mini_serv_tester

A robust, fully-automated Python test suite for the Rank 06 Exam project mini_serv (42 School Cursus).

It uses Python's socket, subprocess, and select libraries to simulate multiple concurrent clients, validating your server against all standard project constraints.

Features

• Basic Connections & Messages: Tests client arrivals, messages spanning multiple lines, partial messages (missing newline), and client departures.
• Stress & Load Testing: Spawns 250 dummy clients to test server file descriptors and stability under load.
• Large Payloads: Transmits continuous message structures larger than typical buffer sizes (8KB+) to ensure chunks are assembled and fragmented efficiently without data corruption.
• Broadcasting Speed Test: Benchmarks latency by sending a single broadcast and monitoring propagation to 200 clients under strict timeouts (simulating the exam requirement: "Warning our tester is expecting that you send the messages as fast as you can").

Requirements

• Python 3.x
• Standard Python libraries (no pip install required)
• Your compiled mini_serv executable located in the same directory running the script.

Usage

1. Compile your server and ensure the executable mini_serv is present.
2. Run the tester:

python3 exam06-tester.py

The script will automatically detect a free port and launch mini_serv, connecting multiple simulated clients to run all checks and validations.

Valgrind Support

To test your server for memory leaks and open file descriptors, you can use the --valgrind flag:

python3 exam06-tester.py --valgrind

When you use this flag, the tester switches to Manual Mode. It will find a free port and pause, giving you the exact Valgrind command to run in a separate terminal.

Why is this needed? During stress tests, when many clients disconnect simultaneously, writing to a closed socket throws a fatal SIGPIPE signal that kills your server instantly. Because the strict exam rules forbid you from using signal() in your C code to ignore it, your server would artificially crash during testing before Valgrind can generate a memory report.

We solve this by having you run the server manually wrapped in a subshell that traps and ignores SIGPIPE. The tester will provide you with a command exactly like this to copy and paste into a new terminal:

(trap '' SIGPIPE; valgrind --leak-check=full --show-leak-kinds=all --track-origins=yes --track-fds=yes -s ./mini_serv <port>)

Once your server is running via this command, press ENTER in the tester to begin the suite. Once finished, you can safely Ctrl+C your Valgrind terminal to read a clean, perfectly un-interrupted leak report!

Tests Overview

• Test 1: Basic connection & arrival notification
• Test 2: Basic message routing
• Test 3: Partial message buffering (simulating multiple recv() calls logic)
• Test 4: Multiple lines concatenated sequentially
• Test 5: Client leaving notification
• Test 6: Stress connections (250 clients without crashing)
• Test 7: Mass messaging & large texts routing to target group
• Test 8: Large payload processing (> 8KB)
• Test 9: Speed / Latency Broadcast Benchmark (ensuring fast message propagation to avoid "too slow" errors during exam grading)