Table of Contents

• Table of Contents
  • About this tutorial
  • Goal of this tutorial
  • How to run it?
    • BinderHub
    • Local installation
      • On Unix systems (Linux, macOS)
      • On Windows
    • Docker
  • For maintainers and contributors

About this tutorial

This tutorial is heavily inspired from a C++ tutorial created by David Chamont (CNRS) that was given as a lecture with the help of Vincent Rouvreau (Inria) in 2016; latest version of this tutorial used as the basis of current one may be found there.

Current version provides two major modifications:

• The tutorial is now in english.
• Jupyter notebooks using homemade Cppyy kernel are now used, thus enabling a sort of interpreted C++ which is rather helpful for teaching it.

We have rewritten entirely and modified heavily several chapters, but the backbone remains heavily indebted to David and Vincent and the hands-on is still very similar to the original one.

As the original tutorial, the present lecture is released under the Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) licence.

Goal of this tutorial

This is an introductory lecture to the modern way of programming C++; at the end of it you should:

• Understand the syntax and basic mechanisms of the C++ language in version 14/17.
• Know the different programming styles.
• Know of and be able to use the most useful part of the standard library.
• Be aware of many good programming practices in C++.

How to run it?

You may run the notebooks by one of the three following methods:

• From your browser through a Binder instance.
• Directly on your computer system.
• Through a Docker image on your computer.

Quick guides for each of these methods are given below.

BinderHub

A link to a BinderHub instance is given at the top of the project page; foresee few minutes to set up properly the notebooks.

The pro of using Binder is that you have basically nothing more to do than click on the link; the cons are the delay to create the instance, the need for an Internet connection and the fact you won't keep locally the possible modifications you did.

Local installation

On Unix systems (Linux, macOS)

• The dependencies to run this tutorial are:

• Jupyterlab
• Our dedicated C++ Jupyter kernel.

pip install jupyterlab
git clone https://gitlab.inria.fr/sed-saclay/cppyy_kernel
pip install ./cppyy_kernel

You may of course choose another way to install JupyterLab (typically using Anaconda or Mamba)

• Then you can run the notebook by going into its root directory (or internal links won't work...) in a terminal and typing:

jupyter lab

NOTE: It is possible to use the notebooks directly from some IDEs like VSCode.

On Windows

None of us is using Windows, and Cppyy kernel was not tested for this OS at all.

Your best option if you're using Windows 10 or 11 is probably to install Ubuntu in your Windows session.

Docker

It is possible to execute the notebooks in a Docker container.

First get the image from Gitlab registry:

docker login registry.gitlab.inria.fr
docker pull registry.gitlab.inria.fr/formations/cpp/gettingstartedwithmoderncpp/xeus-cling:latest

Then run a container with:

docker run --rm -e JUPYTER_TOKEN='easy' -p 8888:8888  --cap-drop=all -v $PWD:/home/dev_cpp/training_cpp  registry.gitlab.inria.fr/formations/cpp/gettingstartedwithmoderncpp/xeus-cling:latest

And in your browser type http://localhost:8888

and then type easy in the token dialog box (of course you may replace by whatever you want).

Few hints for those not familiar with Docker:

• -v creates a mapping between local folder and the /home/dev_cpp/training_cpp folder in the container; this enables you to edit the file from your comfy local environment and see the file edited this way in the Docker container.
• --cap-drop=all is a safety when you're running a Docker image not built by yourself: you're essentially blocking the few remaining operations that might impact your own environment that Docker lets by default open with the run command.
• -p gives the port mapping between the Docker image and your local environment.
• --rm tells docker to delete the container after its use.

The lengthy registry.gitlab.inria.fr/formations/cpp/gettingstartedwithmoderncpp/xeus-cling-and-compilers is the name of the Docker image, if this image is not present locally in your environment Docker will try to fetch it from a registry on the Inria Gitlab.

Then just type http://localhost:8888/ in your browser to run the notebooks.

For maintainers and contributors

Information related to CI are here. All contributions are welcome, but please read the contribution guide first.