Build a Simple HTTP/1.0 Web Server with CGI and Database Support in C

Project Goals

The aim is to implement a minimal HTTP/1.0 web server that can:

Accept GET and POST requests from clients.

Return requested resources as HTML pages and handle errors such as missing files with a 404 page.

Execute simple CGI scripts and return their output.

Perform basic database operations (create, read, update, delete) through web pages.

Basic Framework of the HTTP Server

HTTP Protocol – An application‑layer protocol built on TCP that follows the sequence: connect → request → response → close.

URL – Uniform Resource Locator; includes protocol, host, optional port, and path. Two formats are used in the project: URLs with query parameters for GET and URLs without parameters for POST.

Request & Response Format – The response header contains a status code (e.g., 404 NotFound) indicating the result of the request.

Implementation Approach

1. Because HTTP runs over TCP, the first step is to establish TCP communication between client and server processes. 2. After receiving a request, the server parses the request method (GET or POST) to decide how to handle CGI. 3. The server extracts the URL; for GET the parameters are in the URL, for POST they are in the request body. 4. It checks whether the requested resource exists and determines if it is a directory, a regular file, or an executable program.

Non‑CGI mode – When the request is a GET without parameters, the server calls echo_www() to return the resource as an HTML page.

CGI mode – The server forks a child process to run the CGI program, passes parameters via environment variables, and redirects the child’s I/O so that its output is sent back to the browser.

The responsibilities of the parent and child processes are summarized as follows:

Error Handling

The echo_www() function is reused, but the response header is modified to send the appropriate status code and description. For example, when a resource is missing, a custom 404 page is returned.

Project Files

Directory layout:

cgi – CGI implementation code.

conf – Configuration files (IP and port).

log – Shell and HTTP error logs.

lib – MySQL libraries.

sql_client – MySQL API and CGI code.

wwwroot – Web root containing static pages, 404 page, and executable CGI programs.

Key source files: configure.sh – Shell script that generates the Makefile. http_ctl.sh – Script to start, stop, and restart the server. httpd.pid – Stores the server PID for daemon management. httpd.h – Declarations. httpd.c – Implementations. main.c – Main server logic.

Implementation Results

When the requested resource exists, the server returns the page correctly:

After submitting data to a CGI script, the processed result is displayed:

Source Code

Repository: https://github.com/lybb/Linux/tree/master/httpd

Common Issues

Local loopback works, but external browsers cannot connect without bridge mode; remember to disable the firewall.

If the server sends raw code instead of HTML, check the length passed to send() in echo_www().

Images not displayed because the server only handled directories and executables; add handling for regular files.

Partial image display caused by an insufficient line buffer size.

Incorrect Content-Type header ("text" vs. "text/html") leads browsers to prompt a download.