Flint's official website. The Front-end is in Elm, and the backend is in Haskell. This repo aims to showcase some of the internal tech practices that we adopt, promote openness, and do things in public as we are a remote company. It's an important part of our culture.
The website is used to communicate to the industry what we offer, a solution for hiring internationally educated registered nurses. On the other end, explain how it works for nurses and our offering. Lastly, the third responsibility of the website is to host our open positions at Flint.
Run nix develop to enter a shell environment with all the needed packages and scripts, and run nix run .#withflint to run the website locally.
Use the watch script in that shell to hot reload the Haskell and Elm code.
The newer macOS versions use the port 5000 for AirDrop. You can either disable that feature on your OS or use a different port in the code. We could improve this by making the port an environment variable. But you need to make sure you don't commit to the new port. To change it, search 5000 in the code.
You should know that the SMTP server is IP range protected. So it is convenient if you don't want to send out emails uselessly, but a little more complicated if you work on it.
In the nix run commands, we use elm-format to keep the code tidy, but in Elm, we use elm-review to apply a set of rules. Ultimately it's the programmer's responsibility to uphold the standard, but elm-review is there to help. The principle to think about is if someone else should run elm-review or any Makefile command after you, would that cause the code to change. It's something we want to avoid. So run nix run .#fix before committing to git.
When submitting an increment, you should make a pull request from your form of the repo and use all the best practices promoted in open source. Being a remote company, well, that is how we can be successful.
The repo is open to the public, so you should be mindful of the language you use but also no to expose any secrets. When unsure, ask.
The repo is built with GitHub Actions, and a docker image is pushed to AWS ECR, which is then deployed to Kubernetes using FluxCD.
The build success/failure is available on the Slack automation channel and on GitHub. You can be invited as a contributor.
Deployments are configured and managed in another git repo (GitOps). But in principle, the code runs the exact same way.