ATG servlet and Droplet - java

I'm new for the ATG framework and I want to know different between Servlet and droplet in ATG. I read some tutorials , but still haven't clear idea.

I haven't worked with ATG in a few years, and you may be working with a different version than I did, but a simple way to answer this would be to say that Servlets are more of a J2EE MVC concept, the controller in that case, while Droplets are ATG's "controller". ATG's design patterns aren't exactly MVC (there tends to be quite a bit more logic in the view in most cases), but droplets are meant to be reusable components that the view (your JSPs) use to get data from the model (your xxxManager, xxxService, Repository layer classes, etc.)
For example, if I'm working on a page that is meant to display the shopping cart, I might register a "/Cart" servlet in my web.xml in J2EE. When a doGet is invoked on this servlet, I would call all of the model components I need to gather data about the cart, and use the RequestDispatcher to send that data to a JSP. All of my model/controller logic is done here, and then sent to be rendered by the view.
In ATG, you tend to rely less on web.xml configurations and more on components registered and instantiated by Nucleus. So you might create your cart.jsp and use tags like to grab the data you need ad hoc. You might have a droplet to calculate the current price and display it, or to get promotions applied to the profile and list them. Each piece of data you need for the page, you could use a droplet tag to gather and display.
You might say that droplets tend to be "on page" while servlets are meant to be "between pages."

When you say only Servlet, then that is a generic Java software component that provides a functionality based on a request it receives. A Java servlet needs to run in a servlet container.
Specific to ATG, there are two kinds of components that are termed as servlets - DynamoServlet and PipelineableServlet.
A DynamoServlet is the base class you need to extend to create a Droplet. A Droplet is basically a piece of Java code that you can run on a JSP page running within an ATG application. It is included on a JSP page using the tag that ATG provides. When a JSP executes, it will execute the code within the Droplet and present the output of the Droplet on the page.
A PipelineableServlet is a class (which is part of a chain of classes called a Servlet Pipeline) that provides a functionality when an HTTP request is received. Each servlet in the chain performs a specific operation on the request and relays the modified request to the next servlet in the chain.

Related

Single servlet to handle a single page web application design

I have a single page web application designed to look like a desktop application using plain Jsp/Servlet. In the back end I have one servlet with action param being passed for various actions occurs on the single web page. 90% of the requests are ajax. When the functionality grows on the single web page there will be more actions exist under single servlet. Now my questions are
having single servlet to manage a lot of operations a good design? What would be the better design?
What is the performance benefit when having single vs multiple servlet?
Will the servlet code be unmanageable at one point when it grows?
For company policy reasons I can't use the spring mvc..
having single servlet to manage a lot of operations a good design? What would be the better design?
No. A single class having basically all the responsibilities of your backend is not good design. Unless it only serves as a dispatcher to actions, but then you would reinvent the wheel that all the existing MVC frameworks have already invented for years (Spring MVC, etc.). Frankly, I would fight against the "company policy", and use the appropriate tool for the job: Spring MVC, or JAX-RS, or any other modern framework that can be used to easily implement a proper REST backend.
What is the performance benefit when having single vs multiple servlet?
None.
Will the servlet code be unmanageable at one point when it grows?
Yes. You'd better follow the REST principles and assign different URLs to your different actions instead of using a parameter to pass the action. Moreover, all actions should not use the same HTTP method. Searching or reading information should use GET, while creating stuff should use POST and updating stuff should use PUT.
And of course, use one servlet per URL.
you can use jsf
its a single servlet
called FacesServlet
--
jokes aside,
you could use a single servlet, and create multiple classes to handle different scenarios and keep your servlet class under 100 lines of code with good design
but this is essentially writing mvc yourself
REST is the right approach as suggested by JB Nizet. Use JAX-RS for handling the request. In case you don't want to take that route, you can use a single servlet but use the servlet like a controller that send the request to different modules to handle it. Single servlet is fine as long as you don't put all the code in the same class.
In case you use multiple servlets, the design would get better when you plan to move one of the servlets to a different server. But with multiple servlets, the UI (html or java script) needs to handle routing to the right servlet
Code will be unmanageble only if you write everything in one servlet class... If designed correctly it should be manageable.

Create a html response when user visits internal URL

When a user hits a particular internal URL I need to return some html that is created dynamically.
Some back end services will be called to genrated a list of keys which is then used to generated a list of html
href links which are displayed to the user.
For this I am considering using a servlet. Is this a good methodology ?
Since servlets have been around for some time maybe there are newer/better ways of implementing this ?
There are tons of solutions, most of them (in Java) being based on the servlet API. If you use Spring already, Spring has a module called Spring MVC, which is a framework based on the servlet API to create web applications, based on the MVC pattern:
Model holding the data from the database
View (implemented, most of the time, using JSPs) generating markup based on the data contained in the model
Controller getting the data from the database, and dispatching to the appropriate view
This is a good pattern used by most web frameworks, but every one has its own way of doing, strengths and weaknesses.
You could do with a simple servlet, but generating markup from a servlet is ugly. That's why servlets are usually used in combinations with JSPs, following the MVC pattern. You could implement a micro-MVC framework by yourself, using only servlets and JSPs, but Spring-MVC and other frameworks offer so many additional advanteges that the investment is worth it. I personally like Stripes very much, and it's very simple.
Servlets are the Java way of exposing simple interfaces for HTTP request.
This can also be achieved by a REST framework like Jersey, but its a bit more complicated, so if you need a simple one opertaion interface, I'd go with servlets.

Why in Struts1 there is Servlet where as in case of Struts2 there is Filter

The front-controller was a Servlet but in Struts2, it is a filter. What is the possible reason to change it to a filter?
(This is opinion, you'd need to ask the original WebWork authors.)
IMO it's a bit more intuitive to wrap requests inside a filter, since that's what filters were designed for.
There has been debate about the validity of serving resources from a filter. The spec states:
Filters do not generally create a response or respond to a request as servlets do,
rather they modify or adapt the requests for a resource, and modify or
adapt responses from a resource.
Some have claimed (particularly some WebSphere support tickets, and occasionally myself before re-reading the spec during an email thread on the Struts User mailing list) that the spec disallowed Struts 2's filter usage, but it's clear that nothing disallows their usage in this way.
Filters allow more flexibility in handling other types of requests (forwards, includes, and container errors) by using the <dispatch> element under <filter> configuration.
Note that originally it was a servlet in WebWork--you may be able to look at the commit logs to ferret out a reason whenever it was the change occurred, but it was quite awhile ago, on the order of 7+ years.
Because of the introducing Interceptors in struts2. It becomes necessary for Struts2 contributors to have main controllers from the front so that a user can not breach a threat within the Java EE pattern. However Struts2 dispatchers are built on the top of the hierarchy of the servlet but it reduces a lot effort from the security point of view.
Reasons:
Provide a centralized controller to a framework.
To have interceptors and context which can dance on your fingers.
There are three possibilities why filter is designated as front controller in Strtus2
Servlet made as front controller needs developer to provide a
right value in which lets the framework to initialize many important
aspects( viz. struts configuration file)as the container starts. In
absence of which the framework gets initialized only as the first
request hits.Struts2 makes our life easy by providing
front-controller as a filter and by nature the filters in web.xml
gets initialized automatically as the container starts. There is no
need of such load-on-startup tag.
The second but important one
is, the introduction of Interceptors in Struts2 framework. It not
just reduce our coding effort, but helps us write any code which we
would have used filters for coding and necessary change in the
web.xml as opposed to Struts1.So now any code that fits better in
Filter can now moved to interceptors (which is more controllable
than filters), all configuration can be controlled in struts.xml
file, no need to touch the web.xml file.
The front controller
being a filter also helps towards the new feature of Struts ie UI
Themes. All the static resources in the Themes now served through
the filter

Why is the servlet request not available in the portlet API

The portlet API deos not provide any reference to the enclosing servlet request and response objects. I know it is not the preferred model of interaction with the user, but it seems draconian to remove all access.
I understand that for portlet driven interaction with the user, you want to use portlet URLs, and let the portlet container manage all the complexity.
However if you have a number of portlets which are basically showing variants of the same data, it makes sense for them to be able to use the enclosing request to drive the data.
We ended up using using a Liferay specific call to get the request, and it all seems to work as we wish.
However I do feel the guilt.
So my question really is, is there an underlying deep reason for the prohibition, or is it just to enforce the authors view of the API environment?
The portlet doesn't run right into the Servlet container, but rather what is called a Portlet container.
You should be able to access to the corresponding information, PortletRequest, PortletResponse, and PortletContext.
The reason is that two instances of the same portlet can run next to each other in the same page, but still be isolated with their own lifecycle. The portal will "multiplex" that transparently for you, and it will convert from the servlet world to the portlet world. Portlet bridges are also available to develop portlets with non-portlet technologies (e.g. JSF). I agree that all this is usually (very) complicated to use (because of the many frameworks and implementations available), but when you think of how it works conceptually, it's quite nice.
The exact details will depend on the technologies you chose to develop the portlet. But I feel like there should be a way to do what you want using the portlet API.

When do you use a JSP and when a Servlet? [duplicate]

This question already has answers here:
What is the difference between JSF, Servlet and JSP?
(16 answers)
Closed 7 years ago.
I have an application that sends the customer to another site to handle the payments. The other site, outside of the customer, calls a page on our server to let us know what the status is of the payment. The called page checks the parameters that are given by the payment application and checks to see whether the transaction is known to us. It then updates the database to reflect the status. This is all done without any interaction with the customer.
I have personally chosen to implement this functionality as a JSP since it is easier to just drop a file in the file system than to compile and package the file and then to add an entry into a configuration file.
Considering the functionality of the page I would presume that a servlet would be the preferred option. The question(s) are:
Is my presumption correct?
Is there a real reason to use a servlet over a JSP?
What are those reasons?
A JSP is compiled to a servlet the first time it is run. That means that there's no real runtime difference between them.
However, most have a tradition to use servlets for controllers and JSPs for views. Since controllers are just java classes you can get full tool support (code completion etc.) from all IDEs. That gives better quality and faster development times compared to JSPs. Some more advanced IDE's (IntelliJ IDEA springs to mind) have great JSP support, rendering that argument obsolete.
If you're making your own framework or just making it with simple JSPs, then you should feel free to continue to use JSPs. There's no performance difference and if you feel JSPs are easier to write, then by all means continue.
JSPs: To present data to the user. No business logic should be here, and certainly no database access.
Servlets: To handle input from a form or specific URL. Usually people will use a library like Struts/Spring on top of Servlets to clear up the programming. Regardless the servlet should just validate the data that has come in, and then pass it onto a backend business layer implementation (which you can code test cases against). It should then put the resulting values on the request or session, and call a JSP to display them.
Model: A data model that holds your structured data that the website handles. The servlet may take the arguments, put them into the model and then call the business layer. The model can then interface with back-end DAOs (or Hibernate) to access the database.
Any non-trivial project should implement a MVC structure. It is, of course, overkill for trivial functionality. In your case I would implement a servlet that called a DAO to update the status, etc, or whatever is required.
JSPs should be used in the presentation layer, servlets for business logic and back-end (usually database layer) code.
I don't know any reason why you can't use a JSP as you describe (it gets compiled to a servlet by the containter anyway), but you're right, the preferred method is to make it a servlet in the first place.
There are 2 pretty simple rules:
Whenever you want to write Java code (business logic), do it in a Java class (so, Servlet).
Whenever you want to write HTML/CSS/JS code (view/template logic), do it in a JSP.
Related question:
How to avoid Java code in JSP
JSPs are a shortcut to write a servlet. In fact they are translated to servlet java code before compilation. (You can check it under some tomcat subdir wich I don't remember the name).
To choose between servlet an JSP I use a simple rule: if the page contains more html code than java code, go for JSP, otherwise just write a servlet. In general that translates roughly to: use JSPs for content presentation and servlets for control, validation, etc.
Also, Its easier to organize and structure your code inside a servlet, since it uses the plain java class syntax. JSPs tend to be more monolithic, although its possible to create methods inside then.
JSP's are essentially markup that automatically gets compiled to a servlet by the servlet container, so the compile step will happen in both instances. This is why a servlet container that supports JSP must have the full JDK available as opposed to only needing the JRE.
So the primary reason for JSP is to reduce the amount of code required to render a page. If you don't have to render a page, a servlet is better.
I know this isn't the popular answer today, but: When I'm designing an app from scratch, I always use JSPs. When the logic is non-trivial, I create ordinary Java classes to do the grunt work that I call from the JSP. I've never understood the argument that you should use servlets because, as pure Java classes, they are more maintainable. A JSP can easily call a pure Java class, and of course an ordinary Java class is just as maintainable as any servlet. It's easier to format a page in a JSP because you can put all the markup in-line instead of having to write a bunch of println's. But the biggest advantage of JSPs is that you can just drop them in a directory and they are directly accessible: you don't need to mess with setting up relationships between the URL and the class file. Security is easily handled by having every JSP begin with a security check, which can be a single call statement, so there's no need to put security into a dispatch layer.
The only reason I can see to use a servlet is if you need a complex mapping between URLs and the resulting execution class. Like, if you want to examine the URL and then call one of many classes depending on session state or some such. Personally I've never wanted to do this, and apps I've seen that do do it tend to be hard to maintain because before you can even begin to make a change you have to figure out what code is really being executed.
Most java applications nowadays are build on the MVC pattern...
In the controller side (servlet) you implement business logic. The servlet controller usually forward the request to a jsp that will generate the actual html response (the View in MVC).
The goal is to separate concerns... Thousands of books have been written on that subject.
In an MVC architecture, servlets are used as controller and JSPs as view.
But both are technically the same. JSP will be translated into servlet, either in compile time (like in JDeveloper) or when accessed for the first time (like in Tomcat).
So the real difference is in the ease of use. I'm pretty sure that you'll have a hard time rendering HTML page using servlet; but opposite to common sense, you'll actually find it pretty easy to code even a fairly complex logic all inside JSP (with the help of some prepared helper class maybe). PHP guys do this all the time. And so they fall into the pitfall of creating spaghetti codes.
So my solution to your problem: if you found it easier to code in JSP and it wouldn't involve too many code, feel free to code in JSP. Otherwise, use servlet.
Agreed with all the points above about the differences between JSPs and Servlets, but here are a couple additional considerations. You write:
I have an application that sends the
customer to another site to handle the
payments. The other site, outside of
the customer, calls a page on our
server to let us know what the status
is of the payment. The called page
checks the parameters that are given
by the payment application and checks
to see whether the transaction is
known to us. It then updates the
database to reflect the status. This
is all done without any interaction
with the customer.
Your application is consuming the payment service of another application. Your solution is fragile because if the payment service in the other application changes, that breaks your JSP page. Or if you want to change your application's payment policies, then your page will have to change. The short answer is that your application should be consuming the application's payment service via a web service. Neither a servlet nor a JSP page is appropriate place to put your consumption logic.
Second, along those lines, most usages of servlets/JSP pages in the last few years have been put inside the context of a framework like Spring or Struts. I would recommend Spring, as it offers you the full stack of what you need from the server pages to web service gateway logic to DAOs. If you want to understand the nuts and bolts of Spring, I would recommend Spring in Action. If you need to understand better how to tier an enterprise architecture written in a language like Java (or C#), I would recommend Fowler's Patterns of Enterprise Application Architecture.
Yeah, this should be a servlet. A JSP may be easier to develop, but a servlet will be easier to maintain. Just imagine having to fix some random bug in 6 months and trying to remember how it worked.
In java servlet the HTML tags are embeded in java coding.
In JSP the java codings are embeded in HTML tags.
For big application for big problem the servlet is complex to read,understand,debug,etc because of unreadability of embeding more html tags inside the java coding..So we use jsp.In jsp it is easy to understand,debug,etc.
Thanks & Regards,
Sivakumar.j
i think its up to you?
because
JSP is Java inside of HTML
and Servlet is a Java that can do the HTML inside
hmmm... servlet is more sercure than jsp because if you submit to Servlet and forward to another JSP there is no file extension appear and also you cant see what Page it is..
but the advantage of JSP is you can code there easily.

Categories