Skip to main content

A server backend for interactive online exercises.

Project description

A server backend for interactive online exercises.

Testing and documentation

Testing is done using tox. Install it, and run it from the command line in the repository root. This will create a virtualenv for each supported environment, install the necessary things in it, run the tests, and build the documentation.

Speeding up builds

For a faster experience, it is recommended that you configure pip to use wheel by default, by placing the following in your ~/.pip/pip.conf or equivalent:

use-wheel = True

find-links = /tmp/wheelhouse

wheel-dir = /tmp/wheelhouse

After that, build wheels out of the requirements by running the following command once:

pip wheel -r requirements.txt -r requirements-docs.txt -r requirements-testing.txt

That will build wheels, which are faster to install than regular packages. You can make installations even faster by adding no-index = True to the [install] section; that way, installations won’t even hit PyPI, further reducing latency. Keep in mind that you will then no longer be able to use pip install to install anything, unless you’ve first made a wheel out of it.

Release notes


Added per-user secrets. Once pyca/cryptography will have HKDF support, this will turn into per-user entropy.

Also minor fixes, version upgrades.


Now being continuously tested on Travis, with coverage monitoring thanks to Coveralls.

Several other minor improvements:

  • A useful mixin for resources that represent exercises, for helping them resolve and notify when an exercise is completed.

  • User e-mails are indexed.

  • Persisted factories are now unary callables that take a store. This prevents some otherwise icky global mutable state.

  • New behavior for adding persisted factories: addOrUpdate; don’t just blindly add more factories with the same identifier.


  • Added support for a dhparam.pem, enabling DH-based PFS ciphersuites. ECDH-based PFS ciphersuites is a work in progress, see #6.


  • Moved requirements into for easier installs

  • clarent version requirement bump



  • localhost manhole support for debugging

  • Only support good ciphersuites.


Renamed to merlyn (see “Whence the name” below).


  • Drastically simplified exercise API

  • Authentication API based on SSL certificate verification



  • Basic documentation for steps and exercises

  • Interfaces: IStep, IRenderer, IValidator (see docs)

  • A renderer based on string templating (str.format)


  • repoze.sphinx.autointerface -> sphinxcontrib-zopeext, which appears to be a shinier, more updated version of the same thing

Common things between merlyn and arthur, such as shared AMP command classes, were moved to clarent.



  • Exercise and Step classes

  • Step validation draft

  • Step solution submission interface


Initial public release. Nothing much to see here.

Whence the name?

This project was originally called merlin, because the step-by-step oracle-like model reminded me of Merlin in the AM complexity class and Arthur-Merlin protocols. It’s since been renamed to merlyn, because the primo merlin PyPI real estate has been taken up by some kind of weird setuptools fork.

Since Arthur is the person who performs the protocol together with Merlin, it only made sense to name the client side project arthur. Finally, clarent, named after king Arthur’s ceremonial sword, holds common tools.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

merlyn-0.0.9.tar.gz (21.1 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page