/elixir-styler

An @elixir-lang code-style enforcer that will just FIFY instead of complaining

Primary LanguageElixirApache License 2.0Apache-2.0

Styler

Styler is an Elixir formatter plugin that's combination of mix format and mix credo, except instead of telling you what's wrong, it just rewrites the code for you to fit its style rules.

You can learn more about the history, purpose and implementation of Styler from our talk: Styler: Elixir Style-Guide Enforcer @ GigCity Elixir 2023

Installation

Add :styler as a dependency to your project's mix.exs:

def deps do
  [
    {:styler, "~> 0.11", only: [:dev, :test], runtime: false},
  ]
end

Then add Styler as a plugin to your .formatter.exs file

[
  plugins: [Styler]
]

And that's it! Now when you run mix format you'll also get the benefits of Styler's definitely-always-right style fixes.

Configuration

There isn't any! This is intentional.

Styler is @adobe's internal Style Guide Enforcer - allowing exceptions to the styles goes against that ethos. Happily, it's open source and thus yours to do with as you will =)

Features (or as we call them, "Styles")

At this point, Styler does a lot. We've catalogued a list of Credo rules that it automatically fixes, but it does some things - like shrinking function heads down to a single line when possible - that Credo doesn't care about.

Ultimately, the best way to see what Styler does is to just try it out! What could go wrong? (You're using version control, right?)

Credo Rules Styler Replaces

If you're using Credo and Styler, we recommend disabling these rules in .credo.exs to save on unnecessary checks in CI.

Disabling the rules means updating your .credo.exs depending on your configuration:

  • if you're using checks: %{enabled: [...]}, ensure none of the checks are listed in your enabled checks
  • if you're using checks: %{disabled: [...]}, copy/paste the snippet below into the list
  • if you're using checks: [...], copy/paste the snippet below into the list and ensure none of the checks appear earlier in the list
# Styler Rewrites
#
# The following rules are automatically rewritten by Styler and so disabled here to save time
# Some of the rules have `priority: :high`, meaning Credo runs them unless we explicitly disable them
# (removing them from this file wouldn't be enough, the `false` is required)
#
# Some rules have a comment before them explaining ways Styler deviates from the Credo rule.
#
# always expands `A.{B, C}`
{Credo.Check.Consistency.MultiAliasImportRequireUse, false},
# including `case`, `fn` and `with` statements
{Credo.Check.Consistency.ParameterPatternMatching, false},
# Styler implements this rule with a depth of 3 and minimum repetition of 2
{Credo.Check.Design.AliasUsage, false},
{Credo.Check.Readability.AliasOrder, false},
{Credo.Check.Readability.BlockPipe, false},
# goes further than formatter - fixes bad underscores, eg: `100_00` -> `10_000`
{Credo.Check.Readability.LargeNumbers, false},
# adds `@moduledoc false`
{Credo.Check.Readability.ModuleDoc, false},
{Credo.Check.Readability.MultiAlias, false},
{Credo.Check.Readability.OneArityFunctionInPipe, false},
# removes parens
{Credo.Check.Readability.ParenthesesOnZeroArityDefs, false},
{Credo.Check.Readability.PipeIntoAnonymousFunctions, false},
{Credo.Check.Readability.PreferImplicitTry, false},
{Credo.Check.Readability.SinglePipe, false},
# **potentially breaks compilation** - see **Troubleshooting** section below
{Credo.Check.Readability.StrictModuleLayout, false},
{Credo.Check.Readability.StringSigils, false},
{Credo.Check.Readability.UnnecessaryAliasExpansion, false},
{Credo.Check.Readability.WithSingleClause, false},
{Credo.Check.Refactor.CaseTrivialMatches, false},
{Credo.Check.Refactor.CondStatements, false},
# in pipes only
{Credo.Check.Refactor.FilterCount, false},
# in pipes only
{Credo.Check.Refactor.MapInto, false},
# in pipes only
{Credo.Check.Refactor.MapJoin, false},
{Credo.Check.Refactor.NegatedConditionsInUnless, false},
{Credo.Check.Refactor.NegatedConditionsWithElse, false},
# allows ecto's `from
{Credo.Check.Refactor.PipeChainStart, false},
{Credo.Check.Refactor.RedundantWithClauseResult, false},
{Credo.Check.Refactor.UnlessWithElse, false},
{Credo.Check.Refactor.WithClauses, false},

Your first Styling

Speed: Expect the first run to take some time as Styler rewrites violations of styles.

Once styled the first time, future styling formats shouldn't take noticeably more time.

Troubleshooting: Compilation broke due to Module Directive rearrangement

Styler naively moves module attributes, which can break compilation. For now, the only fix is some elbow grease.

Module Attribute dependency

Another common compilation break on the first run is a @moduledoc that depended on another module attribute which was moved below it.

For example, given the following broken code after an initial mix format:

defmodule MyGreatLibrary do
  @moduledoc make_pretty_docs(@library_options)
  use OptionsMagic, my_opts: @library_options

  @library_options [ ... ]
end

You can fix the code by moving the static value outside of the module into a naked variable and then reference it in the module. (Note that use macros need an unquote wrapping the variable!)

Yes, this is a thing you can do in a .ex file =)

library_options = [ ... ]

defmodule MyGreatLibrary do
  @moduledoc make_pretty_docs(library_options)
  use OptionsMagic, my_opts: unquote(library_options)

  @library_options library_options
end

Thanks & Inspiration

This work was inspired by earlier large-scale rewrites of an internal codebase that used the fantastic tool Sourceror.

The initial implementation of Styler used Sourceror, but Sourceror's AST-embedding comment algorithm slows Styler down to the point that it's no longer an appropriate drop-in for mix format.

Still, we're grateful for the inspiration Sourceror provided and the changes to the Elixir AST APIs that it drove.

The AST-Zipper implementation in this project was forked from Sourceror's implementation.

Similarly, this project originated from one-off scripts doing large scale rewrites of an enormous codebase as part of an effort to enable particular Credo rules for that codebase. Credo's tests and implementations were referenced for implementing Styles that took the work the rest of the way. Thanks to Credo & the Elixir community at large for coalescing around many of these Elixir style credos.