代写 html file system data structure junit Java CS 131a: Operating Systems Programming Assignment 1

CS 131a: Operating Systems Programming Assignment 1
Released: Tuesday, January 28th
Highly Encouraged Project Tutorial: Wednesday, January 29th 7pm, location Gzang 124
Due (hard deadline): Tuesday, Feb 4th (no late submissions are allowed)
Getting Started
Your assignment is going to be to build a basic command line environment for interacting with a file system. This command line environment (or shell) will be written in Java. The first part of this assignment is designed to refresh your coding ability in Java, and ensure that you have the coding pre-requisites to be able to complete this course. You will do this while learning about Unix and practicing your Java-based System calls.
Understanding UNIX shell
Essential to this assignment is understanding the fundamentals of a UNIX shell, and the functionality that is expected from a UNIX like operating system. The following pages can help you get a sufficient understanding of the expected behavior of UNIX to allow you to complete this project. After reading through these resources, it is highly recommended that you play around on a UNIX console before you attempt to construct your own:
Information
What you need to know
Link
What is Unix?
Theoretical
http://www.ee.surrey.ac.uk/Te aching/Unix/unixintro.html
Working Directory Commands
ls, cd, . , .. , pwd
http://www.ee.surrey.ac.uk/Te aching/Unix/unix1.html
Basic Text Operations
cat, head grep, wc, uniq
http://www.ee.surrey.ac.uk/Te aching/Unix/unix2.html
Redirection and Piping
Redirection, Piping, >|
http://www.ee.surrey.ac.uk/Te aching/Unix/unix3.html
Your assignment: Implementing the REPL Loop
Your main program will include a loop that prints a simple prompt (¡°>¡±), accepts a command, and prints the output from that command (if any). This is called a filter (e.g., head hello-world.txt is a filter). Your program should handle also piped filters, i.e., multiple commands separated by pipes (e.g., head hello-world.txt | grep world). This type of loop is sometimes called a Read-Eval-Print Loop, or REPL. The REPL loop should only exit upon user request. For example:
> Welcome to the Unix-ish command line. > head hello-world.txt
hello world!

> not-a-command
The command [not-a-command] was not recognized.
> head hello-world.txt | grep world
world!
> head hello-world.txt | grep world > output.txt
> head output.txt
world!
> ls
src
hello-world.txt
output.txt
> exit
Thank you for using the Unix-ish command line. Goodbye!
Filter Commands
Your program should correctly parse commands in a way that allows the creation of piped filters.YourREPLloopalsomustsupportsimpledirectorycommands:ls,cd,andpwd. These command are a bit different, as cd does not take piped input and does not produce output and ls and pwd create piped output, even though they don’t need piped input (the same is true for the command head). Your program is expected to implement the following commands. A proper implementation of each should refer back to the correct error messages to print, and the behavior expected of each in a Unix-like system.
Command
Piped input
Piped output
Short Description
Notes
File System Navigation Commands
pwd
No
Yes
Pipes the working directory to the output message queue
You may want to use the Java File class (also see the slide about the File class in the slides from the first tutorial). You can also use System calls to achieve this.
ls
No
Yes
Pipes the contents of the current working directory to the output message queue
You do not need to allow arguments to ls, that is, ls can just always output the contents of the current working directory. You do not need to include “.” or “..” in the list. You also do not need to mark directories with any special characters.
cd
No
No
Change to another directory relative to the current directory
Make sure you can accept the special directories “.” (the current directory) and “..” (one directory up in the directory hierarchy). You do not need to support absolute paths. You do not need to support up-down paths (i.e., cd../hello/world). Important note: the cd command is the only command (other than exit) that does not need to participate in the piping mechanism, it can only output errors, it never accepts piped input or sends piped output. Hence, its implementation can be simpler than your implementation of the other commands.

Text Manipulation Commands
head
No
Yes
Output the first 10 lines (or given number) of one or more files to the output message queue
By default, it outputs the first 10 lines of the given file. It should accept an optional parameter specifying the number of lines to be sent to the output (i.e. head -5 file.txt will output the first 5 lines of file.txt). Unlike in UNIX, head does not accept piped input. I.e. for your program, head will always be first in a string of commands separated by pipes. The head command in a proper UNIX shell has significantly more functionality than the version we require you to implement, you need only implement the output capability.
grep
Yes
Yes
Read lines from piped input, and only output lines with a given search string
You do not need to do regular expression matching. You should do a simple case-sensitive string search, and count any line containing that search string as a match. You can assume that this argument (as well as all file names) will never contain >, or |.
wc
Yes
Yes
Read lines from piped input and output the number of lines, words and characters, separated with space
Note that this command heavily relies on figuring out when it is going to stop getting more input lines to count.
Consider the possibility that there are no lines returned. Make sure that you utilize the isDone() method of this filter and the previous filter.
uniq
Yes
Yes
Filters out duplicate lines from the input
Only permits lines of input to pass through if they have not been detected before (the Collections framework can
be useful here).
>
Yes
No
Reads piped input, and writes it to a file specified after the > symbol
Make sure to utilize the java File class for this operation. If a file exists with the name given by the command, it should be overwritten by a new file
created in its place.
General Commands
exit
No
No
Quit the command line
Like cd, the exit command can be very simple, and it should just end your REPL and return from your main method after printing the “goodbye” message.

Error Reporting
Your program should be able to identify and report the errors shows below:
Invalid Property
Expected Behavior
Undefined Command
> not-a-command
The command [not-a-command] was not recognized.
Invalid Arguments
> head
The command [head] is missing an argument.
> head file.txt | grep
The command [grep] is missing an argument.
Invalid Piping Order
> grep world | head hello.txt
The command [grep world] requires input.
> wc | head hello.txt
The command [wc] requires input.
File/Directory Not Found
> head nonExistantFile.txt
At least one of the files in the command
[head nonExistantFile.txt] was not found.
> cd not-a-directory
The directory specified by the command [cd
not-a-directory] was not found.
Implementation Details
The project is intended to be developed and run on Eclipse (a Java IDE). You will be provided the following files:
SequentialREPL.java:
Your REPL implementation should reside in this file, and should begin when the main(String[])method is called. It should take its input from System.in (using a scanner is suggested).
Filter.java:
An abstract class extended by the SequentialFilter class. You should not modify this class.
SequentialFilter.java:
An abstract class extending the Filter class. For each of the commands mentioned above, you need to extend this class. You should not modify this class. A successful implementation of this project will demand a rigorous understanding of what this class is doing. Understand it before writing any code.

SequentialCommandBuilder.java:
The SequentialCommandBuilder manages the parsing and execution of a command. It splits the raw input into separated subcommands, creates subcommand filters, and links them into a list. We have provided you with some suggested method declarations (a good way to break down the task), but you can feel free to approach this code as you see fit.
AllSequentialTests.java, RedirectionTests.java, REPLTests.java, TextProcessingTests.java, WorkingDirectoryTests.java:
These files contain multiple tests that examine your implementation of the various commands. You should not modify these classes (besides for debugging purposes). Note that a failure of any of these tests will certainly lead to deducted points. All perfect assignments will pass all tests, but passing the tests does not guarantee a perfect score.
Implementation Hints
Your REPL must read user commands from System.in. There are two critical functions within this loop. (1) Correctly parsing commands in a way that allows the creation of piped filters. Using Scanner and String.split will help you with this task. (2) Identifying and reporting the different errors we mentioned above.
All commands that you have to implement must implement the abstract SequentialFilter.java class. The sequential filter is a convenient abstraction: it reads from an input queue (linked list), and writes to an output queue, by calling a method called process(). The queue that it writes to is the input queue for the next filter (if there is a next filter). The abstract class SequentialFilter, should greatly help you along in this project. SequentialFilter inherits from Filter, a class that is not displayed here, but is given along with this assignment.
Note that in implementing the filters, some of them can be simple classes of five or six line of code who solely implement the processLine(String) method in the SequentialFilter class. (Examples of this include grep, >, and the default filter to print to the command line, and uniq). Some other filters (head, ls, pwd) will need to override the process() method of the SequentialFilter class, as they don’t need to use the processLine(String) method to process any piped input. These classes and others may also need to override the isDone() method of the SequentialFilter class.

As we mentioned above, your REPL loop must support simple directory commands: ls, cd, and pwd. cd is the only command you will implement that does not use String queues, because it does not take piped input and does not produce output. You may handle this case without inheriting from the SequentialFilter class.
The directory from which you start the command line program should be initialized as the current working directory that your Java program is initialized to. The working directory can be modified with the cd command. You use the command ls to list the contents of the current working directory. This will be a little more challenging than you might think because you are not allowed to modify the invoking environment’s working directory. You’ll need to manage a separate working directory for your shell. This can be as simple as maintaining a String that keeps track of your present working directory.
Testing and Grading
Successful completion of this project is nearly impossible without appropriately testing against the provided JUnit Tests. If you cannot run the tests when you import your project, please see a TA at office hours as soon as possible.
Importing and exporting properly is crucial to allow us to grade your projects properly. Please import and export based on the instructions in the tutorial slides and on LATTE, and submit the zip file that you exported from Eclipse. If you fail to export your project properly, you may be penalized.
You will need to include Javadocs together with the code you submit. The tutorial will also discuss Java Docs and how you can easily generate them. You should consult our guide on how to write and generate Javadocs on Latte and make sure you include the entire ¡®doc¡¯ folder in your submission. Failure to include Javadocs will be penalized.
To allow us to test without changing your code, please name the class which houses your main method SequentialREPL.java, this will be called when the test suite is run. Because we will be using unit tests that will compare your output (System.out) character for character against our samples, please make sure that your error messages are exactly the same as the ones given, and that there are no intentional or unintentional discrepancies (whitespaces, prompt, wording) between your result and the examples provided. Tests will fail if these discrepancies exist, and a good first place to check a test failure is the exact strings being compared.
There is a helpful Boolean variable called ¡°DEBUGGING_MODE¡± (In the Test Suite), which when set to true will not clean up after the created files and directories after unit tests are run. This can allow you to see the error in your output (files stored in the current working directory), and potentially understand any disconnect between our expectations and your

code. Feel free to play around with the tests (modify them as is helpful for your debugging, etc.), but know that you will be graded against the set of tests that we sent out originally. Any attempt to tamper with the tests will be considered a breach of academic integrity, and will result in point penalties and potentially further discipline.
Note on differences between Unit Tests and Integration Tests
Please note that though JUnit is being used to run this suite of tests, these are not unit tests, but integration tests. Do not take this test design as an appropriate way to test the code that you are writing. Unit tests test specific functions and small, incremental pieces of code. These tests are predicated on the entirety of your code being complete, and this assumption breaks most of the conventions of unit test writing.
Note on Platform Independency
Windows and Unix based operating systems differ on whether to use the ¡°\¡± or ¡°/¡± as their preferred file separator. In order to ensure your code will run in either case it is necessary to request this information at runtime, we have already included the necessary command in the starting code. cs131.pa1.Filter contains a static field FILE_SEPARATOR that will be set to this value for you to easily reference. Using a hardcoded ¡°/¡± or ¡°\¡± in your code instead of Filter.FILE_SEPARATOR to handle changing directories may result in a loss of points.
However, if your code is not platform independent when you submit it, do not fret; we will do our best to run your code on both a Windows and a Mac machine. This part of the assignment is not critical to the learning goal, and will be worth less than five points of your final grade. If you want to make sure your code works on both machines, dual boot computers are available in the library to allow you to try your code in the other environment.
Note on Messages
In order for the test cases to properly evaluate your code, everyone must use the exact same messages for when your REPL loop starts, ends, reports an error, or requests a new command. We¡¯ve included the expected messages in an enum for you to use to help with consistency. The enum cs131.pa1.filter.Message contains 11 messages that you will need to use throughout your project to report information to the user. Note that the 8 error messages contain a placeholder where the invalid command should be placed. Calling the function with_parameter(String parameter) for these 8 messages will replace the placeholder %s with parameter i.e. calling Message.REQUIRES_INPUT(¡°grep foobar¡±) will generate the String ¡°The command [grep foobar] requires input.¡± which is expected by the test cases.

A Note on the Future
In this assignment, you will be building a lot of code, but everything that is mentioned should be complete-able if you have taken COSI 12b (focusing on inheritance and file processing), and COSI 21a (Focusing on data structures and code design). Right now, all of this activity is Sequential (so in the command “head hello.txt | grep foo | uniq | wc”, each filter is run in order, and terminates before the next one begins). In the next assignment, you will use the skills that you will learn in this course to begin using concurrent stream processing, where each filter runs when it can, and the tasks of each filter are processed simultaneously (by multiple cores), or simply concurrently by one core.