I have a tomcat6 servlet that manages incoming HttpPosts this way:
public void doPost(HttpServletRequest request, HttpServletResponse response)
throws ServletException, IOException {
if (request.getParameter("cmd") != null) {
eventPool.addEvent(new CommandEvent(new String[] { request.getParameter("cmd"),
request.getParameter("json") }, response));
}
}
The request will now be processed. When this is done, I want to write the result to the requesting client this way:
protected void sendResponse(HttpServletResponse httpResponse, String content){
try {
httpResponse.getWriter().println(CMD + "#" + content);
httpResponse.getWriter().close();
} catch (IOException e) {
e.printStackTrace();
}
}
But it fails to flush and I get a NullPointerException because the HttpResponse was already closed.
How can I prevent the HttpResponse from flushing before I want it to?
You'll need to use Tomcat 7 (or any other container that supports Servlet 3.0 onwards) to use that style of programming. Look at the asynchronous request processing parts of the Servlet 3.0 specification.
Prior to Servlet 3.0, request/response processing is synchronous. i.e. you cannot 'park' a request/response pair and then handle them later in a different thread. Pretty much as soon as your doPost() method exits, Tomcat will recycle the request and response objects ready to use them to handle a new request.
Related
public void doGet(HttpServletRequest req, HttpServletResponse resp)
throws ServletException, IOException {
//some code here
}
public void doPost(HttpServletRequest req, HttpServletResponse resp)
throws ServletException, IOException {
//performTask(req, resp);
//some code here
}
private void insertRequestTemplate() {
HttpSession session = req.getSession();
responsePage = req.getParameter("ResponsePage");
ServletContext ctx = getServletConfig().getServletContext();
ctx.getRequestDispatcher(responsePage).forward(req,resp);
readMessage();
public void readMessage()
{
System.out.println("calling MessageTrigger_ABean");
MessageTrigger_ABean msg = new MessageTrigger_ABean();
msg.read();
}
msg.read() has the code to read messages from MQ. Inside insertRequestTemplate method, I am calling readMessage method after ctx.getRequestDispatcher(responsePage).forward(req,resp);is this the correct way of calling this?
But inside insertRequestTemplate method, the page is not getting forwarded to the next page untill readMessage() is executed because of which the page keeps on loading for a long time until message is read from MQ. Could you please help me on this.
Most examples I have seen of a servlet forwarding the request to another servlet have the dispatcher forward invocation at the end of the method. ie. there is no more code, other than closing braces at the end of the method.
I am guessing that the forwarding doesn't happen until the invoking method completes. So where you have your msg.read() will stop the insertRequestTemplate method from completing. This will more than likely be because the code inside msg.read is being performed synchronously. Leading to http timeouts on the http request.
How you solve this will depend on what you want to do with the messages you obtain from msg.read().
What should be done to manually stop Servlet as calling destroy doesn't help unless all threads exited from service.
Say, If I have n number of Servlets and I want to stop only one of them.
That behavior is very important when dealing with Servlets. Instances can be created after the multi-thread model and are thus not thread-safe.
The container does not allow a thread to invoke the service method after destroy has been called.
This gives you the means to close all resources that your Servlet is using (db, file, memory, etc).
#WebServlet
public class OncePerApplicationServlet extends HttpServlet {
private Connection connection;
#Override
protected void doGet(HttpServletRequest req, HttpServletResponse resp) throws ServletException, IOException {
if(req.getParameter("closeServlet").equals("true"))
this.destroy();
else
this.service(req, resp); // normal flow
}
// this method will never be called by the container after the destroy method has been invoked
#Override
protected void service(HttpServletRequest req, HttpServletResponse resp) throws ServletException, IOException {
// 1.
try {
connection = DriverManager.getConnection("someDbUrl");
Statement stm = connection.createStatement();
stm.execute("select * from someTable");
} catch (SQLException e) {
e.printStackTrace();
}
}
#Override
public void destroy() {
// the point is that when this method is called you should be able to
// clean up and close all resources, you can rest assured that there are no "loose"
// threads that need the connection-instance
try {
connection.close();
} catch (SQLException e) {
e.printStackTrace();
}
}
}
Here is a quote from the API-docs:
This interface defines methods to initialize a servlet, to service
requests, and to remove a servlet from the server. These are known as
life-cycle methods and are called in the following sequence:
The servlet is constructed, then initialized with the init method. Any
calls from clients to the service method are handled. The servlet is
taken out of service, then destroyed with the destroy method, then
garbage collected and finalized.
|
Link to the documentation
Good luck!
I am trying to implement HTML 5 SSE feature. For this I need to keep response stream open for long duration. Right now my implementation is as below.
#Override
protected void doGet(HttpServletRequest req, HttpServletResponse resp) throws ServletException, IOException {
resp.setContentType("text/event-stream");
resp.setCharacterEncoding("UTF-8");
PrintWriter pw = resp.getWriter();
Broadcaster.add(pw);
try {
Thread.sleep(1000*60*60);
} catch (InterruptedException ex) {
Logger.getLogger(TestServelet.class.getName()).log(Level.SEVERE, null, ex);
}
Broadcaster.remove(pw);
}
I the above code i am making thread to sleep for 1 hour, so that I can keep PrintWriter open for 1 hour. I know suspending thread is not good method.
I want to know is there any better why where I can keep PrintWriter open even after call to servlet is returned? Or is there any different approach to achieve this effect?
Here You will find sample application (HTML+Servlets) which will probably meet Your requirements.
You can also use WebSockets
I have a web-app developed with servlet & JSP. I configured my app to throw an IllegalArgumentException if I insert bad parameters.
Then I configured my web.xml file in this way:
<error-page>
<error-code>404</error-code>
<location>/error.jsp</location>
</error-page>
<error-page>
<exception-type>java.lang.Throwable</exception-type>
<location>/error.jsp</location>
</error-page>
When I rise a 404 error, then it works and calls error.jsp, but when I rise a java.lang.IllegalArgumentException, then it does not work and I've a blank page instead of error.jsp. Why?
The server is Glassfish, and logs show really IllegalArgumentException rised.
You should not catch and suppress it, but just let it go.
I.e. do not do:
#Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
try {
doSomethingWhichMayThrowException();
} catch (IllegalArgumentException e) {
e.printStackTrace(); // Or something else which totally suppresses the exception.
}
}
But rather just let it go:
#Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
doSomethingWhichMayThrowException();
}
Or, if you actually intented to catch it for logging or so (I'd rather use a filter for that, but ala), then rethrow it:
#Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
try {
doSomethingWhichMayThrowException();
} catch (IllegalArgumentException e) {
e.printStackTrace();
throw e;
}
}
Or, if it's not an runtime exception, then rethrow it wrapped in ServletException, it will be automatically unwrapped by the container:
#Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
try {
doSomethingWhichMayThrowException();
} catch (NotARuntimeException e) {
throw new ServletException(e);
}
}
See also:
How does server prioritize which type of web.xml error page to use?
Submitting form to Servlet which interacts with database results in blank page
Another (simplified) approach is not to declare multiple handlers for various <error-code> and <exception-type> situations but rather have one, sort of catch-all sink, e.g.
<error-page>
<location>/error-page.jsp</location>
</error-page>
Inside your error-page.jsp you can determine the cause, be it a return status code or an exception as described here: https://www.tutorialspoint.com/servlets/servlets-exception-handling.htm These constants are a part of the standard Servlet 3.0 API.
For instance a primitive error-page.jsp response handler placed into the root of your webapp can look like this:
Server encountered a situation
Status code: <%=(Integer) request.getAttribute(javax.servlet.RequestDispatcher.ERROR_STATUS_CODE)%>
<br>
Exception: <%=(Throwable) request.getAttribute(javax.servlet.RequestDispatcher.ERROR_EXCEPTION)%>
For security reasons I wouldn't recommend sending the exact exception type to the client; this is just an example of how to handle different types of errors and response statuses inside a JSP handler; a servlet can be used instead of JSP.
One common catch-all handler vs one per status code is certainly dependent on the situation and requirements.
I have today the same issue. (JavaEE 7 and Glassfish 4.0)
The problem seems that the framework check it as String instead with the Class.
String based check (the hypothesis)
When a Exception is twrown, e.getClass() is compared with <exception-type> as string.
So you can't use inheritance.
Note that nested classes must be pointed as '$' instead '.' (same as getClass() method).
Class based check
The framework create an instance of the class, and <exception-type> text refer to it, and the class.isInstance() is used to check.
This will need reflection and policy file could break it.
I hope that this response solves future issues.
I am just beginning with Servlets and managed to have some servlets that act as individual URLs for populating a database for some dummy testing. Something of the form:
public class Populate_ServletName extends HttpServlet {
public void doGet(HttpServletRequest req, HttpServletResponse resp) throws IOException {
resp.setContentType("text/plain");
//Insert records
//Print confirmation
}
}
I have about 6 such servlets which I want to execute in a sequence. I was thinking of using setLocation to set the next page to be redirected but was not sure if this is the right approach because the redirects should happen after the records have been inserted. Specifically, I am looking for something like this:
public class Populate_ALL extends HttpServlet {
public void doGet(HttpServletRequest req, HttpServletResponse resp) throws IOException {
resp.setContentType("text/plain");
//Call Populate_1
//Call Populate_2
//Call Populate_3
//...
}
}
Any suggestions?
Use RequestDispatcher#include() on an URL matching the url-pattern of the Servlet.
public class Populate_ALL extends HttpServlet {
public void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
response.setContentType("text/plain");
request.getRequestDispatcher("/populateServlet1").include(request, response);
request.getRequestDispatcher("/populateServlet2").include(request, response);
request.getRequestDispatcher("/populateServlet3").include(request, response);
//...
}
}
Note: if those servlets cannot be used independently, then this is the wrong approach and you should be using standalone Java classes for this which does not extend HttpServlet. In your specific case, I think the Builder Pattern may be of interest.
The RequestDispatcher#forward() is not suitable here since it throws IllegalStateException when the response headers are already committed. This will be undoubtely the case when you pass the request/response through multiple servlets which each writes to the response.
The HttpServletResponse#sendRedirect() is absolutely not suitable here since it implicitly creates a brand new request and response, hereby trashing the original ones.
See also:
How do I call a second JSP servlet while in the first JSP servlet?
RequestDispatcher.forward() vs HttpServletResponse.sendRedirect()
communication between remote servlets
It looks like what you may need is a service that each of the servlets can use to perform some work. Then the servlets are not depending one and another, but rather all using the service.
However, here is an explanation of forwarding or redirecting requests.