Sudoku Solver is a proof of concept-app on how to leverage WebAssembly in combination with existing C++ libraries.
The application consists of a serverless endpoint that basically is a front for a small WebAssembly application that does the actual solving.
The solving is based of functionality from the OR-Tools library (https://developers.google.com/optimization) that has been compiled into WASM.
See https://github.com/kjartanm/wasm-or-tools
The C++ app is based on one of the contributed examples in the OR-Tools repo: https://github.com/google/or-tools/blob/stable/examples/contrib/Sudoku.java
In addition it has an endpoint with a small web-application that uses the serverless endpoint.
Demo: https://sudoku-solver.krmuller.workers.dev/
It uses Cloudflare Workers and is based on worker-emscripten-template
A template for kick starting a Cloudflare worker project with emscripten
index.js
is the content of the Workers script.
main.c
is the c source code that calls into the stb image resizer library.
build.js
holds the command we use to call emscripten.
webpack.config.js
holds the webpack config we use to bundle the emscripten output together with your script.
This template requires Docker for providing the emscripten build environment. While we believe this provides the best developer experience, if you wish to not use Docker you can delete the check for docker and the docker parts of the build command in build.js
. Note this means you must have emscripten installed on your machine.
This template requires version >=1.6.0 of Wrangler
$ wrangler generate myapp https://github.com/cloudflare/worker-emscripten-template
🔧 Creating project called `myapp`...
✨ Done! New project created /path/to/myapp
To demo you can use wrangler dev
$ wrangler dev
👂 Listening on http://localhost:8787