spring's action management framework. Because your Struts actions are oblivious to the existence of Spring, you can swap out Spring for some other inversion of control container without refactoring your Struts code.
While the DelegatingRequestProcessor approach is definitely better than the first one, it does have some problems. If you were using a different RequestProcessor, then you would need to integrate the Spring DelegatingRequestProcessor manually. The added code would become a maintenance hassle and would also reduce your application's flexibility going forward. Moreover, there has been some talk of replacing the Struts RequestProcessor with a chain of command. Such a change would negatively impact the longevity of this solution.

Recipe 3. Delegate action management to Spring:
A much better solution is to delegate Struts action management to the Spring framework. You can do this by registering a proxy in the struts-config action mapping. The proxy is responsible for looking up the Struts action in the Spring context. Because the action is under Spring's control, it populates the action's JavaBean properties and leaves the door open to applying features such as Spring's AOP interceptors.
In Listing 5, the Action class is the same as it was in Listing 4. However, the struts-config is a little different:Listing 5. The delegation method of Spring integration

Listing 5 is a typical struts-config.xml file, except for one small difference. Instead of declaring the action's class name, it registers the name of Spring's proxy class, as shown at (1). The DelegatingActionProxy class uses the action mapping name to look up the action in the Spring context. This is the context that was declared with ContextLoaderPlugIn.
Registering a Struts action as a Spring bean is very straightforward, as shown in Listing 6. I simply create a bean using the name of the action mapping using the tag's name attribute (in this case, "/searchSubmit"). The action's JavaBean properties are populated like any Spring bean:Listing 6. Register a Struts action in the Spring context.

The benefits of action delegation:
The action-delegation solution is the best of the three. The Struts action has no knowledge of Spring and could be used in non-Spring applications without changing a single line of code. It's not at the mercy of a change to the RequestProcessor, and it can take advantage of Spring's AOP features.
The benefits of action delegation don't stop there, either. Once you have your Struts action under Spring's control, you can leverage Spring to give them more pizzazz. For example, without Spring, all Struts actions must be threadsafe. If you set the tag's singleton attribute to "false," however, your application will have a newly minted action object on every request. You might not need this feature, but it's nice to know you have it in your back pocket. You can also take advantage of Spring's lifecycle methods. For example, the tag's init-method attribute is used to run a method when the Struts action is instantiated. Similarly, the destroy-method attribute executes a method just before the bean is removed from the container. These methods are a great way to manage expensive objects in much the same way as the Servlet lifecycle does.


sources can be found in the original article:

http://www.ibm.com/developerworks/java/library/j-sr2.html