xeus-cling, a Jupyter Kernel for C++, has been around for many years. It is based on xeus, a native implementation of the Jupyter Kernel protocol, and cling, a C++ interpreter developed at the CERN, based on an outdated version of Clang. This limits its capabilities and compatibility with modern C++ features and makes it hard to maintain on all the platforms.

More recent versions of Clang have incorporated a Read-Eval-Print-Loop in their code base, making it possible to use them directly in interactive development environments like Jupyter Notebooks. xeus-cpp leverages these advancements to provide a smoother and more powerful C++ programming experience in Jupyter.

In this talk, we will first demonstrate xeus-cpp and give an overview of the new features it provides compared to xeus-cling. We will then dive into its architecture and explain how users can extend it thanks to its plugin system.

In the next section, we will show that xeus-cpp also provides a jupyter-lite kernel running entirely in the browser, enabling larger-scale deployment. We will conclude with a roadmap of future developments, including a debugger for C++ in Jupyter!