This project has been created as part of the 42 curriculum by vblanc, yafahfou, yabokhar.

Description

This project consists of an HTTP server written in C++ (C++98 standard). Its main goal is to recreate the core behavior of a real web server by handling client connections, parsing HTTP requests, and generating appropriate HTTP responses.

The server is configured using a configuration file inspired by nginx, allowing flexible setup of server blocks, routes, ports, error pages, and other directives. It supports the three main HTTP methods: GET, POST, and DELETE.

Additionally, the server implements CGI (Common Gateway Interface) support, enabling the execution of dynamic scripts such as .pl (Perl), .py (Python), .php (PHP), and .cgi files. This allows the server to handle dynamic content alongside static files.

The project focuses on low-level network programming, process management, file handling, and strict compliance with the HTTP/1.1 protocol, while respecting the constraints of the C++98 standard.

Instructions

The project is compiled using the provided Makefile:

make        # Compile the project
make clean  # Remove object files
make fclean # Remove object files and the executable
make re     # Recompile everything

The executable generated is:

./webserv

To run the server:

./webserv [configuration_file] (if no argument is given, the server will use the default configuration file named "config/good/webserv.conf")

Resources

• Beej's Guide to Network Programming
• "99% of Developers Don't Get Sockets"
• Socket Programming in C++
• HTTP Status Codes — Wikipedia (FR)
• Common Gateway Interface
• HTTP Status Codes — MDN
• HTTP Semantics — RFC 9110
• HTTP Messages — MDN
• HTTP/1.1 — RFC 9112
• Cookie Header — MDN
• AI was used to generate HTML pages, server tests and to introduce us to Github Actions

AI Usage

AI tools were used for:

• Improving documentation wording
• Debugging specific errors
• Generating HTML pages, server tests (introduction to Github Actions)

All code was written, reviewed, and validated by team members.
AI was not used to generate complete features without understanding.