IntelliJ Language Plugin

IntelliJ Platform Plugin SDK is the name of the Java API that we can use to write plugin for IntelliJ-based IDEs.

Language Plugins are plugin adding support for a language, such as R, Rust, or OCaml. The best way to learn is to see how others implemented the API "extension points": intellij-platform-explorer.

πŸ“š A plugin using the Language Server Protocol (LSP) maybe a more viable alternative than implementing your own logic.

A few useful links:

Get Started

You can use the Template or directly inside the IntelliJ, you can create IDE Plugin project. I used the latter which is a bit different.

Open the Manifest file at: src/main/resources/META-INF/plugin.xml and define the basic information about the plugin:

    <name>Plugin Name</name>
    <vendor email="" url="">YourCompany</vendor>
        Enter short description for your plugin here.<br>
        <em>most HTML tags may be used</em>

You can also add your plugin icon(s):

  • src/main/resources/META-INF/pluginIcon.svg
  • src/main/resources/META-INF/pluginIcon_dark.svg

Run the Gradle task Run Plugin and see if your plugin is in the list of the new IDE that opened, and each information is properly set.

Basic Utilities


Create a class to centralize access to icons.

Useful links: Working with Icons and Images, IntelliJ Platform UI Guidelines, IntelliJ Platform Icons, intellij-icon-generator.

 * @see com.intellij.icons.AllIcons
 * @see com.intellij.icons.AllIcons.FileTypes
 * @see com.intellij.icons.AllIcons.General
 * @see com.intellij.icons.AllIcons.Gutter
 * @see com.intellij.icons.AllIcons.Nodes
 object OCamlIcons {
    private fun loadIcon(path: String): Icon {
        return IconLoader.getIcon(path,

    object FileTypes {
        val OCAML_SOURCE = loadIcon("/icons/mlFile.svg")


IntelliJ recommends avoiding hard-coded translatable texts. We store translatable strings in a file, and access them from the code:

import com.intellij.DynamicBundle
import org.jetbrains.annotations.PropertyKey

private const val BUNDLE = "messages.OCamlBundle"

// ex: OCamlBundle.message("")
object OCamlBundle : DynamicBundle(BUNDLE) {
    fun message(@PropertyKey(resourceBundle = BUNDLE) key: String,
                vararg params: Any): String {
        return getMessage(key, *params)

With src/main/resources/messages/

# Filetype file

Defining File Types

Add a fileType with all information that is also defined in the class com.ocaml.ide.files.OCamlFileType.

    <extensions defaultExtensionNs="com.intellij">
        <!-- FILES -->
        <fileType extensions="ml" hashBangs="ml" name="OCaml File" language="OCaml"
                  fieldName="INSTANCE" implementationClass="com.ocaml.ide.files.OCamlFileType"/>

You may have to define the language of the file:

import com.intellij.lang.Language

object OCamlLanguage : Language("OCaml") {
    private fun readResolve(): Any = OCamlLanguage
    override fun getDisplayName(): String = "OCaml"
    override fun isCaseSensitive(): Boolean = true

🍰 You should see an icon in files having the selected extension.

πŸ‘‰ You can use Language(OCamlLanguage, "OCamli") to declare an language inheriting from OCamlLanguage (same highlighter, etc.).

Using variables for the extension is not needed, but it's useful if we need this information in other classes.

import com.intellij.openapi.fileTypes.LanguageFileType
import com.ocaml.OCamlBundle
import com.ocaml.icons.OCamlIcons
import javax.swing.Icon

object OCamlFileType : LanguageFileType(OCamlLanguage) {
    // Extension-Related Constants
    private const val DEFAULT_EXTENSION = "ml"

    // LanguageFileType implementation
    override fun getName(): String  = "OCaml File"
    override fun getDescription(): String = OCamlBundle.message("")
    override fun getDefaultExtension(): String = DEFAULT_EXTENSION
    override fun getIcon(): Icon = OCamlIcons.FileTypes.OCAML_SOURCE
    override fun getDisplayName(): String  = description

⚠️ There is no INSTANCE in the Kotlin Class as it's a static object.

Language Parsing

JetBrains uses the Backus–Naur form (BNF) language to define the parsing rules of a language. We create a .bnf syntax file, such as OCaml.bnf, and use JetBrains Grammar-Kit to generate a parser from it.

This section is the hardest ⚠️. The OCaml Manual along the extended page describe language elements and explanations in a BNF-like format, which you may use as a reference.

// Left: an element of the language
// Right: how the element is structured
file ::= expressions*

// Braces can be used to group expressions
// "?" means the expression is optional
// "*" means "0+" times the expression
// "+" means "1+" times the expression
// "|" is used for each alternative structure
expression ::= { A ";" B } * // Must quote characters
| (A | B)*                   // Alternative to braces
| another_expression         // Reference another element

// Brackets can be used to group an optional sequence
another_expression ::= [A B]
| (A B)?            // Alternative to brackets

⚠️ An element can reference itself as long as there is no "left recursive" call, e.g. the first element is not itself.

⚠️ The parser is applied on the Lexed tokens!

At the start of the file, we may configure the parser generation. Refer to the Generated Files Section to learn more about these.

    // Parser class
    // Generated class with all elements+tokens instances

    // Generated files prefix/suffix
    // Where to store generated files

    // Base classes for elements/tokens instances
    // That we can access from the "elementTypeHolderClass"
        AND = "and"
    // ...

Inside the tokens, you can define the language tokens. For instance, SEMI=";" means you can use SEMI instead of ";". See also: Lexer.

Generated Parser Files

Right-click on the BNF syntax file to generate the parser code.

Psi Type Holder

The tool will also generate a elementTypeHolderClass with every an element type for each token and element of the language, allowing us to reference an element or a token in the code.

IElementType COMMENT = new OCamlTokenType("COMMENT");

The type of the tokens/elements in the type folder class can be customized using elementTypeClass and tokenTypeClass.

Example of Type Classes
// Elements are Composite Elements
// elementTypeClass="com.ocaml.language.psi.OCamlElementType"
class OCamlElementType(debugName: String) : IElementType(debugName, OCamlLanguage), ICompositeElementType {
    override fun createCompositeNode(): ASTNode {
        return OCamlTypes.Factory.createElement(this)
// Tokens are IElementType tokens 
// tokenTypeClass="com.ocaml.language.psi.OCamlTokenType"
class OCamlTokenType(debugName: String) : IElementType(debugName, OCamlLanguage)

πŸ‘‰ See also: generateTokens=true (default).

Psi Files And Interfaces

The tool will generate an interface (stored in psiPackage) and its implementation (stored in psiImplPackage) for every element.

You can exclude some by marking them as private:

private abc ::= /* some definition */

Each interface will define getters allowing us to fetch the children elements. Methods/types are computed from the whole rule.

All interfaces extend PsiElement, but you can use your own interface:

// Add: implements='com.ocaml.language.psi.api.OCamlElement'
interface OCamlElement : PsiElement, UserDataHolderEx

All implementations extend ASTWrapperPsiElement, but you can use your own class:

// Add: extends='com.ocaml.language.psi.api.OCamlElementImpl'
abstract class OCamlElementImpl(type: IElementType) : CompositePsiElement(type)

πŸ‘‰ See also: generateTokenAccessors=false (default).

Lexical Analysis

Before parsing a file, we convert the stream of characters to a stream of tokens using a Lexer. Right-click on the BNF syntax file to generate the base lexer file. For instance, if we have /* xxx */ in the file, the lexer could return that this text is a COMMENT token.

  • Integers (5, 0x5, 0X5, 0xF, etc.)
  • Floats (5., 5.0, 0x9.5, etc.)
  • Chars (escape characters, etc.)
  • Strings (multilines string, '' and "", escape characters, etc.)
  • Comments (Multiline comments, documentation comments, etc.)
  • Operators (";", "(", ")", etc.)
  • Keywords ("and", "if", "fun", etc.)
  • ...

Right-click on the lexer file to generate a Lexer. We don't use the lexer directly in the code, so create an adapter. It should have the same package path as generated lexer class.

import com.intellij.lexer.FlexAdapter // generated

class OCamlLexerAdapter : FlexAdapter(_OCamlLexer(null))

An overview of a lexing file:

[...]          // class definition (omitted)
DIGIT=[0-9]    // variables
%state INITIAL // parsing states
%state IN_XXX
<INITIAL> { // Link to the Type Holder Variables
   "and"       { return AND; } // OCamlTypes.AND
   "A" "B"     { return AB; }  // Same as "AB"
   // You can use variables, etc.
   {DIGIT}* ("i"|"I")? { return INTEGER_VALUE; }
   // you can move to custom state for complex tokens
   "/*" { yybegin(XXX); /* ... */ }

<IN_XXX> {
    "*/" { /*...*/; return BLOCK_COMMENT; }
    . { }
    <<EOF>> { /*...*/ }

// Every state can have a default case
[^] { return BAD_CHARACTER; }

Parser Definition

The parser definition connects all classes defined above:

internal class OCamlParserDefinition : ParserDefinition {
    override fun createLexer(project: Project?): Lexer = OCamlLexerAdapter()
    override fun getCommentTokens(): TokenSet = ParserDefinitionUtils.COMMENTS
    override fun getStringLiteralElements(): TokenSet = ParserDefinitionUtils.STRINGS
    override fun createParser(project: Project?): PsiParser = OCamlParser()
    override fun getFileNodeType(): IFileElementType = ParserDefinitionUtils.FILE
    override fun createFile(viewProvider: FileViewProvider): PsiFile = OCamlFile(viewProvider)
    override fun createElement(node: ASTNode?): PsiElement = OCamlTypes.Factory.createElement(node)

    object ParserDefinitionUtils {
        val FILE = IFileElementType(OCamlLanguage)
        val COMMENTS = TokenSet.create(OCamlTypes.COMMENT)
        val STRINGS = TokenSet.create(OCamlTypes.STRING_VALUE)

πŸ“š Add to getCommentTokens() any token that must be ignored by the parser, e.g., treated as a comment.

⚠️ Use OCamlTypes.Factory.createElement(node.elementType) when using CompositePsiElement.

The OCamlFile class:

import com.intellij.extapi.psi.PsiFileBase
import com.intellij.openapi.fileTypes.FileType
import com.intellij.psi.FileViewProvider
import com.ocaml.ide.files.OCamlFileType
import com.ocaml.ide.files.OCamlLanguage

class OCamlFile(viewProvider: FileViewProvider) : PsiFileBase(viewProvider, OCamlLanguage) {
    override fun getFileType(): FileType = OCamlFileType
    override fun toString(): String = "OCaml File"

Add to the Manifest:

    <extensions defaultExtensionNs="com.intellij">
        <!-- PARSER -->
        <lang.parserDefinition language="OCaml" implementationClass="com.ocaml.language.parser.OCamlParserDefinition"/>

Syntax Highlighter

The syntax highlighter is coloring the tokens that were lexed by the Lexer. For parsed elements, we need to use an annotator.

First, define colors:

import com.intellij.openapi.editor.colors.TextAttributesKey
import com.intellij.openapi.options.OptionsBundle
import com.intellij.openapi.options.colors.AttributesDescriptor
import com.intellij.openapi.editor.DefaultLanguageHighlighterColors as Default

// Token(text_shown_in_settings, color_used)
// Use "//" in translations to group colors in folders
enum class OCamlColor(humanText: String, attr: TextAttributesKey? = null) {
    VARIABLE(OCamlBundle.message("settings.ocaml.color.variables.default"), Default.IDENTIFIER),
    // See also: OptionsBundle for existing texts
    BRACES(OptionsBundle.message("options.language.defaults.braces"), Default.BRACES),

    // For The Highlighter
    val textAttributesKey = TextAttributesKey.createTextAttributesKey("com.ocaml.$name", attr)
    // For Highlight Color Settings
    val attributesDescriptor = AttributesDescriptor(humanText, textAttributesKey)

Then we can write an highlighter:

import com.intellij.openapi.fileTypes.SyntaxHighlighter
import com.intellij.openapi.fileTypes.SyntaxHighlighterBase.pack
import com.intellij.psi.tree.IElementType

class OCamlSyntaxHighlighter : SyntaxHighlighter {
    override fun getHighlightingLexer(): Lexer = OCamlLexerAdapter()

    override fun getTokenHighlights(tokenType: IElementType): Array<TextAttributesKey> =

    companion object {
        fun map(tokenType: IElementType): OCamlColor? = when (tokenType) {
            CHAR_VALUE -> OCamlColor.CHAR
            // ...
            LPAREN, RPAREN -> OCamlColor.PARENTHESES
            // ...
            in OCAML_KEYWORDS -> OCamlColor.KEYWORD
            // ...
            else -> null

        private val OCAML_KEYWORDS = setOf<IElementType>(
            AS, CLASS, ELSE, FOR, IF, // ...

You need to create a factory and add it to the Manifest:

// <lang.syntaxHighlighterFactory language="..." implementationClass=""/>
class OCamlSyntaxHighlighterFactory : SyntaxHighlighterFactory() {
    override fun getSyntaxHighlighter(project: Project?, virtualFile: VirtualFile?): SyntaxHighlighter {
        return OCamlSyntaxHighlighter()

To add the highlight color settings page in Editor > Color Scheme:

// <colorSettingsPage implementation=""/>
import com.intellij.openapi.editor.colors.TextAttributesKey
import com.intellij.openapi.fileTypes.SyntaxHighlighter
import com.intellij.openapi.options.colors.AttributesDescriptor
import com.intellij.openapi.options.colors.ColorDescriptor
import com.intellij.openapi.options.colors.ColorSettingsPage
import javax.swing.Icon

class OCamlColorSettingsPage : ColorSettingsPage {
    override fun getDisplayName(): String = "OCaml"
    override fun getIcon(): Icon = OCamlIcons.FileTypes.OCAML_SOURCE
    override fun getAttributeDescriptors() = Constants.ATTRS
    override fun getColorDescriptors(): Array<ColorDescriptor> = ColorDescriptor.EMPTY_ARRAY
    override fun getHighlighter(): SyntaxHighlighter = OCamlSyntaxHighlighter()
    override fun getAdditionalHighlightingTagToDescriptorMap() = Constants.ANNOTATOR_TAGS
    override fun getDemoText() = Constants.DEMO_TEXT

    internal object Constants {
        val ATTRS: Array<AttributesDescriptor> = OCamlColor.values().map{ it.attributesDescriptor }.toTypedArray()
        val ANNOTATOR_TAGS: Map<String, TextAttributesKey> = OCamlColor.values().associateBy({ }, { it.textAttributesKey })
        val DEMO_TEXT: String by lazy {
            (* write some code *)


Customizing Parser Generated Classes

At the top of the file, you can define the extends/implements restrictions on elements and their interface.

    // We use a regex to select elements
    // implements: selected element will implement this interface
    // extends: selected elements implementation will inherit from this class
    // ⚠️ You can have multiple "extends" but an element
    // is only associated to the first extends that matches it

You can alternatively define properties locally, e.g., after an element.

element ::= /* some definition */
    name = "my_element" // Generate class: {Prefix}MyElement
    methods = [getName setName]
    implements = "com.intellij.psi.PsiNamedElement"
    //or: implements = ["com.intellij.psi.PsiNamedElement"]


Using methods, you can inject a method in the generated class. First, in the top-level block, add a psi implementation util class: psiImplUtilClass="".

The first argument is the interface of the associated element. You can add as many argument as you need.

internal object OCamlImplUtils {
    @JvmStatic // my_element ::= and prefix = "XXX"
    fun getName(myElement: XXXMyElement): String? {
        return myElement.text
    // ...


Instead of using method injection, we can mix the generated class with the contents of another "mixin" class.

  • The generated class will inherit from the mixin class
  • The generated class will explicitly copy every constructor
  • The mixin class must extend an element implementation class, such as the OCamlElementImpl in the examples above.
// OCamlLetBinding is a generated interface
// Allowing us to use the implementation methods in the mixin
abstract class LetBindingMixin : OCamlElementImpl, OCamlLetBinding {

    // example of declaring multiple constructors
    constructor(type: IElementType) : super(type) {}
    constructor(node: ASTNode) : super(node.elementType) {}

    // declare methods
    override fun getNameIdentifier(): PsiElement? {
        TODO("Not yet implemented")

    override fun getName(): String? = nameIdentifier?.text

    override fun setName(name: String): PsiElement {
        TODO("Not yet implemented")


  • PsiNameIdentifierOwner: elements that contain a name identifier, e.g., an element that has a name
  • PsiNamedElement: elements that have a name
  • NavigatablePsiElement: elements that we can navigate to

Ensure the relevant elements implement the relevant interfaces.

Structure View

The structure view is a tab listing all variables, types/classes, functions, etc. of a file. Add to the Manifest:

<lang.psiStructureViewFactory language="..." implementationClass=""/>

And create the entrypoint factory:

import com.intellij.ide.structureView.StructureViewBuilder
import com.intellij.ide.structureView.StructureViewModel
import com.intellij.ide.structureView.TreeBasedStructureViewBuilder
import com.intellij.lang.PsiStructureViewFactory
import com.intellij.openapi.editor.Editor
import com.intellij.psi.PsiFile

internal class OCamlStructureViewFactory : PsiStructureViewFactory {
    override fun getStructureViewBuilder(psiFile: PsiFile): StructureViewBuilder {
        return object : TreeBasedStructureViewBuilder() {
            override fun createStructureViewModel(editor: Editor?): StructureViewModel {
                return OCamlStructureViewModel(editor, psiFile)

And the model:

import com.intellij.ide.structureView.StructureViewModel
import com.intellij.ide.structureView.StructureViewModelBase
import com.intellij.ide.structureView.StructureViewTreeElement
import com.intellij.ide.util.treeView.smartTree.Filter
import com.intellij.ide.util.treeView.smartTree.Sorter
import com.intellij.openapi.editor.Editor
import com.intellij.psi.PsiFile

class OCamlStructureViewModel(editor: Editor?, psiFile: PsiFile) :
    StructureViewModelBase(psiFile, editor, OCamlStructureViewElement(psiFile)),
    StructureViewModel.ElementInfoProvider {

    override fun getSorters(): Array<Sorter> = arrayOf(Sorter.ALPHA_SORTER)
    override fun getFilters(): Array<Filter> = super.getFilters()
    override fun isAlwaysShowsPlus(element: StructureViewTreeElement): Boolean = element.value is OCamlFileBase // your Psi File Class
    override fun isAlwaysLeaf(element: StructureViewTreeElement): Boolean = when (element.value) {
        // list of generated interfaces of "always leaf" elements
        is OCamlLetBinding -> true
        else -> false

The structure view element requires quite a lot of language-specific code. The base "template" would be something like this:

class OCamlStructureViewElement(element: PsiElement) : StructureViewTreeElement {
    private val psiAnchor = TreeAnchorizer.getService().createAnchor(element)
    private val myElement: PsiElement? get() = TreeAnchorizer.getService().retrieveElement(psiAnchor) as? PsiElement
    private val childElements: List<PsiElement>
        get() {
            return when (val psi = myElement) {
                else -> emptyList()

    override fun getValue(): PsiElement? = myElement
    override fun navigate(requestFocus: Boolean) { (myElement as? Navigatable)?.navigate(requestFocus) }
    override fun canNavigate(): Boolean = (myElement as? Navigatable)?.canNavigate() == true
    override fun canNavigateToSource(): Boolean = (myElement as? Navigatable)?.canNavigateToSource() == true
    override fun getPresentation(): ItemPresentation {
        return  PresentationData("unknown", "", null, null)
    override fun getChildren(): Array<out TreeElement> =
        childElements.map2Array { OCamlStructureViewElement(it) }

You must then define how, starting from the PsiFile, we can get the list of children nodes in childElements.

// for instance, in OCaml (let_bindings="let x = 5 and y = 7"
is OCamlFile -> {
    psi.childrenOfType<OCamlLetBindings>() // all variables
// letBindingList=["x=5", "y=7"]
is OCamlLetBindings -> psi.letBindingList

And, you need to define how each element is rendered in the view:

override fun getPresentation(): ItemPresentation {
    return myElement?.let(::getPresentationForStructure)
        ?: PresentationData("unknown", null, null, null)
private fun getPresentationForStructure(psi: PsiElement): ItemPresentation {
    val presentation =  when(psi) { // text shown for each element
        is OCamlNamedElement ->
        else -> null
    val icon = when(psi) { // icon shown | implement it in each element
        else -> psi.getIcon(Iconable.ICON_FLAG_VISIBILITY)
    return PresentationData(presentation, null, icon, null)

➑️ See also related classes: SortableTreeElement, Queryable.

Random Features

Spell Checker

You can define which elements should be analyzed by the spell checker using: < language="..." implementationClass=""/>

import com.intellij.psi.PsiElement
import com.intellij.spellchecker.tokenizer.SpellcheckingStrategy
import com.intellij.spellchecker.tokenizer.Tokenizer
import com.ocaml.ide.files.OCamlLanguage
import com.ocaml.language.psi.OCamlTypes

class OCamlSpellcheckingStrategy : SpellcheckingStrategy() {
    override fun isMyContext(element: PsiElement) = OCamlLanguage.isKindOf(element.language)

    override fun getTokenizer(element: PsiElement?): Tokenizer<*> = when {
        element?.node?.elementType == OCamlTypes.STRING_LITERAL -> TEXT_TOKENIZER
        // add variables, etc.
        else -> super.getTokenizer(element)

The default tokenizer can handle comments, but it doesn't handle elements of the language such as STRINGS or variable names etc.

For instance, assuming that OCamlNameIdentifierOwner is an element that own an element that has a name, we would add element is OCamlNameIdentifierOwner -> OCamlNameIdentifierOwnerTokenizer.

We would then develop a custom tokenizer:

import com.intellij.spellchecker.inspections.IdentifierSplitter
import com.intellij.spellchecker.tokenizer.TokenConsumer
import com.intellij.spellchecker.tokenizer.Tokenizer

object OCamlNameIdentifierOwnerTokenizer : Tokenizer<OCamlNameIdentifierOwner>() {
    override fun tokenize(element: OCamlNameIdentifierOwner, consumer: TokenConsumer) {
        val identifier = element.nameIdentifier ?: return
        consumer.consumeToken(identifier, IdentifierSplitter.getInstance())
        // you can add a more complex logic here


We can use <lang.commenter language="..." implementationClass=""/> to support the comment line (CTRL+/) and the comment block (CTRL+SHIFT+/) features.

import com.intellij.lang.Commenter

class OCamlCommenter : Commenter {
    override fun getLineCommentPrefix(): String? = "//"
    override fun getBlockCommentPrefix(): String  = "/*"
    override fun getBlockCommentSuffix(): String  = "*/"
    override fun getCommentedBlockCommentPrefix(): String  = "/*"
    override fun getCommentedBlockCommentSuffix(): String  = "*/"

It automatically supports uncommenting. Related classes: CommenterDataHolder, CustomUncommenter, SelfManagingCommenter<T>.

Braces Matching

You can add support to automatically detect the matching brace/token using <lang.braceMatcher language="..." implementationClass=""/>:

package com.ocaml.ide.typing

import com.intellij.lang.BracePair
import com.intellij.lang.PairedBraceMatcher
import com.intellij.psi.PsiFile
import com.intellij.psi.tree.IElementType
import com.ocaml.language.psi.OCamlTypes

class OCamlBraceMatcher : PairedBraceMatcher {
    override fun getPairs() = Constants.PAIRS

    override fun isPairedBracesAllowedBeforeType(lbraceType: IElementType, next: IElementType?): Boolean = 
    override fun getCodeConstructStart(file: PsiFile?, openingBraceOffset: Int): Int = 

    internal object Constants {
        val PAIRS: Array<BracePair> = arrayOf(
            BracePair(OCamlTypes.LBRACE, OCamlTypes.RBRACE, false),
            BracePair(OCamlTypes.BEGIN, OCamlTypes.END, false),
            // ...

πŸ‘» To-do πŸ‘»

Stuff that I found, but never read/used yet.

idea {
  module {


Split the XML into sub-files
+<idea-plugin xmlns:xi="">

+<xi:include href="/META-INF/other.xml" xpointer="xpointer(/idea-plugin/*)"/>
Set the sandbox Directory

➑️ If you change the IDE version often, you might want to use different sandbox, to start where you left.


intellij {
Add new folders as src/res


sourceSets { = ["src/xxx", "src/main"] += ["src/yyy"]
    main.resources.srcDirs = ["resources/main", "resources/zzz"] = ["test/xxx/", "test/main"]
    test.resources.srcDirs = ["resources/main", "resources/xxx", "test/testData"]


sourceSets {
    main.configure {


  • IStubElementType
  • stubClass=""
  • elementTypeFactory="""

BNF Grammar File

elementTypeFactory(".*") = ""
consumeTokenMethod(".*") = "consumeTokenFast"




import com.intellij.DynamicBundle
import org.jetbrains.annotations.NonNls
import org.jetbrains.annotations.PropertyKey

private const val BUNDLE = ""

object XXXBundle : DynamicBundle(BUNDLE) {

    fun message(@PropertyKey(resourceBundle = BUNDLE) key: String, vararg params: Any) =
        getMessage(key, *params)

    @Suppress("SpreadOperator", "unused")
    fun messagePointer(@PropertyKey(resourceBundle = BUNDLE) key: String, vararg params: Any) =
        getLazyMessage(key, *params)