pipex

A C program that reproduces the behavior of the shell pipe (`|`). This 42 school project uses `fork()`, `pipe()`, and `dup2()` to manage processes and I/O.

---

This project is a C program that reproduces the behavior of the shell pipe (|). It's a 42 school assignment designed to provide a deeper understanding of process creation and inter-process communication in Unix-like operating systems using system calls like pipe(), fork(), dup2(), and execve().

Features

• Mandatory Part: Replicates the behavior of < infile cmd1 | cmd2 > outfile.
• Bonus Part:
  • Supports multiple pipes: < infile cmd1 | cmd2 | ... | cmdn > outfile.
  • Supports here_doc input: cmd1 << LIMITER | cmd2 >> outfile.

Getting Started

Prerequisites

• A C compiler (like gcc or clang)
• make

Building the Project

1. Clone the repository:

   git clone https://github.com/jdecorte-be/pipex.git
   cd pipex

2. The project includes a libft submodule. Initialize and update it:

   git submodule update --init --recursive

3. Compile the program using the Makefile:

   # For the mandatory part
   make
   
   # For the bonus part (multiple pipes and here_doc)
   make bonus

Usage

Mandatory

The program takes four arguments: an input file, two commands, and an output file. It redirects the output of the first command to the input of the second command.

Syntax:

./pipex <infile> <cmd1> <cmd2> <outfile>

Example: This is equivalent to the shell command: < infile grep a1 | wc -l > outfile

./pipex infile "grep a1" "wc -l" outfile

Bonus: Multiple Pipes

The bonus version can handle an arbitrary number of commands chained together.

Syntax:

./pipex <infile> <cmd1> <cmd2> <cmd3> ... <outfile>

Example: Equivalent to: < infile grep a | cat -e | wc -l > outfile

./pipex infile "grep a" "cat -e" "wc -l" outfile

Bonus: Here Document (here_doc)

The bonus version also supports here_doc as an input source.

Syntax:

./pipex here_doc <LIMITER> <cmd1> <cmd2> ... <outfile>

Example: Reads from standard input until EOF is entered, then pipes the result through grep and wc.

./pipex here_doc EOF "grep a" "wc -w" outfile

Core Concepts

This project revolves around orchestrating multiple processes and managing their input and output streams. The goal is to create a data flow like this:

infile -> [stdin] cmd1 [stdout] -> [pipe] -> [stdin] cmd2 [stdout] -> outfile

This is achieved using four key system calls.

1. pipe() - Inter-Process Communication

A pipe is a unidirectional data channel that allows one process to send data to another.

int pipe(int pipefd[2]);

• pipe() creates a pipe and returns two file descriptors in the pipefd array:
  • pipefd[0]: The read end of the pipe.
  • pipefd[1]: The write end of the pipe.
• Data written to pipefd[1] can be read from pipefd[0].

2. fork() - Process Creation

To run two commands simultaneously, we need two separate processes. fork() creates a new process by duplicating the calling process.

pid_t fork(void);

• In the parent process, fork() returns the process ID (PID) of the newly created child process.
• In the child process, fork() returns 0.
• If an error occurs, it returns -1.

This allows us to execute different code blocks for the parent and child, enabling them to handle cmd1 and cmd2 respectively.

3. dup2() - I/O Redirection

By default, a process reads from standard input (stdin, fd 0) and writes to standard output (stdout, fd 1). dup2() allows us to redirect these streams.

int dup2(int oldfd, int newfd);

dup2() makes newfd a copy of oldfd, closing newfd first if necessary. This is the key to connecting our files and pipes to the commands.

• For the first command (child process):
  1. Redirect stdin to read from infile: dup2(infile_fd, STDIN_FILENO).
  2. Redirect stdout to write to the pipe: dup2(pipefd[1], STDOUT_FILENO).
• For the second command (parent process):
  1. Redirect stdin to read from the pipe: dup2(pipefd[0], STDIN_FILENO).
  2. Redirect stdout to write to outfile: dup2(outfile_fd, STDOUT_FILENO).

4. execve() - Command Execution

The execve() system call replaces the current process image with a new program. This is how we run the user-specified commands (e.g., ls, grep, wc).

int execve(const char *pathname, char *const argv[], char *const envp[]);

• pathname: The absolute path to the command's executable (e.g., /bin/ls).
• argv[]: An array of arguments for the command, with the first element being the command name itself (e.g., {"ls", "-l", NULL}).
• envp[]: An array of environment variables, which is needed to find the command pathname if it's not an absolute path.

To find the correct pathname, we parse the PATH variable from envp, split it by :, and test each directory to find the executable.

Debugging Tips

• Hanging Program: If your program hangs, it's almost always because a pipe's file descriptor was not closed correctly. Every process that has access to the pipe must close both ends when it's done with them. The kernel will only signal "end-of-file" on the read end when all write ends have been closed.
• No Output: printf writes to stdout. If you have redirected stdout using dup2, your debug messages will go to a file or a pipe, not the terminal. Use perror() or write directly to stderr (fd 2) for debugging messages that should always appear on the console.

  // perror prints the error message for the last failed system call
  perror("Error executing command");
  
  // ft_putstr_fd is a useful function from libft
  ft_putstr_fd("Debug: In child process\n", STDERR_FILENO);

• Command Not Found: Before calling execve(), use access() to check if the command executable exists and has the correct permissions. This allows you to provide a more informative error message, like bash: cmd: command not found.
• File Permissions: Handle errors from open() gracefully. Your program should mimic the shell's behavior when infile doesn't exist or outfile cannot be written to.