Keep your code Spotless with Gradle
Spotless is a general-purpose formatting plugin used by 15,000 projects on GitHub (Jan 2023). It is completely à la carte, but also includes powerful "batteries-included" if you opt-in.
To people who use your build, it looks like this (IDE support also available):
user@machine repo % ./gradlew build
:spotlessJavaCheck FAILED
The following files had format violations:
src\main\java\com\diffplug\gradle\spotless\FormatExtension.java
-\t\t····if·(targets.length·==·0)·{
+\t\tif·(targets.length·==·0)·{
Run './gradlew spotlessApply' to fix these violations.
user@machine repo % ./gradlew spotlessApply
:spotlessApply
BUILD SUCCESSFUL
user@machine repo % ./gradlew build
BUILD SUCCESSFUL
Spotless supports all of Gradle's built-in performance features (incremental build, remote and local buildcache, lazy configuration, etc), and also automatically fixes idempotence issues, infers line-endings from git, is cautious about misconfigured encoding bugs, and can use git to ratchet formatting without "format-everything" commits.
- Quickstart
- Languages
- Java (google-java-format, eclipse jdt, clang-format, prettier, palantir-java-format, formatAnnotations, cleanthat)
- Groovy (eclipse groovy)
- Kotlin (ktfmt, ktlint, diktat, prettier)
- Scala (scalafmt)
- C/C++ (clang-format, eclipse cdt)
- Protobuf (buf, clang-format)
- Python (black)
- FreshMark aka markdown
- Flexmark aka markdown
- Antlr4 (antlr4formatter)
- SQL (dbeaver, prettier)
- Maven POM (sortPom)
- Typescript (tsfmt, prettier, ESLint, Biome)
- Javascript (prettier, ESLint, Biome)
- JSON (simple, gson, jackson, Biome, jsonPatch)
- YAML
- Shell
- Gherkin
- Multiple languages
- Prettier (plugins, npm detection,
.npmrc
detection, cachingnpm install
results)- javascript, jsx, angular, vue, flow, typescript, css, less, scss, html, json, graphql, markdown, ymaml
- clang-format
- c, c++, c#, objective-c, protobuf, javascript, java
- eclipse web tools platform
- css, html, js, json, xml
- Biome (binary detection, config file, input language)
- Prettier (plugins, npm detection,
- Language independent
- Generic steps
- License header (slurp year from git)
- How can I enforce formatting gradually? (aka "ratchet")
spotless:off
andspotless:on
- Line endings and encodings (invisible stuff)
- Custom steps
- Multiple (or custom) language-specific blocks
- Inception (languages within languages within...)
- Disabling warnings and error messages
- Dependency resolution modes
- How do I preview what
spotlessApply
will do? - Example configurations (from real-world projects)
Contributions are welcome, see the contributing guide for development info.
To use it in your buildscript, just add the Spotless dependency, and configure it like so:
spotless {
// optional: limit format enforcement to just the files changed by this feature branch
ratchetFrom 'origin/main'
format 'misc', {
// define the files to apply `misc` to
target '*.gradle', '.gitattributes', '.gitignore'
// define the steps to apply to those files
trimTrailingWhitespace()
indentWithTabs() // or spaces. Takes an integer argument if you don't like 4
endWithNewline()
}
java {
// don't need to set target, it is inferred from java
// apply a specific flavor of google-java-format
googleJavaFormat('1.8').aosp().reflowLongStrings().skipJavadocFormatting()
// fix formatting of type annotations
formatAnnotations()
// make sure every file has the following copyright header.
// optionally, Spotless can set copyright years by digging
// through git history (see "license" section below)
licenseHeader '/* (C)$YEAR */'
}
}
Spotless consists of a list of formats (in the example above, misc
and java
), and each format has:
- a
target
(the files to format), which you set withtarget
andtargetExclude
- a list of
FormatterStep
, which are justString -> String
functions, such asreplace
,replaceRegex
,trimTrailingWhitespace
,custom
,prettier
,eclipseWtp
,licenseHeader
etc.
All the generic steps live in FormatExtension
, and there are many language-specific steps which live in its language-specific subclasses, which are described below.
Spotless requires JRE 11+ and Gradle 6.1.1 or newer.
- If you're stuck on JRE 8, use
id 'com.diffplug.spotless' version '6.13.0'
or older. - If you're stuck on an older version of Gradle,
id 'com.diffplug.gradle.spotless' version '4.5.1'
supports all the way back to Gradle 2.x.
com.diffplug.gradle.spotless.JavaExtension
javadoc, code
spotless {
java {
// Use the default importOrder configuration
importOrder()
// optional: you can specify import groups directly
// note: you can use an empty string for all the imports you didn't specify explicitly, '|' to join group without blank line, and '\\#` prefix for static imports
importOrder('java|javax', 'com.acme', '', '\\#com.acme', '\\#')
// optional: instead of specifying import groups directly you can specify a config file
// export config file: https://github.com/diffplug/spotless/blob/main/ECLIPSE_SCREENSHOTS.md#creating-spotlessimportorder
importOrderFile('eclipse-import-order.txt') // import order file as exported from eclipse
removeUnusedImports()
// Cleanthat will refactor your code, but it may break your style: apply it before your formatter
cleanthat() // has its own section below
// Choose one of these formatters.
googleJavaFormat() // has its own section below
eclipse() // has its own section below
prettier() // has its own section below
clangFormat() // has its own section below
formatAnnotations() // fixes formatting of type annotations, see below
licenseHeader '/* (C) $YEAR */' // or licenseHeaderFile
}
}
The target is usually inferred automatically from the java source sets. However, Spotless cannot automatically detect android or java-gradle-plugin sources, but you can fix this easily:
spotless {
java {
target 'src/*/java/**/*.java'
spotless {
java {
removeUnusedImports()
// optional: you may switch for `google-java-format` as underlying engine to `cleanthat-javaparser-unnecessaryimport`
// which enables processing any language level source file with a JDK8+ Runtime
removeUnusedImports('cleanthat-javaparser-unnecessaryimport')
spotless {
java {
googleJavaFormat()
// optional: you can specify a specific version (>= 1.8) and/or switch to AOSP style
// and/or reflow long strings
// and/or use custom group artifact (you probably don't need this)
googleJavaFormat('1.8').aosp().reflowLongStrings().formatJavadoc(false).reorderImports(false).groupArtifact('com.google.googlejavaformat:google-java-format')
spotless {
java {
palantirJavaFormat()
// optional: you can specify a specific version and/or switch to AOSP/GOOGLE style
palantirJavaFormat('2.9.0').style("GOOGLE")
// optional: you can also format Javadocs, requires at least Palantir 2.39.0
palantirJavaFormat('2.39.0').formatJavadoc(true)
homepage. See here for screenshots that demonstrate how to get and install the config file mentioned below.
spotless {
java {
eclipse()
// optional: you can specify a specific version and/or config file
eclipse('4.26').configFile('eclipse-prefs.xml')
// if the access to the p2 repositories is restricted, mirrors can be
// specified using a URI prefix map as follows:
eclipse().withP2Mirrors(['https://download.eclipse.org/eclipse/updates/4.29/':'https://some.internal.mirror/4-29-updates-p2/'])
Type annotations should be on the same line as the type that they qualify.
@Override
@Deprecated
@Nullable @Interned String s;
However, some tools format them incorrectly, like this:
@Override
@Deprecated
@Nullable
@Interned
String s;
To fix the incorrect formatting, add the formatAnnotations()
rule after a Java formatter. For example:
spotless {
java {
googleJavaFormat()
formatAnnotations()
}
}
This does not re-order annotations, it just removes incorrect newlines.
A type annotation is an annotation that is meta-annotated with @Target({ElementType.TYPE_USE})
.
Spotless has a default list of well-known type annotations.
You can use addTypeAnnotation()
and removeTypeAnnotation()
to override its defaults:
formatAnnotations().addTypeAnnotation("Empty").addTypeAnnotation("NonEmpty").removeTypeAnnotation("Localized")
You can make a pull request to add new annotations to Spotless's default list.
homepage. CleanThat enables automatic refactoring of Java code. ChangeLog
spotless {
java {
cleanthat()
// optional: you can specify a specific version and/or config file
cleanthat()
.groupArtifact('io.github.solven-eu.cleanthat:java') // Optional. Default is 'io.github.solven-eu.cleanthat:java'
.version('2.8') // You may force a custom version of Cleanthat
.sourceCompatibility('1.7') // default is '1.7'
.addMutator('SafeAndConsensual') // Default includes the SafeAndConsensual composite mutator
.addMutator('your.custom.MagicMutator') // List of mutators: https://github.com/solven-eu/cleanthat/blob/master/MUTATORS.generated.MD
.excludeMutator('UseCollectionIsEmpty') // You may exclude some mutators (from Composite ones)
.includeDraft(false) // You may exclude draft mutators (from Composite ones)
com.diffplug.gradle.spotless.GroovyExtension
javadoc, codecom.diffplug.gradle.spotless.GroovyGradleExtension
javadoc, code
Configuration for Groovy is similar to Java, in that it also supports licenseHeader
and importOrder
.
The groovy formatter's default behavior is to format all .groovy
and .java
files found in the Java and Groovy source sets. If you would like to exclude the .java
files, set the parameter excludeJava
, or you can set the target
parameter as described in the Custom rules section.
apply plugin: 'groovy'
spotless {
groovy {
// Use the default importOrder configuration
importOrder()
// optional: you can specify import groups directly
// note: you can use an empty string for all the imports you didn't specify explicitly, and '\\#` prefix for static imports
importOrder('java', 'javax', 'com.acme', '', '\\#com.acme', '\\#')
// optional: instead of specifying import groups directly you can specify a config file
// export config file: https://github.com/diffplug/spotless/blob/main/ECLIPSE_SCREENSHOTS.md#creating-spotlessimportorder
importOrderFile('eclipse-import-order.txt') // import order file as exported from eclipse
// removes semicolons at the end of lines
removeSemicolons()
// the Groovy Eclipse formatter extends the Java Eclipse formatter,
// so it formats Java files by default (unless `excludeJava` is used).
greclipse() // has its own section below
licenseHeader('/* (C) $YEAR */') // or licenseHeaderFile
//---- Below is for `groovy` only ----
// excludes all Java sources within the Groovy source dirs from formatting
excludeJava()
}
groovyGradle {
target '*.gradle' // default target of groovyGradle
greclipse()
}
}
homepage. changelog. compatible versions. The Groovy formatter uses some of the eclipse jdt configuration parameters in addition to groovy-specific ones. All parameters can be configured within a single file, like the Java properties file greclipse.properties in the previous example. The formatter step can also load the exported Eclipse properties and augment it with the .metadata/.plugins/org.eclipse.core.runtime/.settings/org.codehaus.groovy.eclipse.ui.prefs
from your Eclipse workspace as shown below.
spotless {
groovy {
// Use the default version and Groovy-Eclipse default configuration
greclipse()
// optional: you can specify a specific version or config file(s), version matches the Eclipse Platform
greclipse('4.26').configFile('spotless.eclipseformat.xml', 'org.codehaus.groovy.eclipse.ui.prefs')
Groovy-Eclipse formatting errors/warnings lead per default to a build failure. This behavior can be changed by adding the property/key value ignoreFormatterProblems=true
to a configuration file. In this scenario, files causing problems, will not be modified by this formatter step.
com.diffplug.gradle.spotless.KotlinExtension
javadoc, codecom.diffplug.gradle.spotless.KotlinGradleExtension
javadoc, code
spotless { // if you are using build.gradle.kts, instead of 'spotless {' use:
// configure<com.diffplug.gradle.spotless.SpotlessExtension> {
kotlin {
// by default the target is every '.kt' and '.kts` file in the java sourcesets
ktfmt() // has its own section below
ktlint() // has its own section below
diktat() // has its own section below
prettier() // has its own section below
licenseHeader '/* (C)$YEAR */' // or licenseHeaderFile
}
kotlinGradle {
target('*.gradle.kts') // default target for kotlinGradle
ktlint() // or ktfmt() or prettier()
}
}
spotless {
kotlin {
// version, style and all configurations here are optional
ktfmt("0.51").googleStyle().configure {
it.setMaxWidth(80)
it.setBlockIndent(4)
it.setContinuationIndent(4)
it.setRemoveUnusedImports(false)
it.setManageTrailingCommas(false)
}
}
}
Spotless respects the .editorconfig
settings by providing editorConfigPath
option.
(ktlint docs).
Default value is the .editorconfig
file located in the top project.
Passing null
will clear the option.
Additionally, editorConfigOverride
options will override what's supplied in .editorconfig
file.
spotless {
kotlin {
// version, editorConfigPath, editorConfigOverride and customRuleSets are all optional
ktlint("1.0.0")
.setEditorConfigPath("$projectDir/config/.editorconfig") // sample unusual placement
.editorConfigOverride(
mapOf(
"indent_size" to 2,
// intellij_idea is the default style we preset in Spotless, you can override it referring to https://pinterest.github.io/ktlint/latest/rules/code-styles.
"ktlint_code_style" to "intellij_idea",
)
)
.customRuleSets(
listOf(
"io.nlopez.compose.rules:ktlint:0.3.3"
)
)
}
}
homepage. changelog. You can provide configuration path manually as configFile
.
spotless {
kotlin {
// version and configFile are both optional
diktat('1.0.1').configFile("full/path/to/diktat-analysis.yml")
com.diffplug.gradle.spotless.ScalaExtension
javadoc, code
spotless {
scala {
// by default, all `.scala` and `.sc` files in the java sourcesets will be formatted
scalafmt() // has its own section below
licenseHeader '/* (C) $YEAR */', 'package ' // or licenseHeaderFile
// note the 'package ' argument - this is a regex which identifies the top
// of the file, be careful that all of your sources have a package declaration,
// or pick a regex which works better for your code
}
}
homepage. changelog. config docs.
spotless {
scala {
// version and configFile, scalaMajorVersion are all optional
scalafmt('3.5.9').configFile('scalafmt.conf').scalaMajorVersion('2.13')
com.diffplug.gradle.spotless.CppExtension
javadoc, code
spotless {
cpp {
target 'src/native/**' // you have to set the target manually
clangFormat() // has its own section below
eclipseCdt() // has its own section below
licenseHeader '/* (C) $YEAR */' // or licenseHeaderFile
}
}
homepage. compatible versions.
spotles {
cpp {
// version and configFile are both optional
eclipseCdt('4.13.0').configFile('eclipse-cdt.xml')
}
}
com.diffplug.gradle.spotless.PythonExtension
javadoc, code
spotless {
python {
target 'src/main/**/*.py' // have to set manually
black() // has its own section below
licenseHeader '/* (C) $YEAR */', 'REGEX_TO_DEFINE_TOP_OF_FILE' // or licenseHeaderFile
}
}
black('19.10b0') // version is optional
// if black is not on your path, you must specify its location manually
black().pathToExe('C:/myuser/.pyenv/versions/3.8.0/scripts/black.exe')
// Spotless always checks the version of the black it is using
// and will fail with an error if it does not match the expected version
// (whether manually specified or default). If there is a problem, Spotless
// will suggest commands to help install the correct version.
// TODO: handle installation & packaging automatically - https://github.com/diffplug/spotless/issues/674
com.diffplug.gradle.spotless.ProtobufExtension
javadoc, code
WARNING this step must be the first step in the chain, steps before it will be ignored. Thumbs up this issue for a resolution, see here for more details on the problem.
spotless {
protobuf {
// by default the target is every '.proto' file in the project
buf()
licenseHeader '/* (C) $YEAR */' // or licenseHeaderFile
}
}
When used in conjunction with the buf-gradle-plugin, the buf
executable can be resolved from its bufTool
configuration:
spotless {
protobuf {
buf().pathToExe(configurations.getByName(BUF_BINARY_CONFIGURATION_NAME).getSingleFile().getAbsolutePath())
}
}
// Be sure to disable the buf-gradle-plugin's execution of `buf format`:
buf {
enforceFormat = false
}
com.diffplug.gradle.spotless.FreshMarkExtension
javadoc, code
homepage. changelog. FreshMark lets you generate markdown in the comments of your markdown. This helps to keep badges and links up-to-date (see the source for this file), and can also be helpful for generating complex tables (see the source for the parent readme).
To apply freshmark to all of the .md
files in your project, with all of your project's properties available for templating, use this snippet:
spotless {
freshmark {
target '*.md' // you have to set the target manually
propertiesFile('gradle.properties') // loads all the properties in the given file
properties {
it.put('key', 'value') // specify other properties manually
}
}
}
com.diffplug.gradle.spotless.FlexmarkExtension
javadoc, code
homepage. Flexmark is a flexible Commonmark/Markdown parser that can be used to format Markdown files. It supports different flavors of Markdown and many formatting options.
Currently, none of the available options can be configured yet. It uses only the default options together with COMMONMARK
as FORMATTER_EMULATION_PROFILE
.
To apply flexmark to all of the .md
files in your project, use this snippet:
spotless {
flexmark {
target '**/*.md' // you have to set the target manually
flexmark() // or flexmark('0.64.8') // version is optional
}
}
com.diffplug.gradle.spotless.Antlr4Extension
javadoc, code
spotless {
antlr4 {
target 'src/*/antlr4/**/*.g4' // default value, you can change if you want
antlr4Formatter() // has its own section below
licenseHeader '/* (C) $YEAR */' // or licenseHeaderFile
}
}
antlr4formatter('1.2.1') // version is optional
com.diffplug.gradle.spotless.SqlExtension
javadoc, code
spotless {
sql {
target 'src/main/resources/**/*.sql' // have to set manually
dbeaver() // has its own section below
prettier() // has its own section below
}
}
homepage. DBeaver is only distributed as a monolithic jar, so the formatter used here was copy-pasted into Spotless, and thus there is no version to change.
spotless {
sql {
dbeaver().configFile('dbeaver.properties') // configFile is optional
Default configuration file, other options available here.
# case of the keywords (UPPER, LOWER or ORIGINAL)
sql.formatter.keyword.case=UPPER
# Statement delimiter
sql.formatter.statement.delimiter=;
# Indentation style (space or tab)
sql.formatter.indent.type=space
# Number of indentation characters
sql.formatter.indent.size=4
com.diffplug.gradle.spotless.PomExtension
javadoc, code
spotless {
pom {
target('pom.xml') // default value, you can change if you want
sortPom() // has its own section below
}
}
All configuration settings are optional, they are described in detail here.
spotless {
pom {
sortPom('4.0.0')
.encoding('UTF-8') // The encoding of the pom files
.lineSeparator(System.getProperty('line.separator')) // line separator to use
.expandEmptyElements(true) // Should empty elements be expanded
.spaceBeforeCloseEmptyElement(false) // Should a space be added inside self-closing elements
.keepBlankLines(true) // Keep empty lines
.endWithNewline(true) // Whether sorted pom ends with a newline
.nrOfIndentSpace(2) // Indentation
.indentBlankLines(false) // Should empty lines be indented
.indentSchemaLocation(false) // Should schema locations be indented
.indentAttribute(null) // Should the xml attributes be indented
.predefinedSortOrder('recommended_2008_06') // Sort order of elements: https://github.com/Ekryd/sortpom/wiki/PredefinedSortOrderProfiles
.sortOrderFile(null) // Custom sort order of elements: https://raw.githubusercontent.com/Ekryd/sortpom/master/sorter/src/main/resources/custom_1.xml
.sortDependencies(null) // Sort dependencies: https://github.com/Ekryd/sortpom/wiki/SortDependencies
.sortDependencyManagement(null) // Sort dependency management: https://github.com/Ekryd/sortpom/wiki/SortDependencies
.sortDependencyExclusions(null) // Sort dependency exclusions: https://github.com/Ekryd/sortpom/wiki/SortDependencies
.sortPlugins(null) // Sort plugins: https://github.com/Ekryd/sortpom/wiki/SortPlugins
.sortProperties(false) // Sort properties
.sortModules(false) // Sort modules
.sortExecutions(false) // Sort plugin executions
}
}
spotless {
typescript {
target 'src/**/*.ts' // you have to set the target manually
tsfmt() // has its own section below
prettier() // has its own section below
eslint() // has its own section below
biome() // has its own section below
licenseHeader '/* (C) $YEAR */', '(import|const|declare|export|var) ' // or licenseHeaderFile
// note the '(import|const|...' argument - this is a regex which identifies the top
// of the file, be careful that all of your sources have a suitable top-level declaration,
// or pick a regex which works better for your code
}
}
npm. changelog. Please note: The auto-discovery of config files (up the file tree) will not work when using tsfmt within spotless, hence you are required to provide resolvable file paths for config files, or alternatively provide the configuration inline. See tsfmt's default config settings for what is available.
spotless {
typescript {
tsfmt('7.2.2')
// provide config inline: https://github.com/vvakame/typescript-formatter/blob/7764258ad42ac65071399840d1b8701868510ca7/lib/utils.ts#L11L32
.config(['indentSize': 1, 'convertTabsToSpaces': true])
// or according to tsfmt-parameters: https://github.com/vvakame/typescript-formatter/blob/7764258ad42ac65071399840d1b8701868510ca7/lib/index.ts#L27L34
.tsconfigFile('tsconfig.json')
.tslintFile('tslint.json')
.vscodeFile('vscode.json')
.tsfmtFile('tsfmt.json')
Prerequisite: tsfmt requires a working NodeJS version
For details, see the npm detection, .npmrc
detection and caching results of npm install
sections of prettier, which apply also to tsfmt.
npm. changelog. Please note: The auto-discovery of config files (up the file tree) will not work when using ESLint within spotless, hence you are required to provide resolvable file paths for config files, or alternatively provide the configuration inline.
The configuration is very similar to the ESLint (Javascript) configuration. In typescript, a
reference to a tsconfig.json
is required.
spotless {
typescript {
eslint('8.30.0') // version is optional
eslint(['my-eslint-fork': '1.2.3', 'my-eslint-plugin': '1.2.1']) // can specify exactly which npm packages to use
eslint()
// configuration is mandatory. Provide inline config or a config file.
// a) inline-configuration
.configJs('''
{
env: {
browser: true,
es2021: true
},
extends: 'standard-with-typescript',
overrides: [
],
parserOptions: {
ecmaVersion: 'latest',
sourceType: 'module',
project: './tsconfig.json',
},
rules: {
}
}
''')
// b) config file
.configFile('.eslintrc.js')
// recommended: provide a tsconfig.json - especially when using the styleguides
.tsconfigFile('tsconfig.json')
}
}
Prerequisite: ESLint requires a working NodeJS version
For details, see the npm detection, .npmrc
detection and caching results of npm install
sections of prettier, which apply also to ESLint.
spotless {
javascript {
target 'src/**/*.js' // you have to set the target manually
prettier() // has its own section below
eslint() // has its own section below
biome() // has its own section below
licenseHeader '/* (C) $YEAR */', 'REGEX_TO_DEFINE_TOP_OF_FILE' // or licenseHeaderFile
}
}
npm. changelog. Please note: The auto-discovery of config files (up the file tree) will not work when using ESLint within spotless, hence you are required to provide resolvable file paths for config files, or alternatively provide the configuration inline.
The configuration is very similar to the ESLint (Typescript) configuration. In javascript, no
tsconfig.json
is supported.
spotless {
javascript {
eslint('8.30.0') // version is optional
eslint(['my-eslint-fork': '1.2.3', 'my-eslint-plugin': '1.2.1']) // can specify exactly which npm packages to use
eslint()
// configuration is mandatory. Provide inline config or a config file.
// a) inline-configuration
.configJs('''
{
env: {
browser: true,
es2021: true
},
extends: 'standard',
overrides: [
],
parserOptions: {
ecmaVersion: 'latest',
sourceType: 'module'
},
rules: {
}
}
''')
// b) config file
.configFile('.eslintrc.js')
}
}
Prerequisite: ESLint requires a working NodeJS version
For details, see the npm detection, .npmrc
detection and caching results of npm install
sections of prettier, which apply also to ESLint.
spotless {
json {
target 'src/**/*.json' // you have to set the target manually
simple() // has its own section below
prettier().config(['parser': 'json']) // see Prettier section below
eclipseWtp('json') // see Eclipse web tools platform section
gson() // has its own section below
jackson() // has its own section below
biome() // has its own section below
jsonPatch([]) // has its own section below
}
}
Uses a JSON pretty-printer that optionally allows configuring the number of spaces that are used to pretty print objects:
spotless {
json {
target 'src/**/*.json'
simple()
// optional: specify the number of spaces to use
simple().indentWithSpaces(6)
}
}
Uses Google Gson to also allow sorting by keys besides custom indentation - useful for i18n files.
spotless {
json {
target 'src/**/*.json'
gson()
.indentWithSpaces(6) // optional: specify the number of spaces to use
.sortByKeys() // optional: sort JSON by its keys
.escapeHtml() // optional: escape HTML in values
.version('2.8.1') // optional: specify version
}
}
Notes:
- There's no option in Gson to leave HTML as-is (i.e. escaped HTML would remain escaped, raw would remain raw). Either
all HTML characters are written escaped or none. Set
escapeHtml
if you prefer the former. sortByKeys
will apply lexicographic order on the keys of the input JSON. See the javadoc of String for details.
Uses Jackson for json files.
spotless {
json {
target 'src/**/*.json'
jackson()
.spaceBeforeSeparator(false) // optional: add a whitespace before key separator. False by default
.feature('INDENT_OUTPUT', true) // optional: true by default
.feature('ORDER_MAP_ENTRIES_BY_KEYS', true) // optional: false by default
.feature('ANY_OTHER_FEATURE', true|false) // optional: any SerializationFeature can be toggled on or off
.jsonFeature('ANY_OTHER_FEATURE', true|false) // any JsonGenerator.Feature can be toggled on or off
}
}
Uses zjsonpatch to apply JSON Patches as per RFC 6902 to JSON documents.
This enables you to add, replace or remove properties at locations in the JSON document that you specify using JSON Pointers.
In Spotless Gradle, these JSON patches are represented as a List<Map<String, Object>>
, or a list of patch operations.
Each patch operation must be a map with the following properties:
"op"
- the operation to apply, one of"replace"
,"add"
or"remove"
."path"
- a JSON Pointer string, for example"/foo"
"value"
- the value to"add"
or"replace"
at the specified path. Not needed for"remove"
operations.
For example, to apply the patch from the JSON Patch homepage:
spotless {
json {
target 'src/**/*.json'
jsonPatch([
[op: 'replace', path: '/baz', value: 'boo'],
[op: 'add', path: '/hello', value: ['world']],
[op: 'remove', path: '/foo']
])
}
}
Or using the Kotlin DSL:
spotless {
json {
target("src/**/*.json")
jsonPatch(listOf(
mapOf("op" to "replace", "path" to "/baz", "value" to "boo"),
mapOf("op" to "add", "path" to "/hello", "value" to listOf("world")),
mapOf("op" to "remove", "path" to "/foo")
))
}
}
spotless {
yaml {
target 'src/**/*.yaml' // you have to set the target manually
jackson() // has its own section below
prettier() // has its own section below
}
}
Uses Jackson for yaml
files.
spotless {
yaml {
target 'src/**/*.yaml'
jackson()
.spaceBeforeSeparator(false) // optional: add a whitespace before key separator. False by default
.feature('INDENT_OUTPUT', true) // optional: true by default
.feature('ORDER_MAP_ENTRIES_BY_KEYS', true) // optional: false by default
.feature('ANY_OTHER_FEATURE', true|false) // optional: any SerializationFeature can be toggled on or off
.yamlFeature('ANY_OTHER_FEATURE', true|false) // any YAMLGenerator.Feature can be toggled on or off
}
}
com.diffplug.gradle.spotless.ShellExtension
javadoc, code
spotless {
shell {
target 'scripts/**/*.sh' // default: '**/*.sh'
shfmt() // has its own section below
}
}
When formatting shell scripts via shfmt
, configure shfmt
settings via .editorconfig
.
Refer to the shfmt
man page for .editorconfig
settings.
shfmt('3.8.0') // version is optional
// if shfmt is not on your path, you must specify its location manually
shfmt().pathToExe('/opt/homebrew/bin/shfmt')
// Spotless always checks the version of the shfmt it is using
// and will fail with an error if it does not match the expected version
// (whether manually specified or default). If there is a problem, Spotless
// will suggest commands to help install the correct version.
// TODO: handle installation & packaging automatically - https://github.com/diffplug/spotless/issues/674
spotless {
gherkin {
target 'src/**/*.feature' // you have to set the target manually
gherkinUtils() // has its own section below
}
}
Uses a Gherkin pretty-printer that optionally allows configuring the number of spaces that are used to pretty print objects:
spotless {
gherkin {
target 'src/**/*.feature' // required to be set explicitly
gherkinUtils()
.version('9.0.0') // optional: custom version of 'io.cucumber:gherkin-utils'
}
}
com.diffplug.gradle.spotless.CssExtension
javadoc, code
spotless {
css {
target 'css/**/*.css' // default: '**/*.css'
biome('1.8.3') // has its own section below
}
}
Note regarding biome: Biome supports formatting CSS as of 1.8.0 (experimental, opt-in) and 1.9.0 (stable).
homepage. changelog. official plugins. community plugins. Prettier is a formatter that can format almost every anything - JavaScript, JSX, Angular, Vue, Flow, TypeScript, CSS, Less, SCSS, HTML, JSON, GraphQL, Markdown (including GFM and MDX), and YAML. It can format even more using plugins (PHP, Ruby, Swift, XML, Apex, Elm, Java (!!), Kotlin, pgSQL, .properties, solidity, svelte, toml, shellscript, ...).
You can use prettier in any language-specific format, but usually you will be creating a generic format.
spotless {
format 'styling', {
// you have to set the target manually
target 'src/*/webapp/**/*.css', 'src/*/webapp/**/*.scss'
prettier('2.0.4') // version is optional
prettier(['my-prettier-fork': '1.16.4']) // can specify exactly which npm packages to use
// by default, prettier uses the file's extension to guess a parser
// but you can override that and specify the parser manually
prettier().config(['parser': 'css'])
// you can also set some style options
// https://prettier.io/docs/en/configuration.html
prettier().config(['tabWidth': 4])
// you can also slurp from a file or even provide both (inline always takes precedence over file)
prettier().config(['tabWidth': 4]).configFile('path-to/.prettierrc.yml')
}
}
Limitations:
- The auto-discovery of config files (up the file tree) will not work when using prettier within spotless.
- Prettier's override syntax is not supported when using prettier within spotless.
To apply prettier to more kinds of files, just add more formats
Since spotless uses the actual npm prettier package behind the scenes, it is possible to use prettier with plugins or community-plugins in order to support even more file types.
spotless {
java {
prettier(['prettier': '2.8.8', 'prettier-plugin-java': '2.2.0']).config(['parser': 'java', 'tabWidth': 4])
}
format 'php', {
target 'src/**/*.php'
prettier(['prettier': '2.8.8', '@prettier/plugin-php': '0.19.6']).config(['parser': 'php', 'tabWidth': 3])
}
}
With version 3 prettier it is required to pass in an additional 'plugins' parameter to the config block with a list of plugins you want to use.
spotless {
java {
prettier(['prettier': '3.0.3', 'prettier-plugin-java': '2.3.0'])
.config(['parser': 'java', 'tabWidth': 4, 'plugins': ['prettier-plugin-java']])
}
format 'php', {
target 'src/**/*.php'
prettier(['prettier': '3.0.3', '@prettier/plugin-php': '0.20.1'])
.config(['parser': 'php', 'tabWidth': 3, 'plugins': ['@prettier/plugin-php']])
}
}
Prettier is based on NodeJS, so a working NodeJS installation (especially npm) is required on the host running spotless. Spotless will try to auto-discover an npm installation. If that is not working for you, it is possible to directly configure the npm and/or node binary to use.
spotless {
format 'javascript', {
prettier().npmExecutable('/usr/bin/npm').nodeExecutable('/usr/bin/node').config(...)
If you provide both npmExecutable
and nodeExecutable
, spotless will use these paths. If you specify only one of the
two, spotless will assume the other one is in the same directory.
If you use the gradle-node-plugin
(github), it is possible to use the
node- and npm-binaries dynamically installed by this plugin. See
this
or this example.
Spotless picks up npm configuration stored in a .npmrc
file either in the project directory or in your user home.
Alternatively you can supply spotless with a location of the .npmrc
file to use. (This can be combined with
npmExecutable
and nodeExecutable
, of course.)
spotless {
typescript {
prettier().npmrc("$projectDir/config/.npmrc").config(...)
Spotless uses npm
behind the scenes to install prettier
. This can be a slow process, especially if you are using a slow internet connection or
if you need large plugins. You can instruct spotless to cache the results of the npm install
calls, so that for the next installation,
it will not need to download the packages again, but instead reuse the cached version.
spotless {
typescript {
prettier().npmInstallCache() // will use the default cache directory (the build-directory of the respective module)
prettier().npmInstallCache("${rootProject.rootDir}/.gradle/spotless-npm-cache") // will use the specified directory (creating it if not existing)
Depending on your filesystem and the location of the cache directory, spotless will use hardlinks when caching the npm packages. If that is not possible, it will fall back to copying the files.
homepage. changelog. clang-format
is a formatter for c, c++, c#, objective-c, protobuf, javascript, and java. You can use clang-format in any language-specific format, but usually you will be creating a generic format.
spotless {
format 'csharp', {
// you have to set the target manually
target 'src/**/*.cs'
clangFormat('10.0.1') // version is optional
// can also specify a code style
clangFormat().style('LLVM') // or Google, Chromium, Mozilla, WebKit
// TODO: support arbitrary .clang-format
// if clang-format is not on your path, you must specify its location manually
clangFormat().pathToExe('/usr/local/Cellar/clang-format/10.0.1/bin/clang-format')
// Spotless always checks the version of the clang-format it is using
// and will fail with an error if it does not match the expected version
// (whether manually specified or default). If there is a problem, Spotless
// will suggest commands to help install the correct version.
// TODO: handle installation & packaging automatically - https://github.com/diffplug/spotless/issues/673
}
}
changelog. compatible versions.
spotless {
format 'xml', {
target 'src/**/*.xml' // must specify target
eclipseWtp('xml') // must specify a type (table below)
eclipseWtp('xml', '11.0') // optional version, others at https://download.eclipse.org/tools/cdt/releases/
// you can also specify an arbitrary number of config files
eclipseWtp('xml').configFile('spotless.xml.prefs', 'spotless.common.properties'
}
}
The WTP formatter accept multiple configuration files. All Eclipse configuration file formats are accepted as well as simple Java property files. Omit the configFile
entirely to use the default Eclipse configuration. The following formatters and configurations are supported:
Type | Configuration | File location |
---|---|---|
CSS | editor preferences | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.css.core.prefs |
cleanup preferences | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.css.core.prefs | |
HTML | editor preferences | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.html.core.prefs |
cleanup preferences | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.html.core.prefs | |
embedded CSS | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.css.core.prefs | |
embedded JS | Use the export in the Eclipse editor configuration dialog | |
JS | editor preferences | Use the export in the Eclipse editor configuration dialog |
JSON | editor preferences | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.json.core.prefs |
XML | editor preferences | .metadata/.plugins/org.eclipse.core.runtime/.settings/org.eclipse.wst.xml.core.prefs |
Note that HTML
should be used for X-HTML
sources instead of XML
.
The Eclipse XML catalog cannot be configured for the Spotless WTP formatter, instead a
user defined catalog file can be specified using the property userCatalog
. Catalog versions
1.0 and 1.1 are supported by Spotless.
Unlike Eclipse, Spotless WTP ignores per default external URIs in schema location hints and
external entities. To allow the access of external URIs, set the property resolveExternalURI
to true.
homepage. changelog. Biome is a formatter that for the frontend written in Rust, which has a native binary, does not require Node.js and as such, is pretty fast. It can currently format JavaScript, TypeScript, JSX, and JSON, and may support more frontend languages such as CSS in the future.
You can use Biome in any language-specific format for supported languages, but usually you will be creating a generic format.
Note regarding biome: Biome supports formatting CSS as of 1.8.0 (experimental, opt-in) and 1.9.0 (stable).
spotless {
format 'styling', {
// you have to set the target manually
target 'src/*/webapp/**/*.js'
// Download Biome from the network if not already downloaded, see below for more info
biome('1.2.0')
// (optional) Path to the directory with the biome.json conig file
biome('1.2.0').configPath("path/config/dir")
// (optional) Biome will auto detect the language based on the file extension.
// See below for possible values.
biome('1.2.0').language("js")
}
}
To apply Biome to more kinds of files with a different configuration, just add more formats:
spotless {
format 'biome-js', {
target '**/*.js'
biome('1.2.0')
}
format 'biome-ts', {
target '**/*.ts'
biome('1.2.0')
}
format 'biome-json', {
target '**/*.json'
biome('1.2.0')
}
}
Limitations:
- The auto-discovery of config files (up the file tree) will not work when using Biome within spotless.
- The
ignore
option of thebiome.json
configuration file will not be applied. Include and exclude patterns are configured in the spotless configuration in the Gradle settings file instead.
Note: Due to a limitation of biome, if the name of a file matches a pattern in
the ignore
option of the specified biome.json
configuration file, it will not be
formatted, even if included in the biome configuration section of the Gradle settings
file.
You could specify a different biome.json
configuration file without an ignore
pattern to circumvent this.
Note 2: Biome is hard-coded to ignore certain special files, such as package.json
or tsconfig.json
. These files will never be formatted.
To format with Biome, spotless needs to find the Biome binary. By default, spotless downloads the binary for the given version from the network. This should be fine in most cases, but may not work e.g. when there is not connection to the internet.
To download the Biome binary from the network, just specify a version:
spotless {
format 'biome', {
target '**/*.js','**/*.ts','**/*.json'
biome('1.2.0')
}
}
Spotless uses a default version when you do not specify a version, but this
may change at any time, so we recommend that you always set the Biome version
you want to use. Optionally, you can also specify a directory for the downloaded
Biome binaries (defaults to ~/.m2/repository/com/diffplug/spotless/spotless-data/biome
):
spotless {
format 'biome', {
target '**/*.js','**/*.ts','**/*.json'
// Relative paths are resolved against the project's base directory
biome('1.2.0').downloadDir("${project.gradle.gradleUserHomeDir}/biome")
}
}
To use a fixed binary, omit the version
and specify a pathToExe
:
spotless {
format 'biome', {
target '**/*.js','**/*.ts','**/*.json'
biome().pathToExe("${project.layout.buildDirectory.asFile.get().absolutePath}/bin/biome")
}
}
Absolute paths are used as-is. Relative paths are resolved against the project's base directory. To use a pre-installed Biome binary on the user's path, specify just a name without any slashes / backslashes:
spotless {
format 'biome', {
target '**/*.js','**/*.ts','**/*.json'
// Uses the "biome" command, which must be on the user's path. -->
biome().pathToExe('biome')
}
}
Biome is a biased formatter and linter without many options, but there are a few basic options. Biome uses a file named biome.json for its configuration. When none is specified, the default configuration from Biome is used. To use a custom configuration:
spotless {
format 'biome', {
target '**/*.js','**/*.ts','**/*.json'
// Must point to the directory with the "biome.json" config file -->
// Relative paths are resolved against the project's base directory -->
biome('1.2.0').configPath('./config')
}
}
By default, Biome detects the language / syntax of the files to format automatically from the file extension. This may fail if your source code files have unusual extensions for some reason. If you are using the generic format, you can force a certain language like this:
spotless {
format 'biome', {
target 'src/**/typescript/**/*.mjson'
biome('1.2.0').language('json')
}
}
The following languages are currently recognized:
js
-- JavaScriptjsx
-- JavaScript + JSX (React)js?
-- JavaScript, with or without JSX, depending on the file extensionts
-- TypeScripttsx
-- TypeScript + JSX (React)ts?
-- TypeScript, with or without JSX, depending on the file extensionjson
-- JSON
Prettier, eclipse wtp, and license header are available in every format, and they each have their own section. As mentioned in the quickstart, there are a variety of simple generic steps which are also available in every format, here are examples of these:
spotless {
// run a native binary
format 'terraform', {
target 'src/**/*.tf', 'src/**/*.tfvars' // you have to set the target manually
nativeCmd('terraform', '/opt/homebrew/bin/terraform', ['fmt', '-']) // name, path to binary, additional arguments
}
}
If the license header (specified with licenseHeader
or licenseHeaderFile
) contains $YEAR
or $today.year
, then that token will be replaced with the current 4-digit year. For example, if Spotless is launched in 2020, then /* Licensed under Apache-2.0 $YEAR. */
will produce /* Licensed under Apache-2.0 2020. */
Once a file's license header has a valid year, whether it is a year (2020
) or a year range (2017-2020
), it will not be changed. If you want the date to be updated when it changes, enable the ratchetFrom
functionality, and the year will be automatically set to today's year according to the following table (assuming the current year is 2020):
- No license header ->
2020
2017
->2017-2020
2017-2019
->2017-2020
See the javadoc for a complete listing of options.
If your project has not been rigorous with copyright headers, and you'd like to use git history to repair this retroactively, you can do so with -PspotlessSetLicenseHeaderYearsFromGitHistory=true
. When run in this mode, Spotless will do an expensive search through git history for each file, and set the copyright header based on the oldest and youngest commits for that file. This is intended to be a one-off sort of thing.
Some files have fixed header lines (e.g. <?xml version="1.0" ...
in XMLs, or #!/bin/bash
in bash scripts). Comments cannot precede these, so the license header has to come after them, too.
To define what lines to skip at the beginning of such files, fill the skipLinesMatching
option with a regular expression that matches them (e.g. .skipLinesMatching("^#!.+?\$")
to skip shebangs).
If your project is not currently enforcing formatting, then it can be a noisy transition. Having a giant commit where every single file gets changed makes the history harder to read. To address this, you can use the ratchet
feature:
spotless {
ratchetFrom 'origin/main' // only format files which have changed since origin/main
In this mode, Spotless will apply only to files which have changed since origin/main
. You can ratchet from any point you want, even HEAD
. You can also set ratchetFrom
per-format if you prefer (e.g. spotless { java { ratchetFrom ...
).
However, we strongly recommend that you use a non-local branch, such as a tag or origin/main
. The problem with HEAD
or any local branch is that as soon as you commit a file, that is now the canonical formatting, even if it was formatted incorrectly. By instead specifying origin/main
or a tag, your CI server will fail unless every changed file is at least as good or better than it was before the change.
This is especially helpful for injecting accurate copyright dates using the license step.
Many popular CI systems (GitHub, GitLab, BitBucket, and Travis) use a "shallow clone". This means that ratchetFrom 'origin/main'
will fail with No such reference
. You can fix this by:
- calling
git fetch origin main
before you call Spotless - disabling the shallow clone like so
Sometimes there is a chunk of code which you have carefully handcrafted, and you would like to exclude just this one little part from getting clobbered by the autoformat. Some formatters have a way to do this, many don't, but who cares. If you setup your spotless like this:
spotless {
java { // or kotlin, or c, or python, or whatever
toggleOffOn()
Then whenever Spotless encounters a pair of spotless:off
/ spotless:on
, it will exclude the code between them from formatting, regardless of all other rules. If you want, you can change the tags to be whatever you want, e.g. toggleOffOn('fmt:off', 'fmt:on')
. If you decide to change the default, be sure to read this for some gotchas.
Spotless uses UTF-8 by default, but you can use any encoding which the JVM supports. You can set it globally, and you can also set it per-format.
spotless {
encoding 'UTF-8' // all formats will be interpreted as UTF-8
java {
encoding 'Cp1252' // except java, which will be Cp1252
Line endings can also be set globally or per-format using the lineEndings
property. Spotless supports four line ending modes: UNIX
, WINDOWS
, MAC_CLASSIC
, PLATFORM_NATIVE
, GIT_ATTRIBUTES
, and GIT_ATTRIBUTES_FAST_ALLSAME
. The default value is GIT_ATTRIBUTES_FAST_ALLSAME
, and we highly recommend that you do not change this value. Git has opinions about line endings, and if Spotless and git disagree, then you're going to have a bad time. FAST_ALLSAME
just means that Spotless can assume that every file being formatted has the same line endings (more info).
You can easily set the line endings of different files using a .gitattributes
file. Here's an example .gitattributes
which sets all files to unix newlines: * text eol=lf
.
As described in the quickstart, Spotless is just a set of files ("target"), passed through a list of String -> String
functions. The string each function gets will always have unix \n
endings, and Spotless doesn't care which endings the function provides back, it will renormalize regardless. You can easily make a new step directly in your buildscript, like so:
spotless {
format 'misc', {
custom 'lowercase', { str -> str.toLowerCase() }
However, custom rules will disable up-to-date checking and caching, unless you read this javadoc and follow its instructions carefully.
Another option is to create proper FormatterStep
in your buildSrc
, and then call addStep
. The contributing guide describes how to do this. If the step is generally-useful, we hope you'll open a PR to share it!
spotless {
format 'misc', {
addStep(MyFormatterStep.create())
Ideally, your formatter will be able to silently fix any problems that it finds, that's the beauty of the String -> String
model. But sometimes that's not possible. If you throw
AssertionError
or a subclass -> Spotless reports as a problem in the file being formatted- anything else -> Spotless reports as a bug in the formatter itself
If you want to have two independent java
blocks, you can do something like this:
spotless { java { ... } }
spotless { format 'javaFoo', com.diffplug.gradle.spotless.JavaExtension, { ... } }
// has to be 'javaFoo' not 'java' because each format needs a unique name
That's how the real spotless { java {
works anyway. As a follow-on, you can make your own subclass to FormatExtension
in the buildSrc
directory, and then use it in your buildscript like so:
spotless {
format 'foo', com.acme.FooLanguageExtension, {
If you'd like to create a one-off Spotless task outside of the check
/apply
framework, see FormatExtension.createIndependentApplyTask
.
In very rare cases, you might want to format e.g. javascript which is written inside JSP templates, or maybe java within a markdown file, or something wacky like that. You can specify hunks within a file using either open/close tags or a regex with a single capturing group, and then specify rules within it, like so. See javadoc for more details.
import com.diffplug.gradle.spotless.JavaExtension
spotless {
format 'templates', {
target 'src/templates/**/*.foo.html'
prettier().config(['parser': 'html'])
withinBlocks 'javascript block', '<script>', '</script>', {
prettier().config(['parser': 'javascript'])
}
withinBlocksRegex 'single-line @(java-expresion)', '@\\((.*?)\\)', JavaExtension, {
googleJavaFormat()
}
The check
task is Gradle's built-in task for grouping all verification tasks - unit tests, static analysis, etc. By default, spotlessCheck
is added as a dependency to check
.
You might want to disable this behavior. We recommend against this, but it's easy to do if you'd like:
spotless {
enforceCheck false
}
When a misformatted file throws an exception, it will be for one of two reasons:
- Spotless calculated the properly formatted version, and it is different than the current contents.
- One of the formatters threw an exception while attempting to calculate the properly formatted version.
You can fix (1) by excluding the file from formatting using the targetExclude
method, see the quickstart section for details. You can fix (2) and turn these exceptions into warnings like this:
spotless {
java {
custom 'my-glitchy-step', { ... }
ignoreErrorForStep('my-glitchy-step') // ignore errors on all files thrown by a specific step
ignoreErrorForPath('path/to/file.java') // ignore errors by all steps on this specific file
By default, Spotless resolves dependencies on a per-project basis. For very large parallel builds, this can sometimes cause problems. As an alternative, Spotless can be configured to resolve all dependencies in the root project like so:
spotless {
...
predeclareDeps()
}
spotlessPredeclare {
java { eclipse() }
kotlin { ktfmt('0.28') }
}
Alternatively, you can also use predeclareDepsFromBuildscript()
to resolve the dependencies from the buildscript repositories rather than the project repositories.
If you use this feature, you will get an error if you use a formatter in a subproject which is not declared in the spotlessPredeclare
block.
- Save your working tree with
git add -A
, thengit commit -m "Checkpoint before spotless."
- Run
gradlew spotlessApply
- View the changes with
git diff
- If you don't like what spotless did,
git reset --hard
- If you'd like to remove the "checkpoint" commit,
git reset --soft head~1
will make the checkpoint commit "disappear" from history, but keeps the changes in your working directory.
- A few thousand github projects
- JUnit 5 (aka JUnit Lambda)
- Apache Beam (direct link to spotless section in its build.gradle)
- opentest4j
- Durian (direct link to spotless section in its build.gradle)
- DurianRx (direct link to spotless section in its build.gradle)
- DurianSwt (direct link to spotless section in its build.gradle)
- MatConsoleCtl (direct link to spotless section in its build.gradle)
- MatFileRW (direct link to spotless section in its build.gradle)
- Goomph (direct link to spotless section in its build.gradle)
- FreshMark (direct link to spotless section in its build.gradle)
- JScriptBox (direct link to spotless section in its build.gradle)
- (Your project here)