View Source Simple (Single Node) Styles

Function Performance & Readability Optimizations

Optimizing for either performance or readability, probably both! These apply to the piped versions as well

Strings to Sigils

This addresses Credo.Check.Readability.StringSigils. This is not configurable.

Rewrites strings with 4 or more escaped quotes to string sigils with an alternative delimiter. The delimiter will be one of " ( { | [ ' < /, chosen by which would require the fewest escapes, and otherwise preferred in the order listed.

# Before
"{\"errors\":[\"Not Authorized\"]}"
# Styled
~s({"errors":["Not Authorized"]})

Large Base 10 Numbers

This addresses Credo.Check.Readability.LargeNumbers. Quokka will respect the :only_greater_than Credo opt.

Style base 10 numbers with 5 or more digits to have a _ every three digits. Formatter already does this except it doesn't rewrite "typos" like 100_000_0.

If you're concerned that this breaks your team's formatting for things like "cents" (like "$100" being written as 100_00), consider using a library made for denoting currencies rather than raw elixir integers.

BeforeAfter
1000010_000
1_0_0_0_010_000 (elixir's formatter leaves the former as-is)
-543213-543_213
123456789123_456_789
55333.2255_333.22
-123456728.0001-123_456_728.0001

Efficient Function Rewrites

All these rewrites are configurable by setting the :inefficient_function_rewrites option in your .formatter.exs file. See the README for more information.

Enum.into -> X.new

This rewrite is applied when the collectable is a new map, keyword list, or mapset via Enum.into/2,3.

This is an improvement for the reader, who gets a more natural language expression: "make a new map from enum" vs "enumerate enum and collect its elements into a new map"

Note that all of the examples below also apply to pipes (enum |> Enum.into(...))

BeforeAfter
Enum.into(enum, %{})Map.new(enum)
Enum.into(enum, Map.new())Map.new(enum)
Enum.into(enum, Keyword.new())Keyword.new(enum)
Enum.into(enum, MapSet.new())Keyword.new(enum)
Enum.into(enum, %{}, fn x -> {x, x} end)Map.new(enum, fn x -> {x, x} end)
Enum.into(enum, [])Enum.to_list(enum)
Enum.into(enum, [], mapper)Enum.map(enum, mapper)

Map/Keyword.merge w/ single key literal -> X.put

Keyword.merge and Map.merge called with a literal map or keyword argument with a single key are rewritten to the equivalent put, a cognitively simpler function.

# Before
Keyword.merge(kw, [key: :value])
# Styled
Keyword.put(kw, :key, :value)

# Before
Map.merge(map, %{key: :value})
# Styled
Map.put(map, :key, :value)

# Before
Map.merge(map, %{key => value})
# Styled
Map.put(map, key, value)

# Before
map |> Map.merge(%{key: value}) |> foo()
# Styled
map |> Map.put(:key, value) |> foo()

Map/Keyword.drop w/ single key -> X.delete

In the same vein as the merge style above, [Map|Keyword].drop/2 with a single key to drop are rewritten to use delete/2

# Before
Map.drop(map, [key])
# Styled
Map.delete(map, key)

# Before
Keyword.drop(kw, [key])
# Styled
Keyword.delete(kw, key)

Enum.reverse/1 and concatenation -> Enum.reverse/2

Enum.reverse/2 optimizes a two-step reverse and concatenation into a single step.

# Before
Enum.reverse(foo) ++ bar
# Styled
Enum.reverse(foo, bar)

# Before
baz |> Enum.reverse() |> Enum.concat(bop)
# Styled
Enum.reverse(baz, bop)

Timex.now/0 ->DateTime.utc_now/0

Timex certainly has its uses, but knowing what stdlib date/time struct is returned by now/0 is a bit difficult!

We prefer calling the actual function rather than its rename in Timex, helping the reader by being more explicit.

This also hews to our internal styleguide's "Don't make one-line helper functions" guidance.

DateModule.compare/2 -> DateModule.[before?|after?]

Again, the goal is readability and maintainability. before?/2 and after?/2 were implemented long after compare/2, so it's not unusual that a codebase needs a lot of refactoring to be brought up to date with these new functions. That's where Quokka comes in!

The examples below use DateTime.compare/2, but the same is also done for NaiveDateTime|Time|Date.compare/2

# Before
DateTime.compare(start, end_date) == :gt
# Styled
DateTime.after?(start, end_date)

# Before
DateTime.compare(start, end_date) == :lt
# Styled
DateTime.before?(start, end_date)

Filter count

This addresses Credo.Check.Refactor.FilterCount. This is not configurable.

[1, 2, 3, 4, 5]
|> Enum.filter(fn x -> rem(x, 3) == 0 end)
|> Enum.count()

# Styled:
Enum.count([1, 2, 3, 4, 5], fn x -> rem(x, 3) == 0 end)

Map into

This addresses Credo.Check.Refactor.MapInto. This is not configurable.

[:apple, :banana, :carrot]
|> Enum.map(&({&1, to_string(&1)}))
|> Enum.into(%{})

# Styled:
Map.new([:apple, :banana, :carrot], &{&1, to_string(&1)})

Map join

This addresses Credo.Check.Refactor.MapJoin. This is not configurable.

["a", "b", "c"]
|> Enum.map(&String.upcase/1)
|> Enum.join(", ")

# Styled:
Enum.join(["a", "b", "c"], ", ", &String.upcase/1)

Implicit Try

Quokka will rewrite functions whose entire body is a try/do to instead use the implicit try syntax. Addresses Credo.Check.Readability.PreferImplicitTry. This is not configurable.

The following example illustrates the most complex case, but Quokka happily handles just basic try do/rescue bodies just as easily.

Before

def foo() do
  try do
    uh_oh()
  rescue
    exception -> {:error, exception}
  catch
    :a_throw -> {:error, :threw!}
  else
    try_has_an_else_clause? -> {:did_you_know, try_has_an_else_clause?}
  after
    :done
  end
end

After

def foo() do
  uh_oh()
rescue
  exception -> {:error, exception}
catch
  :a_throw -> {:error, :threw!}
else
  try_has_an_else_clause? -> {:did_you_know, try_has_an_else_clause?}
after
  :done
end

Add parenthesis to 0-arity functions and macro definitions

This addresses Credo.Check.Readability.ParenthesesOnZeroArityDefs. Quokka will add or remove parens from function calls in pipes when the function has no arguments based on the Credo config. If the Credo check is disabled, Quokka will not add or remove parens.

# Behavior if .credo.exs has `Credo.Check.Readability.ParenthesesOnZeroArityDefs, parens: true`
# Before
def foo
defp foo
defmacro foo
defmacrop foo

# Styled
def foo()
defp foo()
defmacro foo()
defmacrop foo()
# Behavior if .credo.exs has `Credo.Check.Readability.ParenthesesOnZeroArityDefs, parens: false`
# Before
def foo()
defp foo()
defmacro foo()
defmacrop foo()

# Styled
def foo
defp foo
defmacro foo
defmacrop foo

Elixir Deprecation Rewrites

BeforeAfterElixir Version
Logger.warnLogger.warning1.15+
Path.safe_relative_to/2Path.safe_relative/21.15+
~R/my_regex/~r/my_regex/1.15+
Enum/String.slice/2 with decreasing rangesadd explicit steps to the range *1.15+
Date.range/2 with decreasing rangeDate.range/3 *1.15+
IO.read/bin_read with :all optionreplace :all with :eof1.15+
File.stream!(path, [encoding: :utf8, trim_bom: true], :line) (:line :bytes deprecated)File.stream!(path, :line, encoding: :utf8, trim_bom: true)1.16+
:timer.units(x)to_time(unit: x)1.17+
first..last = ...first..last//_ = ...1.18+
List.zip(list1, list2)Enum.zip(list1, list2)1.18+
%Foo{x | y} => %{x | y}%{x | y}1.19+

* For both of the "decreasing range" changes, the rewrite can only be applied if the range is being passed as an argument to the function.

Putting variable matching on the right

# Before
case foo do
  bar = %{baz: baz? = true} -> :baz?
  opts = [[a = %{}] | _] -> a
end
# Styled:
case foo do
  %{baz: true = baz?} = bar -> :baz?
  [[%{} = a] | _] = opts -> a
end

# Before
with {:ok, result = %{}} <- foo, do: result
# Styled
with {:ok, %{} = result} <- foo, do: result

# Before
def foo(bar = %{baz: baz? = true}, opts = [[a = %{}] | _]), do: :ok
# Styled
def foo(%{baz: true = baz?} = bar, [[%{} = a] | _] = opts), do: :ok

Drops superfluous = _ in pattern matching

# Before
def foo(_ = bar), do: bar
# Styled
def foo(bar), do: bar

# Before
case foo do
  _ = bar -> :ok
end
# Styled
case foo do
  bar -> :ok
end

Use Implicit Try

# before
def foo d
  try do
    throw_ball()
  catch
    :ball -> :caught
  end
end

# Styled:
def foo d
  throw_ball()
catch
  :ball -> :caught
end

Shrink Function Definitions to One Line When Possible

# Before

def save(
       # Socket comment
       %Socket{assigns: %{user: user, live_action: :new}} = initial_socket,
       # Params comment
       params
     ),
     do: :ok

# Styled

# Socket comment
# Params comment
def save(%Socket{assigns: %{user: user, live_action: :new}} = initial_socket, params), do: :ok

Parameter Pattern Matching Consistency

This addresses Credo.Check.Consistency.ParameterPatternMatching. Note that while this is configurable in credo, Quokka will rewrite all matches to be on the right hand side of the = sign.

# Before
def process(user = %User{age: age}) when age >= 18 do
  # ...
end

# Styled
def process(%User{age: age} = user) when age >= 18 do
  # ...
end

# Before - match on left
def process(opts = [foo: foo]) do
  # ...
end

# Styled - match on right
def process([foo: foo] = opts) do
  # ...
end

Line Length

This addresses Credo.Check.Readability.MaxLineLength. Quokka will respect the minimum of the :line_length configuration between .credo.exs and .formatter.exs when determining whether to split lines. When possible, will compress code onto a single line if it fits within the configured length.

# Before - Multiple lines when it could fit on one
def process_user(
  name,
  email
) do
  UserProcessor.handle(name, email)
end

# Styled - Compressed to one line since it fits
def process_user(name, email) do
  UserProcessor.handle(name, email)
end

# Before - Long line exceeding configured length
def process_user_data(user_name, email, age, occupation, address, phone_number, preferences), do: UserProcessor.handle_data(user_name, email, age, occupation, address, phone_number, preferences)

# Styled - Split across multiple lines to respect length limit
def process_user_data(
  user_name,
  email,
  age,
  occupation,
  address,
  phone_number,
  preferences
), do: UserProcessor.handle_data(
  user_name,
  email,
  age,
  occupation,
  address,
  phone_number,
  preferences
)