Standard - Ruby style guide, linter, and formatter

Tests Ruby Style Guide Gem Version

This gem is a spiritual port of StandardJS and aims to save you (and others!) time in the same three ways:

  • No configuration. The easiest way to enforce consistent style in your project. Just drop it in.
  • Automatically format code. Just run standardrb --fix and say goodbye to messy or inconsistent code.
  • Catch style issues & programmer errors early. Save precious code review time by eliminating back-and-forth between reviewer & contributor.

No decisions to make. It just works. Here's a ⚡ lightning talk ⚡ about it.

Install Standard by adding it to your Gemfile and running bundle install:

gem "standard", group: [:development, :test]

You can then run Standard from the command line with:

$ bundle exec standardrb

And if you'd like, Standard can autocorrect your code by tacking on a --fix flag.

If your project uses Rake, adding require "standard/rake" adds two tasks: standard and standard:fix. In most new projects, we tend to add the fixer variant to our default rake task after our test command, similar to this:

task default: [:test, "standard:fix"]

StandardRB — The Rules

  • 2 spaces – for indentation
  • Double quotes for string literals - because pre-committing to whether you'll need interpolation in a string slows people down
  • 1.9 hash syntax - When all the keys in a hash literal are symbols, Standard enforces Ruby 1.9's {hash: syntax}
  • Braces for single-line blocks - Require {/} for one-line blocks, but allow either braces or do/end for multiline blocks. Like using do/end for multiline blocks? Prefer {/} when chaining? A fan of expressing intent with Jim Weirich's semantic block approach? Standard lets you do you!
  • Leading dots on multi-line method chains - chosen for these reasons.
  • Spaces inside blocks, but not hash literals - In Ruby, the { and } characters do a lot of heavy lifting. To visually distinguish hash literals from blocks, Standard enforces that (like arrays), no leading or trailing spaces be added to pad hashes
  • And a good deal more

If you're familiar with RuboCop, you can look at Standard's current base configuration in config/base.yml. In lieu of a separate changelog file, significant changes to the configuration will be documented as GitHub release notes.

Usage

Once you've installed Standard, you should be able to use the standardrb program. The simplest use case would be checking the style of all Ruby files in the current working directory:

$ bundle exec standardrb
standard: Use Ruby Standard Style (https://github.com/testdouble/standard)
standard: Run `standardrb --fix` to potentially fix one problem.
  /Users/jill/code/cli.rb:31:23: Style/Semicolon: Do not use semicolons to terminate expressions.

You can optionally pass in a directory (or directories) using a glob pattern. Be sure to quote paths containing glob patterns so that they are expanded by standardrb instead of your shell:

$ bundle exec standardrb "lib/**/*.rb" test

Note: by default, StandardRB will look for all *.rb files (and some other files typically associated with Ruby like *.gemspec and Gemfile)

If you want to add Standard to an existing project, but don't want to stop all development until you've fixed every violation in every file, you can create a backlog of to-be-converted files by generating a TODO file:

$ bundle exec standardrb --generate-todo

This will create a .standard_todo.yml that lists all the files that contain errors. When you run Standard in the future, it will ignore these files as if they were listed under the ignore section in the .standard.yml file.

As you refactor your existing project you can remove files from the list. You can also regenerate the TODO file at any time by re-running the above command.

Using with Rake

Standard also ships with Rake tasks. If you're using Rails, these should autoload and be available after installing Standard. Otherwise, just require the tasks in your Rakefile:

require "standard/rake"

Here are the tasks bundled with Standard:

$ rake standard     # equivalent to running `standardrb`
$ rake standard:fix # equivalent to running `standardrb --fix`
$ rake standard:fix_unsafely # equivalent to running `standardrb --fix-unsafely`

You may also pass command line options to Standard's Rake tasks by embedding them in a STANDARDOPTS environment variable (similar to how the Minitest Rake task accepts CLI options in TESTOPTS).

# equivalent to `standardrb --format progress`:
$ rake standard STANDARDOPTS="--format progress"

# equivalent to `standardrb lib "app/**/*"`, to lint just certain paths:
$ rake standard STANDARDOPTS="lib \"app/**/*\""

What you might do if you're clever

If you want or need to configure Standard, there are a handful of options available by creating a .standard.yml file in the root of your project.

Here's an example yaml file with every option set:

fix: true               # default: false
parallel: true          # default: false
format: progress        # default: Standard::Formatter
ruby_version: 2.3.3     # default: RUBY_VERSION
default_ignores: false  # default: true

ignore:                 # default: []
  - 'db/schema.rb'
  - 'vendor/**/*'
  - 'test/**/*':
    - Layout/AlignHash

Note: If you're running Standard in a context where your .standard.yml file cannot be found by ascending the current working directory (i.e., against a temporary file buffer in your editor), you can specify the config location with --config path/to/.standard.yml. (Similarly, for the .standard_todo.yml file, you can specify --todo path/to/.standard_todo.yml.)

What you might do if you're REALLY clever

Because StandardRB is essentially a wrapper on top of RuboCop, it will actually forward the vast majority of CLI and ENV arguments to RuboCop.

You can see a list of RuboCop's CLI flags here.

Why should I use Ruby Standard Style?

(This section will look familiar if you've used StandardJS.)

The beauty of Ruby Standard Style is that it's simple. No one wants to maintain multiple hundred-line style configuration files for every module/project they work on. Enough of this madness!

This gem saves you (and others!) time in four ways:

  • No configuration. The easiest way to enforce consistent style in your project. Just drop it in.
  • Automatically format code. Just run standardrb --fix and say goodbye to messy or inconsistent code.
  • Catch style issues & programmer errors early. Save precious code review time by eliminating back-and-forth between reviewer & contributor.
  • Deliberate pace. We strive to take the hassle of upgrading Rubocop out of each individual team's hands and shoulder it ourselves. We enable about ~20% of new cops and generally choose conservative configurations for them.

Adopting Standard style means ranking the importance of code clarity and community conventions higher than personal style. This might not make sense for 100% of projects and development cultures, however open source can be a hostile place for newbies. Setting up clear, automated contributor expectations makes a project healthier.

Who uses Ruby Standard Style?

Here are a few examples of Ruby Standard-compliant teams & projects:

Does your team use Standard? Add your name to the list!

Is there a readme badge?

Yes! If you use Standard in your project, you can include one of these badges in your readme to let people know that your code is using the StandardRB style.

Ruby Style Guide

[![Ruby Style Guide](https://img.shields.io/badge/code_style-standard-brightgreen.svg)](https://github.com/testdouble/standard)

I disagree with rule X, can you change it?

No. The whole point of Standard is to save you time by avoiding bikeshedding about code style. There are lots of debates online about tabs vs. spaces, etc. that will never be resolved. These debates just distract from getting stuff done. At the end of the day you have to 'just pick something', and that's the whole philosophy of Standard -- it's a bunch of sensible 'just pick something' opinions. Hopefully, users see the value in that over defending their own opinions.

Pro tip: Just use Standard and move on. There are actual real problems that you could spend your time solving! :P

Is there an automatic formatter?

Yes! You can use standardrb --fix to fix most issues automatically.

standardrb --fix is built into standardrb for maximum convenience. Most problems are fixable, but some errors must be fixed manually.

If you're feeling sinister, you can also use standardrb --fix-unsafely, which will also perform fixes, even if they run the risk of altering program behavior. If you read your git diffs closely and have good tests, this is often okay, but YMMV… it's called unsafe for a reason!

Can I override the fix: true config setting?

Also yes! You can use standardrb --no-fix. Not fixing is the default behavior, but this flag will override the fix: true setting in your .standard.yml config. This is especially useful for checking your project's compliance with standardrb in CI environments while keeping the fix: true option enabled locally.

How do I ignore files?

Sometimes you need to ignore additional folders or specific minified files. To do that, add a .standard.yml file to the root of your project and specify a list of files and globs that should be excluded:

ignore:
  - 'some/file/in/particular.rb'
  - 'a/whole/directory/**/*'

You can see the files Standard ignores by default here

How do I hide a certain warning?

In rare cases, you'll need to break a rule and hide the warning generated by Standard.

Ruby Standard Style uses RuboCop under-the-hood and you can hide warnings as you normally would if you used RuboCop directly.

To ignore only certain rules from certain globs (not recommended, but maybe your test suite uses a non-standardable DSL, you can specify an array of RuboCop rules to ignore for a particular glob:

ignore:
  - 'test/**/*':
    - Layout/EndAlignment

How do I disable a warning within my source code?

You can also use special comments to disable all or certain rules within your source code.

Given this source listing foo.rb:

baz = 42

Running standard foo.rb would fail:

foo.rb:1:1: Lint/UselessAssignment: Useless assignment to variable - `baz`.

If we wanted to make an exception, we could add the following comment:

baz = 42 # standard:disable Lint/UselessAssignment

The comment directives (both standard:disable and rubocop:disable) will suppress the error and Standard would succeed.

If, however, you needed to disable standard for multiple lines, you could use open and closing directives like this:

# standard:disable Layout/IndentationWidth
def foo
    123
end
# standard:enable Layout/IndentationWidth

And if you don't know or care which rule is being violated, you can also substitute its name for "all". This line actually triggers three different violations, so we can suppress them like this:

baz = ['a'].each do end # standard:disable all

How do I specify a Ruby version? What is supported?

Because Standard wraps RuboCop, they share the same runtime requirements—currently, that's MRI 2.3 and newer. While Standard can't avoid this runtime requirement, it does allow you to lint codebases that target Ruby versions older than 2.3 by narrowing the ruleset somewhat.

Standard will default to telling RuboCop to target the currently running version of Ruby (by inspecting RUBY_VERSION at runtime. But if you want to lock it down, you can specify ruby_version in .standard.yml.

ruby_version: 1.8.7

See testdouble/suture for an example.

It's a little confusing to consider, but the targeted Ruby version for linting may or may not match the version of the runtime (suppose you're on Ruby 2.5.1, but your library supports Ruby 2.3.0). In this case, specify ruby_version and you should be okay. However, note that if you target a newer Ruby version than the runtime, RuboCop may behave in surprising or inconsistent ways.

If you are targeting a Ruby older than 2.3 and run into an issue, check out Standard's version-specific RuboCop configurations and consider helping out by submitting a pull request if you find a rule that won't work for older Rubies.

How do I use Standard with RuboCop extensions or custom rules?

If you want to use Standard in conjunction with RuboCop extensions or custom cops, you can specify them in your own RuboCop configuration YAML files and .standard.yml using the "extend_config` setting.

For a simple example, you could include rubocop-rails when Standard runs by first specifying a file in .standard.yml:

# .standard.yml

extend_config:
  - .standard_rubocop_extensions.yml

And a minimal RuboCop configuration file:

# .standard_rubocop_extensions.yml

require:
  - rubocop-rails

That's it! Now, in addition to all of Standard's built-in rules, standardrb and rake standard will also execute the default configuration of the rubocop-rails gem without needing to invoke rubocop separately.

For a slightly more complex example, we could add the betterlint gem from our friends at Betterment, first by telling Standard where our configuration file is:

# .standard.yml

extend_config:
  - .betterlint.yml

But if we only wanted to enable a particular rule, we could configure it more narrowly, like this:

# .betterlint.yml

require:
  - rubocop/cop/betterment.rb

AllCops:
  DisabledByDefault: true

Betterment/UnscopedFind:
  Enabled: true

  unauthenticated_models:
    - SystemConfiguration

This same approach works for more than just gems! Just require a Ruby file that defines or loads your custom RuboCop implementation and configure it using extend_config.

When Standard encounters an extend_config property, it will merge your configuration files with Standard's base ruleset. To prevent Standard's built-in rules from being modified, any configuration of rules includued in the rubocop or rubocop-performance gems will be ignored. Most settings under AllCops: can be configured, however, unless they'd conflict with a setting used by Standard (like TargetRubyVersion) or prevent Standard's own rules from running (like StyleGuideCopsOnly). If you specify multiple YAML files under extend_config, note that their resulting RuboCop configurations will be merged in order (i.e. last-in-wins).

Usage via RuboCop

If you find that Standard's extend_config feature doesn't meet your needs or if you only want to use Standard's rules while continuing to use RuboCop's CLI (e.g., to continue using your favorite IDE/tooling/workflow with RuboCop support) Evil Martians also maintains a regularly updated guide on how to configure RuboCop to load and execute Standard's ruleset.

In short, you can configure this in your .rubocop.yml to load Standard's ruleset like any other gem:

require: standard

inherit_gem:
  standard: config/base.yml

How do I change the output?

Standard's built-in formatter is intentionally minimal, printing only unfixed failures or (when successful) printing nothing at all. If you'd like to use a different formatter, you can specify any of RuboCop's built-in formatters or write your own.

For example, if you'd like to see colorful progress dots, you can either run Standard with:

$ bundle exec standardrb --format progress
Inspecting 15 files
...............

15 files inspected, no offenses detected

Or, in your project's .standard.yml file, specify:

format: progress

Refer to RuboCop's documentation on formatters for more information.

How do I run Standard in my editor?

It can be very handy to know about failures while editing to shorten the feedback loop.

Language Server Protocol support

To provide immediate feedback of Standard violations and support autofixing of your code while avoiding the performance cost of starting and stopping the standardrb binary repeatedly, Standard Ruby ships with a built-in Language Server Protocol server, which is powered by the language_server-protocol gem and can be activated from the command line with the --lsp flag.

Most likely, you'd instantiate this server indirectly in your editor's configuration, as can be demonstrated easily with neovim. Theoretically, this feature could be leveraged by a purpose-built editor plugin to performantly format and fix your code. (If you're looking for a project, we'd love to see one created for VS Code!)

Editor-specific guides

Why aren't frozen_string_literal: true magic comments enforced?

Standard does not take a stance on whether you should plaster a frozen_string_literal magic comment directive at the top of every file. Enforcing use of the comment became popular when it was believed that string literals would be frozen by default in a future version of Ruby, but according to Matz there are no (longer any) such plans.

Aside from one's personal opinion on the degree to which the comment is an eyesore, the decision to include the magic comment at the top of every file listing ought to be made based on the performance characteristics of each project (e.g. whether it defines a significant number of string literals, whether the commensurate memory usage is a material constraint, whether the code is run as a one-off command or a long-lived server application). These tend to indicate whether the magic comment might lead to meaningful reductions in memory usage.

Because Standard is intended to be used as a default for every kind of Ruby file—from shell scripts to Rails apps—it wouldn't be appropriate for Standard to either enforce or preclude the magic comment. Instead, you might consider either:

  • Measuring memory performance by enabling frozen string literals as the default at runtime (with RUBYOPT=--enable-frozen-string-literal)
  • Introducing the magic_frozen_string_literal gem to your build, which will automatically ensure that the comment is prepended for every applicable file in your project

How often is Standard updated?

We aim to update Standard once a month, in the first week of the month. In between releases, we'll be considering RuboCop updates, RuboCop Performance updates, and community contributions.

Does Standard work with [Insert other tool name here]?

Maybe! Start by searching the repository to see if there's an existing issue open for the tool you're interested in. That aside, here are other known integrations aside from editor plugins:

Contributing

Follow the steps below to setup standard locally:

$ git clone https://github.com/testdouble/standard
$ cd standard
$ gem install bundler # if working with ruby version below 2.6.0
$ bundle install
$ bundle exec rake # to run test suite

Code of Conduct

This project follows Test Double's code of conduct for all community interactions, including (but not limited to) one-on-one communications, public posts/comments, code reviews, pull requests, and GitHub issues. If violations occur, Test Double will take any action they deem appropriate for the infraction, up to and including blocking a user from the organization's repositories.