Skip to content

Latest commit

 

History

History
572 lines (437 loc) · 21.9 KB

README.md

File metadata and controls

572 lines (437 loc) · 21.9 KB
Raw

Tests Ruby Style Guide Gem Version

The standard gem brings the ethos of StandardJS to Ruby. It's a linter & formatter built on RuboCop and provides an unconfigurable configuration to all of RuboCop's built-in rules as well as those included in rubocop-performance. It also supports plugins built with lint_roller, like standard-rails and standard-sorbet.

Standard Ruby was created and is maintained by the team at Test Double, because we appreciate the importance of balancing predictable, consistent code with preserving developer autonomy. 💚

Topics covered in this README:

Wait, did you say unconfigurable configuration?

Yes, Standard is unconfigurable. See, pretty much every developer can agree that automatically identifying and fixing slow, insecure, and error-prone code is a good idea. People also agree it's easier to work in codebases that exhibit a consistent style and format. So, what's the problem? No two developers will ever agree on what all the rules and format should be.

This has resulted in innumerable teams arguing about how to configure their linters and formatters over literal decades. Some teams routinely divert time and energy from whatever they're building to reach consensus on where commas should go. Other teams have an overzealous tech lead who sets up everything his favorite way and then imposes his will on others. It's not uncommon to witness passive-aggressive programmers change a contentious rule back-and-forth, resulting in both acrimony and git churn (and acrimony about git churn). Still, most developers give way to whoever cares more, often resulting in an inconsistent configuration that nobody understands and isn't kept up-to-date with changes to their linter and its target language. Whatever the approach, time spent bikeshedding is time wasted.

But you can't configure Standard Ruby. You can take it or leave it. If this seems presumptive, constraining, and brusque: you're right, it usually does feel that way at first. But since 2018, the Ruby community has debated Standard Ruby's ruleset, ultimately landing us on a sensible set of defaults that—while no one person's favorite way to format Ruby—seems to be good enough for most of the ways people use Ruby, most of the time.

If you adopt Standard Ruby, what's in it for you? Time saved that would've been spent arguing over how to set things up. Also, seamless upgrades: we stay on top of each new RuboCop rule and update our configuration on a monthly release cadence. But the best part of adopting Standard as your linter and formatter? You'll spend a whole lot less time thinking about linters and formatters.

So please, give Standard Ruby a try. If you're like these folks, you'll soon realize that the value of a linter is in using one at all and not in the particulars of how it's configured.

Usage

Install

Getting started is as easy as gem install standard or throwing it in your project's Gemfile and running bundle install:

gem "standard"

Running Standard Ruby

Once installed, you can run Standard from the command line via its built-in executable or as a Rake task.

Standard Ruby's binary is named standardrb to distinguish it from StandardJS:

$ standardrb

And the Rake task can be run if your Rakefile includes require "standard/rake". This will load the standard task, allowing you to run:

$ rake standard

Either way, Standard will inspect any Ruby files found in the current directory tree. If any errors are found, it'll report them. If your code is standard-compliant, it will get out of your way by quietly exiting code 0.

Fixing errors

A majority of Standard's rules can be safely fixed automatically.

# CLI
$ standardrb --fix

# Rake
$ rake standard:fix

Because we're confident Standard's fixes won't change the behavior of our code, we typically run with fix enabled all the time because it saves us from having to look at and think about problems the computer can solve for us.

Applying unsafe fixes

There are a number of other rules that can be automatically remedied, but not without the risk of changing your code's behavior. While we wouldn't recommend running it all the time, if the CLI suggests that additional errors can be fixed unsafely, here's how to do that:

# CLI
$ standardrb --fix-unsafely

# Rake
$ rake standard:fix_unsafely

So long as your code is checked into source control, there's no mortal harm in running with unsafe fixes enabled. If the changes look good to you and your tests pass, then it's probably no more error prone than manually editing every change by hand.

Integrating Standard into your workflow

Because code formatting is so integral to programming productivity and linter violations risk becoming bugs if released into production, tools like Standard Ruby are only as useful as their integration into code editors and continuous integration systems.

Editor support

We've added a number of editing guides for getting started:

If you'd like to help by creating a guide, please draft one in an issue and we'll get it added!

Language Server Protocol support

If you don't see your preferred editor above, Standard Ruby also offers robust language server support, in two ways, both included with the standard gem:

  1. A Ruby LSP add-on, which (if the standard gem is in your Gemfile) will be loaded automatically
  2. A language server executable, which can be run from the command line with standardrb --lsp
Capability Ruby LSP Add-on Internal Server
Diagnostics (Linting) ✅ (Pull) ✅ (Push)
Formatting ✅ ✅
Code Actions ✅ ❌
Everything else ❌ ❌

Due to the advantages of Pull diagnostics and its inclusion of Code Actions ("Quick Fixes" in VS Code parlance), we recommend using Standard's Ruby LSP add-on where possible.

Regardless of which language server you use, many editors have added LSP support, with each bringing their own approach to configuration. Many LSP features should "just work", but when in doubt, please consult our editor guides above as well as your editor's own documentation on how to configure LSP formatters and linters.

CI support

Various continuous integration and quality-checking tools have been made to support Standard Ruby, as well.

Of course, no special plugin is necessary to run Standard Ruby in a CI environment, as standardrb and rake standard work just fine on their own!

Other tools

These tool integrations are also available:

Ignoring errors

While Standard is very strict in how each formatting and linting rule is configured, it's mercifully flexible when you need to ignore a violation to focus on a higher priority (like, say, keeping the build running). There are a number of ways to ignore any errors, with the right answer depending on the situation.

Ignoring a line with a comment

Individual lines can be ignored with a comment directive at the end of the line. As an example, the line text = 'hi' violates two rules, Lint/UselessAssignment and Style/StringLiterals.

You could ignore one, both, or all errors with the following comments:

# Disable one error but keep Lint/UselessAssignment
text = 'hi' # standard:disable Style/StringLiterals

# Disable both errors explicitly
text = 'hi' # standard:disable Style/StringLiterals, Lint/UselessAssignment

# Disable all errors on the line with "all"
text = "hi" # standard:disable all

Ignoring multiple lines with comments

Similarly to individual lines, you can also disable multiple lines by wrapping them in comments that disable and re-enable them:

# standard:disable Style/StringLiterals, Lint/UselessAssignment
text = "hi"
puts 'bye'
# standard:enable Style/StringLiterals, Lint/UselessAssignment

Ignoring entire files and globs

You can ignore entire files and file patterns by adding them to ignore: in your project's .standard.yml file:

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

Ignoring specific rules in files and globs

For a given file or glob, you can even ignore only specific rules by nesting an array of the rules you'd like to ignore:

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

Ignoring every violation and converting them into a todo list

If you're adopting Standard in a large codebase and you don't want to convert everything all at once, you can work incrementally by generating a "todo" file which will cause Standard to ignore every violation present in each file of the codebase.

This way, you can gradually work through the todo list, removing ignore directives and fixing any associated errors, while also being alerted to any regressions if they're introduced into the project.

Here are the commands you might run to get started:

# Start by clearing any auto-fixable errors:
$ standardrb --fix

# Generate a `.standard_todo.yml' to work from
$ standardrb --generate-todo

Configuring Standard

While the rules aren't configurable, Standard offers a number of options that can be configured as CLI flags and YAML properties.

CLI flags

The easiest way to summarize the available CLI flags is to invoke standardrb -h:

Usage: standardrb [--fix] [--lsp] [-vh] [--format <name>] [--] [FILE]...

Options:

  --fix             Apply automatic fixes that we're confident won't break your code
  --fix-unsafely    Apply even more fixes, including some that may change code behavior
  --no-fix          Do not automatically fix failures
  --format <name>   Format output with any RuboCop formatter (e.g. "json")
  --generate-todo   Create a .standard_todo.yml that lists all the files that contain errors
  --lsp             Start a LSP server listening on STDIN
  -v, --version     Print the version of Standard
  -V, --verbose-version   Print the version of Standard and its dependencies.
  -h, --help        Print this message
  FILE              Files to lint [default: ./]

Standard also forwards most CLI arguments to RuboCop. To see them, run:

  $ rubocop --help

If you run Standard via Rake, you can specify your CLI flags in an environment variable named STANDARDOPTS like so:

$ rake standard STANDARDOPTS="--format progress"

YAML options

In addition to CLI flags, Standard will search for a .standard.yml file (ascending to parent directories if the current working directory doesn't contain one). If you find yourself repeatedly running Standard with the same CLI options, it probably makes more sense to set it once in a YAML file:

fix: true               # default: false
parallel: true          # default: false
format: progress        # default: Standard::Formatter
ruby_version: 3.0       # default: RUBY_VERSION
default_ignores: false  # default: true

ignore:                 # default: []
  - 'vendor/**/*'

plugins:                # default: []
  - standard-rails

extend_config:                # default: []
  - .standard_ext.yml

Configuring ruby_version

One notable YAML setting is ruby_version, which allows you to set the oldest version of Ruby the project needs to support RuboCop's TargetRubyVersion setting explicitly. Because this value is inferred from any .ruby-version, .tool-versions, Gemfile.lock, and *.gemspec files that might be present, most applications won't need to set this.

However, gems and libraries that support older versions of Ruby will want to lock the ruby-version: explicitly in their .standard.yml file to ensure new rules don't break old rubies:

ruby_version: 2.7

Extending Standard

Standard includes two extension mechanisms: plugins and configuration extensions. While neither can change the rules configured out-of-the-box by Standard, they can define, require, and configure additional RuboCop rules.

Both are "first-in-wins", meaning once a rule is configured by a plugin or extension, it can't be changed or reconfigured by a later plugin or extension. This way, each Standard plugin you add becomes a de facto "standard" of its own. Plugins have precedence over extensions as they are loaded first.

Plugins

Adding a plugin to your project is as easy as adding it to your Gemfile and specifying it in .standard.yml in the root of your project. For example, after installing standard-rails, you can configure it by adding it to plugins:

plugins:
  - standard-rails

Each plugin can be passed configuration options by specifying them in a nested hash. For example, standard-rails allows you to configure its rules for a particular version of Rails (though this will usually be detected for you automatically):

plugins:
  - standard-rails:
      target_rails_version: 7.0

You can develop your own plugins, too! Check out the lint_roller gem to learn how. For a simple example, you can look at standard-custom, which is one of the default plugins included by Standard.

Extended config files

Of course, you may want to add more rules without going to the trouble of packaging them in a plugin gem. For cases like this, you can define one or more RuboCop configuration files and then list them in your .standard.yml file under extend_config.

For example, after installing the betterlint gem from our friends at Betterment, we could create a RuboCop config file named .betterlint.yml:

require:
  - rubocop/cop/betterment.rb

Betterment/UnscopedFind:
  Enabled: true

  unauthenticated_models:
    - SystemConfiguration

And then reference it in our .standard.yml:

extend_config:
  - .betterlint.yml

Running Standard's rules via RuboCop

Please note that the following usage is not supported and may break at any time. Use at your own risk and please refrain from opening GitHub issues with respect to loading Standard or its plugins' YAML configurations for use by the rubocop CLI.

If you find that neither plugins or configuration extensions meet your needs or if you have some other reason to run Standard's rules with 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 three default rulesets, just as you would any other gem:

require:
  - standard
  - standard-custom
  - standard-performance
  - rubocop-performance

inherit_gem:
  standard: config/base.yml
  standard-custom: config/base.yml
  standard-performance: config/base.yml

Who uses Standard Ruby?

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

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

If you really want to show off, you can also add a badge to your project's README, like this one:

Ruby Code Style

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

Help, I'm seeing an install error!

This section is really only here to feed Google, but if you see an error like any of the following:

ERROR:  Could not find a valid gem 'standard' (= 0.0.36) in any repository

Could not find gem 'standard (= 0.0.36)' in rubygems repository https://rubygems.org/ or installed locally.

Your bundle is locked to standard (0.0.36) from rubygems repository https://rubygems.org/ or installed locally, but that version can no longer be found in that source. That means the author of standard (0.0.36) has removed it. You'll need to update your bundle to a version other than standard (0.0.36) that hasn't been removed in order to install.

This is because on August 18th, 2023, we yanked versions 0.0.1~0.0.36.1 from RubyGems.org for the reasons discussed in this issue. Because these versions are now over four years old and no longer track supported versions of Ruby or RuboCop, the correct fix for any of the above errors is probably to upgrade to the latest version of Standard Ruby.

If for whatever reason you need to install one of these versions, you can change your Gemfile to point to the corresponding git tag from the source repository:

gem "standard", git: "https://github.com/testdouble/standard.git", tag: "v0.0.36"

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.