A minimal snapshot testing tool
to harden the bark of your code.

Info • Usage • Screenshots • License

---

Info

General

I built this tool to integrate nicely with the various things I'm building in different languages - for the projects where simple unit tests are not enough. My main constraints when building this tool was that it needed to be simple to use and (most importantly) language- and toolchain-agnostic.

Snapshot testing

bark.py uses the stdout/stderr of your test-subject to determine its success-condition. It achieves this by first recording a "source of truth" onto disk, and then comparing that to your subsequent tests. By setting up this kind of workflow you can introduce more certainty into your development progression.

"There are already hundreds of scripts and programs that implement this, so why should I use bark.py?" you may say - and I don't care.

Requirements

• Python 3.10+

---

Usage

Setup

Choose a directory in your codebase where you want to use bark.py - create a new file called bark_test in that directory:

cd my/program/test
curl -O https://raw.githubusercontent.com/simon-danielsson/bark.py/refs/heads/main/bark.py
touch bark_test

Within bark_test you write the commands of each test you wish to execute, one command per line! Following the pattern <name of test>|<test command>:

add_nums() | gcc -o main main.c && ./main --test_nums
print_list() | gcc -o main main.c && ./main --test_print

CLI

bark.py comes with two commands:

record
    * Reads a file 'bark_test' with shell commands to be executed.
    * Executes each command and saves their respective stdout/err to a file
      inside a generated directory '.bark'.

compare
    * Runs the same file of shell commands again and compares their
      stdout/err to their recorded counterparts in the '.bark' directory.
    * Prints a summary.

Within the directory of your bark_test file, you run ./bark.py record to record the output of your tests and generate a folder .bark.

Once you've progressed enough in your development process to warrant a test, you simply cd back into the directory of your bark_test file and run ./bark.py compare - to compare your current output with what you had previously recorded.

See the tests directory for a simple showcase

---

Screenshots

---

License

This project is licensed under the MIT License.