A C++ implementation of the bytecode virtual machine from "Crafting Interpreters" by Robert Nystrom.
This uses some elements of the C++ standard library, but tries to stay close to the philosophy of "doing things from scratch" that the book uses.
For example, that means I use <iostream> and std::format, but wrote a
simple resizable array type instead of using std::vector. It also means
using C-style tagged unions instead of std::variant.
I have never written a bytecode VM before, so I'm doing it as close to the book in a more type-safe C++ way, and then refactoring into more idiomatic C++ as I go. This won't ever be a production bytecode VM, but it should eventually show my C++ version of how the book does things.
This project currently uses cmake, so you can build as follows.
Use an out-of-tree build:
mkdir build
cd build
cmake .. # Generate project
cmake --build . # Build project
ctest # Run unit tests
This will build the bytecode virtual machine and run the associated file:
Example:
python .\scripts\run_sample.py .\samples\expressions.lox --config Release
Acceptance tests compare the standard output of running the VM on a file ending
in .lox, against the expected output in a a similarly named .expected file.
The acceptance test script builds the VM and then runs the tests in Release mode.
Run the test battery like:
python .\scripts\run_acceptance_test.py