Cognitive Complexity

Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

A method's cognitive complexity is based on a few simple rules:

• Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
• Code is considered more complex for each "break in the linear flow of the code"
• Code is considered more complex when "flow breaking structures are nested"

Further reading

• Cognitive Complexity docs
• Cognitive Complexity: A new way of measuring understandability

Cognitive Complexity

Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

A method's cognitive complexity is based on a few simple rules:

• Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
• Code is considered more complex for each "break in the linear flow of the code"
• Code is considered more complex when "flow breaking structures are nested"

Further reading

• Cognitive Complexity docs
• Cognitive Complexity: A new way of measuring understandability

Cognitive Complexity

Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

A method's cognitive complexity is based on a few simple rules:

• Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
• Code is considered more complex for each "break in the linear flow of the code"
• Code is considered more complex when "flow breaking structures are nested"

Further reading

• Cognitive Complexity docs
• Cognitive Complexity: A new way of measuring understandability

Cognitive Complexity

Cognitive Complexity is a measure of how difficult a unit of code is to intuitively understand. Unlike Cyclomatic Complexity, which determines how difficult your code will be to test, Cognitive Complexity tells you how difficult your code will be to read and comprehend.

A method's cognitive complexity is based on a few simple rules:

• Code is not considered more complex when it uses shorthand that the language provides for collapsing multiple statements into one
• Code is considered more complex for each "break in the linear flow of the code"
• Code is considered more complex when "flow breaking structures are nested"

Further reading

• Cognitive Complexity docs
• Cognitive Complexity: A new way of measuring understandability

This cop makes sure that predicates are named properly.

Example:

# bad
def is_even(value)
end

def is_even?(value)
end

# good
def even?(value)
end

# bad
def has_value
end

def has_value?
end

# good
def value?
end

This cop enforces empty line after guard clause

Example:

# bad
def foo
  return if need_return?
  bar
end

# good
def foo
  return if need_return?

  bar
end

# good
def foo
  return if something?
  return if something_different?

  bar
end

# also good
def foo
  if something?
    do_something
    return if need_return?
  end
end

This cop enforces empty line after guard clause

Example:

# bad
def foo
  return if need_return?
  bar
end

# good
def foo
  return if need_return?

  bar
end

# good
def foo
  return if something?
  return if something_different?

  bar
end

# also good
def foo
  if something?
    do_something
    return if need_return?
  end
end

This cop enforces the use of $stdout/$stderr/$stdin instead of STDOUT/STDERR/STDIN. STDOUT/STDERR/STDIN are constants, and while you can actually reassign (possibly to redirect some stream) constants in Ruby, you'll get an interpreter warning if you do so.

Safety:

Autocorrection is unsafe because STDOUT and $stdout may point to different objects, for example.

Example:

# bad
STDOUT.puts('hello')

hash = { out: STDOUT, key: value }

def m(out = STDOUT)
  out.puts('hello')
end

# good
$stdout.puts('hello')

hash = { out: $stdout, key: value }

def m(out = $stdout)
  out.puts('hello')
end

Checks for if and unless statements that would fit on one line if written as modifier if/unless. The cop also checks for modifier if/unless lines that exceed the maximum line length.

The maximum line length is configured in the Layout/LineLength cop. The tab size is configured in the IndentationWidth of the Layout/IndentationStyle cop.

Example:

# bad
if condition
  do_stuff(bar)
end

unless qux.empty?
  Foo.do_something
end

do_something_with_a_long_name(arg) if long_condition_that_prevents_code_fit_on_single_line

# good
do_stuff(bar) if condition
Foo.do_something unless qux.empty?

if long_condition_that_prevents_code_fit_on_single_line
  do_something_with_a_long_name(arg)
end

if short_condition # a long comment that makes it too long if it were just a single line
  do_something
end

This cop enforces the use of $stdout/$stderr/$stdin instead of STDOUT/STDERR/STDIN. STDOUT/STDERR/STDIN are constants, and while you can actually reassign (possibly to redirect some stream) constants in Ruby, you'll get an interpreter warning if you do so.

Safety:

Autocorrection is unsafe because STDOUT and $stdout may point to different objects, for example.

Example:

# bad
STDOUT.puts('hello')

hash = { out: STDOUT, key: value }

def m(out = STDOUT)
  out.puts('hello')
end

# good
$stdout.puts('hello')

hash = { out: $stdout, key: value }

def m(out = $stdout)
  out.puts('hello')
end

Checks for if and unless statements that would fit on one line if written as modifier if/unless. The cop also checks for modifier if/unless lines that exceed the maximum line length.

The maximum line length is configured in the Layout/LineLength cop. The tab size is configured in the IndentationWidth of the Layout/IndentationStyle cop.

Example:

# bad
if condition
  do_stuff(bar)
end

unless qux.empty?
  Foo.do_something
end

do_something_with_a_long_name(arg) if long_condition_that_prevents_code_fit_on_single_line

# good
do_stuff(bar) if condition
Foo.do_something unless qux.empty?

if long_condition_that_prevents_code_fit_on_single_line
  do_something_with_a_long_name(arg)
end

if short_condition # a long comment that makes it too long if it were just a single line
  do_something
end

This cop enforces the use of $stdout/$stderr/$stdin instead of STDOUT/STDERR/STDIN. STDOUT/STDERR/STDIN are constants, and while you can actually reassign (possibly to redirect some stream) constants in Ruby, you'll get an interpreter warning if you do so.

Safety:

Autocorrection is unsafe because STDOUT and $stdout may point to different objects, for example.

Example:

# bad
STDOUT.puts('hello')

hash = { out: STDOUT, key: value }

def m(out = STDOUT)
  out.puts('hello')
end

# good
$stdout.puts('hello')

hash = { out: $stdout, key: value }

def m(out = $stdout)
  out.puts('hello')
end

Checks for if and unless statements that would fit on one line if written as modifier if/unless. The cop also checks for modifier if/unless lines that exceed the maximum line length.

The maximum line length is configured in the Layout/LineLength cop. The tab size is configured in the IndentationWidth of the Layout/IndentationStyle cop.

Example:

# bad
if condition
  do_stuff(bar)
end

unless qux.empty?
  Foo.do_something
end

do_something_with_a_long_name(arg) if long_condition_that_prevents_code_fit_on_single_line

# good
do_stuff(bar) if condition
Foo.do_something unless qux.empty?

if long_condition_that_prevents_code_fit_on_single_line
  do_something_with_a_long_name(arg)
end

if short_condition # a long comment that makes it too long if it were just a single line
  do_something
end

This cop checks for uses of fail and raise.

Example: EnforcedStyle: only_raise (default)

# The `only_raise` style enforces the sole use of `raise`.
# bad
begin
  fail
rescue Exception
  # handle it
end

def watch_out
  fail
rescue Exception
  # handle it
end

Kernel.fail

# good
begin
  raise
rescue Exception
  # handle it
end

def watch_out
  raise
rescue Exception
  # handle it
end

Kernel.raise

Example: EnforcedStyle: only_fail

# The `only_fail` style enforces the sole use of `fail`.
# bad
begin
  raise
rescue Exception
  # handle it
end

def watch_out
  raise
rescue Exception
  # handle it
end

Kernel.raise

# good
begin
  fail
rescue Exception
  # handle it
end

def watch_out
  fail
rescue Exception
  # handle it
end

Kernel.fail

Example: EnforcedStyle: semantic

# The `semantic` style enforces the use of `fail` to signal an
# exception, then will use `raise` to trigger an offense after
# it has been rescued.
# bad
begin
  raise
rescue Exception
  # handle it
end

def watch_out
  # Error thrown
rescue Exception
  fail
end

Kernel.fail
Kernel.raise

# good
begin
  fail
rescue Exception
  # handle it
end

def watch_out
  fail
rescue Exception
  raise 'Preferably with descriptive message'
end

explicit_receiver.fail
explicit_receiver.raise

This cop enforces empty line after guard clause

Example:

# bad
def foo
  return if need_return?
  bar
end

# good
def foo
  return if need_return?

  bar
end

# good
def foo
  return if something?
  return if something_different?

  bar
end

# also good
def foo
  if something?
    do_something
    return if need_return?
  end
end

This cop transforms usages of a method call safeguarded by a non nil check for the variable whose method is being called to safe navigation (&.). If there is a method chain, all of the methods in the chain need to be checked for safety, and all of the methods will need to be changed to use safe navigation. We have limited the cop to not register an offense for method chains that exceed 2 methods.

The default for ConvertCodeThatCanStartToReturnNil is false. When configured to true, this will check for code in the format !foo.nil? && foo.bar. As it is written, the return of this code is limited to false and whatever the return of the method is. If this is converted to safe navigation, foo&.bar can start returning nil as well as what the method returns.

Safety:

Autocorrection is unsafe because if a value is false, the resulting code will have different behaviour or raise an error.

x = false
x && x.foo  # return false
x&.foo      # raises NoMethodError

Example:

# bad
foo.bar if foo
foo.bar.baz if foo
foo.bar(param1, param2) if foo
foo.bar { |e| e.something } if foo
foo.bar(param) { |e| e.something } if foo

foo.bar if !foo.nil?
foo.bar unless !foo
foo.bar unless foo.nil?

foo && foo.bar
foo && foo.bar.baz
foo && foo.bar(param1, param2)
foo && foo.bar { |e| e.something }
foo && foo.bar(param) { |e| e.something }

# good
foo&.bar
foo&.bar&.baz
foo&.bar(param1, param2)
foo&.bar { |e| e.something }
foo&.bar(param) { |e| e.something }
foo && foo.bar.baz.qux # method chain with more than 2 methods
foo && foo.nil? # method that `nil` responds to

# Method calls that do not use `.`
foo && foo < bar
foo < bar if foo

# When checking `foo&.empty?` in a conditional, `foo` being `nil` will actually
# do the opposite of what the author intends.
foo && foo.empty?

# This could start returning `nil` as well as the return of the method
foo.nil? || foo.bar
!foo || foo.bar

# Methods that are used on assignment, arithmetic operation or
# comparison should not be converted to use safe navigation
foo.baz = bar if foo
foo.baz + bar if foo
foo.bar > 2 if foo

This cop checks for uses of rescue in its modifier form.

The cop to check rescue in its modifier form is added for following reasons:

• The syntax of modifier form rescue can be misleading because it might led us to believe that rescue handles the given exception but it actually rescue all exceptions to return the given rescue block. In this case, value returned by handle_error or SomeException.

• Modifier form rescue would rescue all the exceptions. It would silently skip all exception or errors and handle the error. Example: If NoMethodError is raised, modifier form rescue would handle the exception.

Example:

# bad
some_method rescue handle_error

# bad
some_method rescue SomeException

# good
begin
  some_method
rescue
  handle_error
end

# good
begin
  some_method
rescue SomeException
  handle_error
end