attached
Project 4: My Shell (aka Command Line Interpreter) Learning Objectives Upon completion of this assignment, you should understand how to: 1. apply basic C/UNIX process management system calls 2. manipulate and redirect a program’s standard input, output and error streams 3. do basic inter-process communication operations The mechanisms you will practice using include: • Process management: fork(), wait(), exec() • File I/O: open(), close(), dup2() • Inter-process communication: pipe() Program Specifications NAME mysh – run a command line interpreter SYNOPSIS mysh [prompt] DESCRIPTION mysh invokes a command line interpreter that supports command execution and input/output redi- rection. mysh takes a single optional command line argument that specifies the prompt string. If no arguments are given, the prompt defaults to “mysh: ”. If “-” is given, do not print a prompt. The basic mysh operation is: 1. print prompt, “mysh: ” 2. read a single input line from standard input 3. execute foreground or background command(s) as specified 4. wait for command(s) executed in the foreground to complete 5. goto 1 OPERATORS On an input command line, commands, operators, and operands are always separated by whitespace. mysh supports the following command line operators: & Place commands into the background: after invoking the specified commands, mysh immediately prints its prompt and waits for another command line. ‘&’ may only be specified as the final command line argument. < redirect="" the="" current="" command’s="" standard="" input="" stream="" from="" the="" file="" named="" immediately="" after="" the=""><’ operator.=""> Redirect the current command’s standard output stream to the file named immediately after the ‘>’ operator. mysh never “clobbers” an existing output file. >> Redirect the current command’s standard output stream to the file named immediately after the ‘>>’ operator. Append to the file if it already exists. | Redirect the current command’s standard output stream to the standard input stream of the suc- ceeding command. There may be any number of pipe-connected processes. EXITING mysh exits with 0 when the user inputs “exit” or CNTL-D. These are the only circumstances under which mysh exits. When command lines are malformed or fail to execute, mysh prints an appropriate error message then prompts the user for another command line. ERRORS Upon error, mysh prints one of the following statements to the standard error stream: “Error: Usage: mysh [prompt]\n” “Error: \”&\” must be last token on command line\n” “Error: Ambiguous input redirection.\n” “Error: Missing filename for input redirection.\n” “Error: Ambiguous output redirection.\n” “Error: Missing filename for output redirection.\n” “Error: Invalid null command.\n” “Error: open(\”%s\”): %s\n”, file name, strerror(errno) Implementation Details mysh searches the PATH environment variables for executables; i.e., mysh uses execvp(). When opening files with O_CREAT, use “S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH” for the mode, i.e., the third argument to open(). Requirements and Constraints 1. Command lines are restricted to 1024 characters or less. 2. You may use wait(), but not waitpid(). 3. When executed in the foreground, mysh waits for all processes in a pipe to complete. 4. Transient zombies may exists, but mysh should periodically clean up zombie processes. 5. Close unneeded file descriptors. When a child process calls execvp(), there should only be three open file descriptors: 0, 1 and 2. Likewise, when the main mysh process is fetching a command line. mysh should not use a file descriptor greater than 5. 6. Comprehensively handle all error conditions, including operator misuse (e.g., trying to redirect output to a pipe and a file at the same time) and bad commands. 7. If a command line is malformed (i.e., syntactically incorrect), no part of it should be executed. (You do not need ot check whether commands are valid or executable.) Tips 1. Use fgets() to read input lines. 2. Check fgets() and wait() for premature returns due to system interruption: if fgets() or wait() fails and errno == EINTR, try the call again. 3. You may use the provided tokens.[c,h] files to “tokenize” your command lines. 4. Always check to make sure execvp() did not unexpectedly fail. 5. Remember in execvp()’s argument vector, the first element must be the command itself, and the last element must be NULL. 6. You will need to track the pids of all foreground command processes in order to make sure they have all terminated before fetching a new command line. 7. Kill all stray processes left around after quitting mysh 8. Consider implementing and testing in the following order: a. Parsing command line into individual commands. b. Executing commands in the foreground. c. Executing commands in the background. d. Redirecting input/output in the foreground e. Redirecting input/output in the background f. Implementing a single pipe g. Implementing two or more pipes 9. A good strategy is to separate the command line processing phase from the command line execution phase, and check for all command line issues, including bad I/O redirection, during command line processing, i.e., before executing any part of the command line. 10. Consider the case when there is a foreground process running, and a background process terminates. Make sure mysh doesn’t try to fetch a new command line prematurely. Submission You will submit your main C file and on the department server, using the turnin program. If you would like to request an extension, please use the reqext program. Details on how to use both can be found on Canvas.