Scalariform is now officially maintained by scala-ide. All further developments will be happening in their branch.
Scalariform is a code formatter for Scala. It's available as a library, a stand-alone command line tool, or via integrations with various editors and build tools (listed below).
The plan is to add preferences and features as and when people ask for them, so please do raise a Github issue if it doesn't format your code the way you'd like it, and I'll see what I can do.
Scalariform is licenced under The MIT Licence.
Mac OS X users can install the scalariform CLI tool using the Homebrew package manager.
brew install scalariform
Or, if you would like to install the latest development release:
brew install --HEAD scalariform
If you would like to package scalariform for use on the command line with java -jar, clone the repo and perform the following simple steps:
sbt "project cli" "assembly"
sbt will build one jar with all the dependencies and put it in
cli/target/scala-$your_scala_version/cli-assembly-$scalariform_version.jar
You can copy this to a location in your path and execute it as follows:
java -jar /home/me/bin/cli-assembly-$scalariform_version.jar -f -q +compactControlReadability +alignParameters +alignSingleLineCaseStatements +doubleIndentClassDeclaration +preserveDanglingCloseParenthesis +rewriteArrowSymbols +preserveSpaceBeforeArguments --stdout ~/myproject/src/main/scala/Stuff.scala > Stuff.scala
A version for sbt >= 0.13.x has been written by Peter Vlugter: https://github.com/daniel-trinh/sbt-scalariform
Please see https://github.com/sbt/sbt-scalariform for older versions of sbt.
Have a use for the scalariform source code directly? You can use it as a build dependency:
"org.scalariform" %% "scalariform" % "0.1.8"
Scala IDE for Eclipse uses Scalariform for code formatting:
- Right click in the editor -> Source -> Format
- Press Ctrl-Shift-F
If you select some lines, only those will be formatted.
You can also configure formatting to be run as a save action (Window -> Preferences -> Java -> Editor -> Save Actions).
To set preferences, go to either
- Window -> Preferences -> Scala -> Editor -> Formatter
- Project -> Properties -> Scala Formatter
"ENSIME uses the Scalariform library to format Scala sources. Type C-c C-v f to format the current buffer."
http://aemon.com/file_dump/ensime_manual.html#tth_sEc4.8
See ScalaSidekick by Stefan Ettrup:
Run Plugins -> scalaSidekickPlugin -> Format Scala File
There is a Maven plugin to run Scalariform contributed by Adam Crain on scala-tools.
Usage:
<plugin> <groupId>org.scalariform</groupId> <artifactId>scalariform-maven-plugin</artifactId> <version>0.1.4</version> <executions> <execution> <phase>process-sources</phase> <goals> <goal>format</goal> </goals> <configuration> <rewriteArrowSymbols>true</rewriteArrowSymbols> </configuration> </execution> </executions> </plugin>
See Mads Jensen's Scala TextMate bundle:
http://github.com/mads379/scala.tmbundle
Reformat using Ctrl-Shift-H.
While there is no specific Vim integration at present, you can use
Scalariform as an external formatter for the gq
command by adding
the following to .vimrc
au BufEnter *.scala setl formatprg=java\ -jar\ /home/me/bin/scalariform.jar\ -f\ -q\ +compactControlReadability\ +alignParameters\ +alignSingleLineCaseStatements\ +doubleIndentClassDeclaration\ +preserveDanglingCloseParenthesis\ +rewriteArrowSymbols\ +preserveSpaceBeforeArguments\ --stdin\ --stdout
Or, if you don't like escaping spaces, you can set up a mapping:
map ,st :%!java -jar /home/me/bin/scalariform.jar -f -q +compactControlReadability +alignParameters +alignSingleLineCaseStatements +doubleIndentClassDeclaration +preserveDanglingCloseParenthesis +rewriteArrowSymbols +preserveSpaceBeforeArguments --stdin --stdout <CR>
You can create your own executable scalariform.jar by following the instructions at the top of this file, in "Packaging an executable JAR."
https://github.com/mdr/scalariform/wiki/Command-line-tool
https://github.com/mdr/scalariform/wiki/Library
Default: false
Align class/function parameters (modifiers and name, type, and defaults) in three columns.
For example, if false
, then:
class Person(name: String, age: Int = 24, birthdate: Date, astrologicalSign: String = "libra", shoeSize: Int, favoriteColor: java.awt.Color )
If true
, then:
class Person( name: String, age: Int = 24, birthdate: Date, astrologicalSign: String = "libra", shoeSize: Int, favoriteColor: java.awt.Color )
This will also place the "implicit" keyword in parameters on its own line, whenever the parameter being formatted contains a newline:
For example, if false
, then:
def formatBirthDate( implicit birthdate: Date = Date("11/11/11"), birthtime: Time ): DateTime
If true
, then:
def formatBirthDate( implicit birthdate: Date = Date("11/11/11"), birthtime: Time ): DateTime
This option is disabled if indentWithTabs
is true
.
Default: false
Aligns multi-line arguments
For example, if false
, then:
Cake(candles = 10, frostingFlavor = Vanilla, layerFlavor = Chocolate, iceCream = true )
If true
, then:
Cake( candles = 10, frostingFlavor = Vanilla, layerFlavor = Chocolate, iceCream = true )
This option is disabled if indentWithTabs
is true
.
Default: false
Align the arrows of consecutive single-line case statements. For example, if true
, then:
a match { case b => 1 case ccc => 2 case dd => 3 }
Is reformatted as:
a match { case b => 1 case ccc => 2 case dd => 3 }
This option is disabled if indentWithTabs
is true
.
Default: 40
When alignSingleLineCaseStatements
is true
, this is a limit on
the number of spaces that can be inserted before an arrow to align it
with other case statements. This can be used to avoid very large gaps,
e.g.:
a match { case Some(wibble, wobble) if wibble + wibble > wobble * wibble => 1 case ccc => 2 }
Default: false
When compactControlReadability
is true
, then if
/else
and
try
/catch
/finally
control structures will be formatted
using Compact Control Readability style
if (x == y) { foo() } else if (y == z) { bar() } else { baz() } try { foo() } catch { case _ => bar() } finally { baz() }
Default: false
Omit spaces when formatting a '+' operator on String literals. For example, if false
, then:
"Hello " + name + "!"
If true
, then:
"Hello "+name+"!"
The Scala Style Guide recommends that operators, "should always be invoked using infix notation with spaces separated the target".
Default: false
With this set to true
, class (and trait / object) declarations
will be formatted as recommended by the Scala Style Guide. That is,
if the declaration section spans multiple lines, it will be formatted
so that either the parameter section or the inheritance section is
doubly indented. This provides a visual distinction from the members
of the class. For example:
class Person( name: String, age: Int, birthdate: Date, astrologicalSign: String, shoeSize: Int, favoriteColor: java.awt.Color) extends Entity with Logging with Identifiable with Serializable { def firstMethod = ... }
Or:
class Person( name: String, age: Int, birthdate: Date, astrologicalSign: String, shoeSize: Int, favoriteColor: java.awt.Color) { def firstMethod = ... }
Default: true
Format embedded XML literals; if false
they will be left untouched.
Default: false
If true
, indent local methods an extra level, with the intention of distinguishing them from other statements. For example,:
class A { def find(...) = { val x = ... def find0() = { ... } find0(...) } }
Default: true
Whether to indent package blocks. For example, if true
:
package foo { package bar { class Baz } }
Else if false
:
package foo { package bar { class Baz } }
Default: 2
The number of spaces to use for each level of indentation.
This option is ignored if indentWithTabs
is true
.
Default: false
Use a tab for each level of indentation. When set to true
, this
ignores any setting given for indentSpaces
. In addition, for the
moment, alignSingleLineCaseStatements
, alignArguments
, and alignParameters
options are not supported when indenting with tabs, and XML
indentation is handled differently.
Default: false
If true
, start a multi-line Scaladoc comment body on same line as the opening comment delimiter:
/** This method applies f to each * element of the given list. */
If false
, start the comment body on a separate line below the opening delimiter:
/** * This method applies f to each * element of the given list. */
Default: false
If true
, Scaladoc asterisks will be placed beneath the second asterisk:
/** Wibble * wobble */ class A
Otherwise, if false
, beneath the first asterisk:
/** Wibble * wobble */ class A
Default: false
If true
, the formatter will keep an existing space before a parenthesis argument. For example:
stack.pop() should equal (2)
Otherwise, if false
, spaces before arguments will always be removed.
Default: Force
If Force
, any closing parentheses will be set to dangle. For example:
Box( contents: List[Thing])
becomes:
Box( contents: List[Thing] )
If Prevent
, all dangling parenthesis are collapsed. For example:
Box( contents: List[Thing] )
becomes:
Box( contents: List[Thing])
If Preserve
, scalariform will try to match what unformatted source code is already doing per parenthesis,
either forcing or preventing.
Default: false
Replace arrow tokens with their unicode equivalents: =>
with ⇒
, and <-
with ←
. For example:
for (n <- 1 to 10) n % 2 match { case 0 => println("even") case 1 => println("odd") }
is formatted as:
for (n ← 1 to 10) n % 2 match { case 0 ⇒ println("even") case 1 ⇒ println("odd") }
Default: false
Whether to ensure a space before colon. For example, if false
, then:
def add(a: Int, b: Int): Int = a + b
If true
, then:
def add(a : Int, b : Int) : Int = a + b
Default: false
Whether to use a space inside type brackets. For example, if true
, then:
Array[ String ]
If false
, then:
Array[String]
Default: false
Whether to use a space inside non-empty parentheses. For example, if true
, then:
def main( args : Array[String] )
If false
, then:
def main(args : Array[String])
Default: true
Whether to add a space around the @ token in pattern binders. For example, if true
,:
case elem @ Multi(values @ _*) =>
If false
,:
case elem@Multi(values@_*) =>
Default: false
Whether or not to add spaces around multi-imports.
For example, if false
, then:
import a.{b,c,d} import foo.{bar => baz}
If true
, then:
import a.{ b, c, d } import foo.{ bar => baz }
Older versions of Scalariform used true
,
but the standard Scala formatting requires false
.
See the examples given in "Chapter 13 - Packages and Imports.", page 244 of Programming in Scala 2nd ed. (2010) by Odersky, Spoon and Venners.
Scalariform is compatible with the Scala Style Guide in the sense that, given the right preference settings, source code that is initially compliant with the Style Guide will not become uncompliant after formatting. In a number of cases, running the formatter will make uncompliant source more compliant.
Preference | Value | Default? |
---|---|---|
alignParameters | false |
|
compactStringConcatenation | false |
|
doubleIndentClassDeclaration | true |
No |
indentSpaces | 2 |
|
placeScaladocAsterisksBeneathSecondAsterisk | true |
No |
preserveSpaceBeforeArguments | false |
|
rewriteArrowSymbols | false |
|
spaceBeforeColon | false |
|
spaceInsideBrackets | false |
|
spaceInsideParentheses | false |
|
spacesAroundMultiImports | false |
As well as global preferences, formatting can be tweaked at the source level through comments.
Disables the formatter for selective portions of a source file:
// format: OFF <-- this directive disables formatting from this point class AsciiDSL { n ¦- "1" -+ { n: Node => n ¦- "i" n ¦- "ii" n ¦- "iii" n ¦- "iv" n ¦- "v" } n ¦- "2" n ¦- "3" -+ { n: Node => n ¦- "i" n ¦- "ii" -+ { n: Node => n ¦- "a" n ¦- "b" n ¦- "c" } n ¦- "iii" n ¦- "iv" n ¦- "v" } // format: ON <-- formatter resumes from this point ... } // (see: http://dev.day.com/microsling/content/blogs/main/scalajcr2.html)
Sets a preference for the entirety of the source file, overriding the global formatting settings:
// format: +preserveSpaceBeforeArguments class StackSpec extends FlatSpec with ShouldMatchers { // ... stack.pop() should equal (2) }