Notes and Updates:

• If you get an error on the import line for "edu.umd.cs.findbugs.annotations.SuppressWarnings" (e.g. in WebServer.java), it is likely that your "annotations.jar" file (located in your project's lib/ folder) is corrupted. Download a fresh copy here and replace the one in your project. You may need to update the project build path:
  

  Right-Click -> Project Properties -> Java Build Path -> Libraries Tab

The API documentation contains most of the details for this project. Here you will find a general overview.

For this project, you will write a simple web server in Java. Your web sever will execute Servlets that create web pages on-the-fly. When users connect to your web server, they will indicate which Servlet(s) to run as part of the URL. Your web server will create and invoke those Servlet(s) and produce a response.

There are several tasks to complete for this project, all described below in this writeup:

1. Write a simple (web) server to accept requests on a port and process those requests
2. Write a Servlet factory class to "manage" Servlets
3. Create a GoogleQuery Servlet
4. Create a Browse Servlet
5. Create a Cache Servlet decorator

WebServer

For this project, you will write a web server that listens on a particular port and produces a response. Your web server will be invoked from the command line. Because we may have multiple students running on the same machine, we can't have everyone using port 80 for their web server. So to avoid any conflicts, your web server should listen on port 33xxx, where xxx is the last three digits of your class account number. For example, if your account is cs433014, your server should listen on port 33014. Your WebServer should also accept an override port as a command-line argument, e.g.

The WebServer class that we have provided is currently implemented to echo back the received HTTP command. You will have to change this to invoke Servlets instead, as described more below. Please refer to the API documentation for the WebServer's runWebServer method for details on the HTTP request format. As an example, given the following HTTP request

  GET /MyServlet/A/Path?Option1+Option2 HTTP/1.0

  /MyServlet/A/Path?Option1+Option2

One subtlety is that the option argument to a Servlet's doGet method must be decoded using the java.net.URLDecoder.decode method found in the Java API. Further details can be found in the API documentation for the Servlet interface.

ServletFactory

After parsing the HTTP request, your web server will obtain an instance of the appropriate Servlet from the ServletFactory class -- specifically by invoking the getServlet method with the servlet name as an argument. The ServletFactory should implement the Singleton design pattern. In particular, the constructor of the factory should be private, and the static method getInstance should be invoked to return an instance of of ServletFactory. Every time this method is called, it should return the same instance (i.e., it shouldn't create a new instance with each call).

The ServletFactory class manages Servlet creation so that every unique servlet name has one and only one instance ever created for it. So if a client requests the servlet HelloWorld twice, the first time ServletFactory will create a new instance, but the second time ServletFactory should return the same instance it created before. Be careful: Cache:Browse and Cache:Cache:Browse are two different servlets.

Servlets

You must implement three new Servlets:

• GoogleQuery - Converts the provided options into an HTTP redirect that sends a query to Google. Typical web browsers, upon receiving a redirect, will contact the web site to which they have been redirected, in this case Google with the desired query.
• Browse - Takes a path to a file (residing on the server) as the argument, reads the corresponding file, and prints it to the output stream with the appropriate Content-Type. For security reasons, the Print servlet should reject requests for files in directories other than the "web" directory.
• Cache - A Servlet which employs the Decorator pattern to cache results from another "nested" Servlet and reuse the cached output on subsequent identical requests. This Servlet is a little bit tricky, and there's more than way to do it, so think before you start coding.

  As an example, the following URL

    /Cache:Browse/index.html
  

  would invoke Browse and output the contents of index.html. This output would be stored in some data structure (of your own design.)

  A subsequent request of the same form

    /Cache:Browse/index.html
  

  would also return the same output. However, this time, Cache should not have delegated to the Browse servlet. Instead, it should have checked its internal data structure for previously cached output for that URL.

StudentTests

You must also write test cases to turn in as part of your project. These should be stored in the class StudentTests, whose skeleton we have provided. You should write tests that exercise the functionality not already covered by the PublicTests. The aim is for you to get used to writing formal test cases when you program.

To evaluate the quality of your test cases, we will use the Clover code coverage tool. The higher the coverage of your test cases, the better your grade. In particular, the points awarded on the submit server for Public and Release tests total to 96 points. The remaining 4 points are assigned based on code coverage and quality of the unit tests that you provide in StudentTests.

FindBugs