*Do not use popen()

For this assignment you will implement your own Shell or Command Line Interpreter (e.g. to replace /bin/bash for simple interactions with the Linux Kernel.). Your shell will be character-oriented, and will fork off processes to execute user commands. Your shell should read lines of user input into a 256-byte buffer, then parse and execute the commands (be sure to clear the buffer between successive commands!) It should be possible for the user to specify the command to execute by giving an absolute path to the file containing the executable (e.g. ./hw1); or to use path expansion to locate the file containing the executable by using the environment PATH variable to construct a series of absolute paths and executing the first file found in this way (note that the execvp() command performs this processing automatically, you do not need to program this yourself!) Your code should parse the input string and separate it into a collection of sub-strings (stored in myargv[]) along with a count of the number of strings encountered (stored in myargc). Note that piped commands will require multiple argc/argv instances!

Your shell should support the following functions (Note : this does not mean your shell supports LS but rather the ability to the LS program):

Execute a single command with up to four command line arguments (including command line arguments with associated flags). For example:

Myshell> ls l Myshell> cat myfile

Myshell> ls al /usr/src/linux

Execute a command in background. For example: Myshell> ls -l &

Myshell> ls al /usr/src/linux &

Redirect the standard output of a command to a file. For example: Myshell> ls -l > outfile

Myshell> ls -l >> outfile

Myshell> ls al /usr/src/linux > outfile2 Myshell> ls al /usr/src/linux >>outfile2

Redirect the standard input of a command to come from a file. For example: Myshell> grep disk < outfile

Myshell> grep linux < outfile2

Execute multiple commands connected by a single shell pipe. For example: Myshell> ls al /usr/src/linux | grep linux

Execute the cd and pwd commands

Myshell> cd some_path

Myshell> pwd

**NOTE** that in most Linux distros CD and PWD are not a program like ls but rather they are shell built-ins. Built-ins are shell commands that are implemented in the shell and not some external binary. For giving your shell the cd and pwd commands, you need to implement these functions in the shell with your code, even if it is provided by your OS. This can be done with the chdir() and getcwd() functions in unistd.h

Suggested implementation strategy to implement a shell with multiple command line arguments (using iterative refinement):

1. Implement shell0 to initialize variables and then go into an infinite loop until stdin detects EOF (i.e. the user enters CTL-D). Each iteration through the loop, the shell program should prompt the user for input, read the input & echo it back to the user and then print out another prompt.

2. Implement shell1 by extending shell0 to parse the users line of input into a series of strings that are stored into a char myargv ** array, along with a count of the number of strings in an int myargc variable. Print out the contents of myargv and myargc after you have parsed the comment. Allow the user to enter an exit command to terminate the shell (instead of typing CTL-D and sending a kill signal to terminate execution).

3. Implement shell2 by extending shell1 to create a child process to execute the command (passing in myargv/myargc as arguments) and then wait for the child process to complete execution.