Trying the Rails or Lint options in RuboCop

We’d like to introduce you a static analytical tool for Ruby called RuboCop.

You can use the automatic code review function of RuboCop even in Sider now. This is a good opportunity for you to try out RuboCop, why not!?

This introduces RuboCop, RuboCop’s Rails option, lint option which detects the contents that cause bugs, and other options which are useful for other than rules, please take your time to read!

What is RuboCop
Benefits for using RuboCop
RuboCop installation
RuboCop basics
-R(--rails) option
-l(--lint) option
How to write .rubocop.yml file
How to use .rubocop_todo.yml
-a(--auto-correct) option
Example of .rubocop.yml

What is RuboCop

RuboCop is a static code analyzing tool to check whether the ruby code in your project “is written as per the coding rules”.

The default coding rules are based on Ruby style guide.
(Japanese version)

By editing the configuration file (.rubocop.yml), you can add or delete your own coding rules.

Benefits for using RuboCop

The setting of Rubocop is collated in the .rubocop.yml file, then it clearly indicates the project coding rules.

By using Rubocop to keep the codes clean, you’ll have less mistakes and it leads to effective review.

RuboCop installation

RuboCop is provided via gem.

When you use Bundler, add below into the Gemfile.

gem 'rubocop', require: false

In case of web applications such as rails,
Since most of the case they are executed in the development environment, followings are sufficient.

group 'development' do
  ...
  gem 'rubocop', require: false
end

You can install

$ bundle

after editing the GemFile.

In case you don’t use Bunfler, you can install it in $ gem install rubocop.

RuboCop basics

The description prior to this part has been verified by using the following versions.

ruby 2.2.0p0 (2014–12–25 revision 49005) [x86_64-darwin14]
rubocop 0.29.1

`rubocop` commands

Execute commands are as follows.
In case of using Bundler, kindly add bundle exec in the top.

$ rubocop [options] [file1, file2, ...]

In order to verify all ruby files in the project, simply follow the below.

$ rubocop

To refer to the Help, kindly follow the below.

$ rubocop --help

`rubocop` execution – interpretation of output results

For example, prepare a ruby file (test.rb) as shown below.

def badName
  if something
    test
    end
end

Execute RuboCop for this test.rb.

$ rubocop test.rb

Output would be as follows.

Inspecting 1 file
W

Offenses:

test.rb:1:5: C: Use snake_case for method names.
def badName
    ^^^^^^^
test.rb:2:3: C: Use a guard clause instead of wrapping the code inside a conditional expression.
  if something
  ^^
test.rb:2:3: C: Favor modifier if usage when having a single-line body. Another good alternative is the usage of control flow &&/||.
  if something
  ^^
test.rb:4:5: W: end at 4, 4 is not aligned with if at 2, 2
    end
    ^^^

1 file inspected, 4 offenses detected

The output above can be interpreted as follows.

Inspecting 1 file in the first line indicates the number of verified ruby files.
The W in the second line indicates the first letter of the most severe violation (sort by Convention, Warning, Error, Fatal severity levels ) that was found in the verified ruby files.
4 violations have been found in this example, the most severe violation was Warning. Hence W has been output..
For the 3rd line and later show detailed location for respective violation (file name: row: column) and description of the violations.
The last line shows the number of verified files and the number of violations

Let’s have a look the output when all the violations have been fixed.

def good_name
  test if something
end

$ rubocop test.rb
Inspecting 1 file
.

1 file inspected, no offenses detected

no offenses detected , it means coding violation is now 0.

`R(--rails)` option

For rails projects, by adding -R option for execution, it executes Rails Cop(special Cop in rails) additionally.
There are no coding violations found in the following ruby( test.rb ) file using the default RuboCop,

def good_name
  puts 'test' if something
end

but by placing it under app/models of rails and executing it with the -R option, it shows

$ rubocop -R app/models/test.rb
Inspecting 1 file
C

Offenses:

app/models/test.rb:2:3: C: Do not write to stdout. Use Rails' logger if you want to log.
  print 'test' if something
  ^^^^^

1 file inspected, 1 offense detected

As you see, it outputs Rails specific violation “Use Rails logger rather than the standard output”.
Rails Cop is defined in gem’s lib/rubocop/cop/ below rails .
Moreover, by setting the followings in .rubocop.yml file,

AllCops:
 RunRailsCops: true

you can keep Rails Cop enabled which is same as specifying the -R option.

l( — lint) option

The verification rules (Cop) in RuboCop are classified into below 4 items.

Style (Cop for style)
Lint (Cop for high error possibility area or bad practice)
Metrics (Cop for the number of class row or the number of characters in a line)
Rails (Special Cop for Rails)

The -l option only Lint.

How to write .rubocop.yml file

By placing the configuration file .rubocop.yml in the root directory of the project, RuboCop will read this automatically. Or it can also specify a yml file in any location as configuration file by -c option.
Below are the representative codes in .rubocop.yml file.

Specify the file to exclude from RuboCop

Use Exclude: to exclude files like (db/schema.rb) which is automatically generated by Rails,
or gem placed under vendor/bundle from RuboCop.

For example, use the followings to exclude the db/schema.rb file from RuboCop.

AllCops:
  Exclude:
    - db/schema.rb

However, by doing this, Exclude which is configured by default will be disabled.
Hence if you would like to continue to use 'vendor/**/*' which is default exclude, you can write as below.

AllCops:
  Exclude:
    - db/schema.rb
    - 'vendor/**/*'

Reference: The one important thing to use rubocop gem

Enabling/Disabling Cop (Enabled: false/true)

When you feel a certain cop isn’t suitable for the project, you can disable it individually.
For example, there is a Cop named Style/StringLiterals .
This rule indicates that 「Always use single quotes for the literals of string that variables expansion is not required」.
To disable this, set Enabled: false after the Cop name.

Style/TrailingComma:
  Enabled: false

Setting example for Cop with option

Some Cop have configurable parameter ( Configuration parameters ) .
For instance, the previous Style/StringLiterals Cop has 2 options for the literals of String that doesn’t require variable expansion. One is always use single quotes (default), or always use double quotes.

EnforcedStyle: single_quotes # default
EnforcedStyle: double_quotes

Although the default is single_quotes , if you prefer double_quotes you can set as shown below.

Style/StringLiterals:
  EnforcedStyle: double_quotes

Other settings

There are other default rules in the config/default.yml(github source)
of the rubocop gem..

How to use .rubocop_todo.yml

If you have too many violations that need to be fixed, there is a method to generate the .rubocop_todo.yml file and use it as a TODO list.

.rubocop_todo.ymlcan be easily generated by using --auto-gen-configoption.
Recreate the test.rb using the following.

def badName
  if something
    test
    end
end

Then, execute rubocop with the --auto-gen-config option.

$ rubocop --auto-gen-config test.rb

It generates the file .rubocop_todo.yml Contents will be as follows.

$ cat .rubocop_todo.yml

# This configuration was generated by `rubocop --auto-gen-config`
# on 2015-03-09 19:07:44 +0900 using RuboCop version 0.29.1.
# The point is for the user to remove these configuration records
# one by one as the offenses are removed from the code base.
# Note that changes in the inspected code, or installation of new
# versions of RuboCop, may require this file to be generated again.

# Offense count: 1
# Configuration parameters: SupportedStyles.
Lint/EndAlignment:
  AlignWith: variable

# Offense count: 1
# Configuration parameters: MinBodyLength.
Style/GuardClause:
  Enabled: false

....

This yml file has the same format as the .rubocop.yml file, and it outputs the configuration which disables the violating Cops.
Call the generated file from .rubocop.yml

inherit_from: .rubocop_todo.yml

Based on this
You can distinguish

「rules to follow」 which is .rubocop.ymlfile

and

「violations to be fixed」 which is .rubocop_todo.yml file.

When you start fixing the violations, you better fix by deleting items in .rubocop_todo.yml one by one.

-a( — auto-correct) option

For some violations, by executing rubocop by adding -a option, the codes will fix automatically.

Cop supports --auto-correct is written on items that support auto-correct located in the .rubocop_todo.yml which is generated with the --auto-gen-config option.
However, this function is in the experimental stage, you need to take care.

.rubocop.yml example

Here is the examples of open source .rubocop.yml

BBC-News/rubocop-config
https://github.com/BBC-News/rubocop-config/blob/master/.rubocop.yml
Configuration file which is uploaded by the British Broadcasting Corporation BBC in GitHub.
thoughtbot/hound
https://github.com/thoughtbot/hound/blob/master/config/style_guides/ruby.yml
This is a configuration file committed in the Rails application of the automated review service 「hound」 operated by the thoughtbot company.
thoughtbot also publishes style guide and it might also be interesting to compare with this.
spree/spree
https://github.com/spree/spree/blob/master/.rubocop.yml
spree is a Rails application for open source EC site.

Summary

You can use RuboCop via SideCI. When you open or update a Pull Request in GitHub, RuboCop will be automatically run in SideCI and it will provide comments in the Pull Request.
If you do not customize the .rubocop.yml file, the RuboCop standard setting will be used to perform the analysis. Moreover, if the project is Rails, SideCI will determine whether this is Rails or not automatically and execute RuboCop by -R option for Rails.

Please have a try our SideCI.