Giter VIP home page Giter VIP logo

haskell-nix's Introduction

Nix and Haskell in production

This guide documents how I use Nix for Haskell development. Feel free to open issues or pull requests if you would like to contribute or suggest improvements

The purpose of this project is to support two Haskell workflows:

  • Workflow #1: Nix provisions the development environment
    • Nix provides all dependencies and the Haskell toolchain
    • You still build the root project using cabal
    • This approach is ideal for development as it supports incremental builds
  • Workflow #2: Nix builds the root project for you
    • This approach is ideal for continuous integration (especially Hydra)

The emphasis of this guide is to be as robust as possible and gracefully handle writing Haskell projects at scale. Some of the suggestions in this guide might be overkill for a small Haskell project but are essential when managing multiple private Haskell projects across a team of developers.

This guide is based partly on the Haskell section of the nixpkgs manual and partly on experience using Nix and Haskell in production at Awake Security.

Background

Nix is not a cabal replacement and Nix actually complements cabal quite well. Nix is much more analogous to a stack replacement. stack does provide some support for Nix integration, but this document does not cover that. Instead, this document describes how to use Nix in conjunction with cabal for Haskell development

The main benefits of using Nix over stack are:

  • Binary caches

    Nix lets you download precompiled Hackage packages whereas stack compiles them on your computer the first time you depend on them

  • Space efficiency

    stack creates a copy of each package for each resolver. This means that if you have two projects with different resolvers then they will not use the same copy of shared dependencies

  • Generality

    Nix is a language-independent build tool. This means you can use Nix to also build and customize non-Haskell dependencies (like gtk). This uniform language simplifies build tooling and infrastructure.

  • Larger ecosystem

    Nix provides a large ecosystem of tools that integrate with anything that Nix can build, such as Hydra (continuous integration), NixOS (an operating system), and NixOps (a deploy tool)

  • Flexibility

    Nix is a powerful tool in the hands of advanced users. You can make very deep and sweeping changes to your toolchain, such as recompiling everything with security hardening

The main disadvantage of using Nix over stack are:

  • Poor OS X support for stable NixOS releases

    The public binary cache only caches OS X packages for unstable nixpkgs releases. If you try to use a stable NixOS release you will spend hours building GHC multiple times since Nix bootstraps GHC starting all the way from GHC 7.4

  • Verbosity

    Nix derivations for Haskell projects are significantly more complex than their corresponding stack.yaml files. The release.nix files in this repository are the Nix analog of a stack.yaml file and you can see for yourself the increase in complexity as the examples progress in difficulty.

  • Poor error messages

    Nix is an untyped language with no special Haskell integration, so error messages are unhelpful

  • Nix cannot incrementally compile Haskell libraries

    Note that you can still use Nix to provision a development environment and incrementally compile a Haskell package using cabal. However, if you use Nix to build the package then Nix will build the package from scratch for every minor change. In theory, this could be fixed to have Nix directly support incremental Haskell builds but this has not been done yet.

  • Worse user experience

    Nix does not provide many conveniences that stack does such as bootstrapping new projects or "file watch"

Both Nix and stack use curated package sets instead of version bounds for dependency management. stack calls these package sets "resolvers" whereas Nix calls these package sets "channels". Nix provides stable channels with names like NixOS-16.09 (analogous to stack's LTS releases) and then an unstable channel named nixpkgs-unstable (analogous to stack's nightly releases)

Related tools

Before continuing, I'd like to mention some other tools for mixing Haskell with Nix:

  • tinc - this uses cabal's solver to select which Haskell packages to use instead of the curated Haskell package set from nixpkgs
  • styx - This tool provides a stack-like interface to managing Haskell dependencies using Nix

Setup

Before you begin, you must install Nix if you haven't already:

$ curl https://nixos.org/nix/install | sh

You must also install cabal2nix and nix-prefetch-git:

$ nix-env -i cabal2nix
$ nix-env -i nix-prefetch-git

You also need to install cabal if you haven't done so already. You can either use your installed cabal or you can use nix to install cabal for you:

$ nix-env -i cabal-install

Make sure that you have a fairly recent version of cabal installed since these examples will use GHC 8 which requires version 1.24 or later of cabal. You can check what version you have installed by running:

$ cabal --version

Finally, run cabal update if you haven't done so already

Organization

This tutorial is split into several tutorial projects in the project*/ subdirectories. Read the README.md file in each subdirectory in order to follow the tutorial:

haskell-nix's People

Contributors

alexeyzab avatar alogic0 avatar domenkozar avatar gabriella439 avatar hlian avatar novemberkilo avatar ryanglscott avatar tmvector avatar tomsmalley avatar vyacheslavhashov avatar

Watchers

 avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.