Rewriting to Haskell–Project Setup
Riccardo Odone
Posted on March 3, 2020
You can keep reading here or jump to my blog to get the full experience, including the wonderful pink, blue and white palette.
This is part of a series:
We decided to go with Servant to rewrite the current Rails API for Stream. We don't really have a strong reason for that. We just like it and we believe it will allow us to do some cool stuff down the line!
Also, we picked Stack because it's what we are used to and seems to be less cryptic than Cabal.
We created the project with the following command:
stack new haskell servant
# ^ Create a new project..
# ^ ..in a new haskell/ folder..
# ^ ..using the Servant template.
The Servant template creates a dummy application with an hardcoded endpoint and some tests that allow to start playing with code right away.
CI
Stream is already using CircleCI, so we opened the Haskell languge guide and started copy / pasting like there was no tomorrow.
We ended up with the following config.yml file that supports Rails, Elm and Haskell. See inlined comments for more info.
---
version: 2.1
commands:
install_dependencies:
# ^ Define a reusable command (see invokations below) to install the dependencies needed for Rails and Elm.
steps:
- restore_cache:
name: Restore bundle cache
key: stream-{{ checksum "Gemfile.lock" }}
- restore_cache:
name: Restore yarn cache
key: stream-yarn-{{ checksum "yarn.lock" }}
- run: bundle install --path vendor/bundle
- run: yarn install
- save_cache:
name: Store bundle cache
key: stream-{{ checksum "Gemfile.lock" }}
paths:
- vendor/bundle
- save_cache:
name: Store yarn cache
key: stream-yarn-{{ checksum "yarn.lock" }}
paths:
- ~/.cache/yarn/v1
- node_modules
jobs:
build:
# ^ Test Rails and Elm code.
working_directory: ~/stream
docker:
- image: circleci/ruby:2.4.1-node-browsers
environment:
PGHOST: 127.0.0.1
PGUSER: stream
RAILS_ENV: test
- image: circleci/postgres:9.5-alpine
environment:
POSTGRES_USER: stream
POSTGRES_DB: stream_test
POSTGRES_PASSWORD: ""
steps:
- checkout
- install_dependencies
# sysconfcpus is a hack for elm-make
# https://github.com/elm-lang/elm-compiler/issues/1473
- run: |
git clone https://github.com/obmarg/libsysconfcpus.git ~/libsysconfcpus
cd ~/libsysconfcpus
./configure
make
sudo make install
cd -
- run: yarn test
- run: yarn run elm-format --validate app/javascript/
- run: sysconfcpus --num 2 yarn run elm-test app/javascript/tests
- run: dockerize -wait tcp://localhost:5432 -timeout 1m
- run: cp config/database.yml.example config/database.yml
- run: bin/rails db:setup
- run: sysconfcpus --num 2 bin/rspec spec
- run: bundle exec bundle audit check --update
- run: bundle exec brakeman --ensure-latest -A -5
- run: bundle exec bundle outdated || true
build_haskell:
# ^ Test Haskell code.
working_directory: ~/stream/haskell
docker:
- image: fpco/stack-build:lts
steps:
- checkout:
path: ~/stream
- restore_cache:
name: Restore Cached Dependencies
keys:
- stream-haskell-{{ checksum "stack.yaml" }}-{{ checksum "package.yaml" }}
- stream-haskell-{{ checksum "stack.yaml" }}
- run:
name: Resolve/Update Dependencies
command: stack --no-terminal setup
- run:
name: Run tests
command: stack --no-terminal test
- run:
name: Install executable
command: stack --no-terminal install
- save_cache:
name: Cache Dependencies
key: stream-haskell-{{ checksum "stack.yaml" }}-{{ checksum "package.yaml" }}
paths:
- "/root/.stack"
- ".stack-work"
- store_artifacts:
path: ~/.local/bin/haskell-exe
destination: haskell-exe
deploy:
# ^ Deploy the Rails and Elm application. We will see in a later post how to deploy the Servant application.
working_directory: ~/stream
docker:
- image: circleci/ruby:2.4.1-node-browsers
steps:
- checkout
- install_dependencies
- run : |
if [ "${CIRCLE_BRANCH}" == "master" ]; then
bundle exec cap staging deploy
elif [ "${CIRCLE_BRANCH}" == "production" ]; then
bundle exec cap production deploy
else
echo "${CIRCLE_BRANCH} is a feature branch so no deploy"
fi
workflows:
version: 2
build_and_deploy:
jobs:
- build
- build_haskell
- deploy:
requires:
- build
- build_haskell
Get the latest content via email from me personally. Reply with your thoughts. Let's learn from each other. Subscribe to my PinkLetter!
Posted on March 3, 2020
Join Our Newsletter. No Spam, Only the good stuff.
Sign up to receive the latest update from our blog.