TypeScript icon, indicating that this package has built-in type declarations

2.0.0 • Public • Published


🎉 This package has been adopted by the vsce-toolroom GitHub collective.

This package is in maintenance mode & the technology is superseded by vscode-anycode, a quicker language service which leverages the tree-sitter symbolic-expression parser technology.

Generate language service providers driven entirely by your Textmate grammar and one configuration file.

In order to generate language providers from this module, the Textmate grammar must include the following features:

  • meta declaration scopes for block level declarations
  • variable assignment scopes differentiated between multiple and single
  • granular keyword control tokens with begin and end scopes


npm install vscode-textmate-languageservice

Browser support:

  • This package supports Webpack and ESBuild.
  • If you use a bundler, you need to set crypto as a external (commonjs crypto one in webpack). This allows the library to avoid polyfilling the node:crypto module.


This package is stable with browser compatibility (1.1.0). But I recommend you watch out for tree-sitter native integration into vscode (issue). Maintainable & with faster retokenization, it is a Holy Grail ...

Whereas this package depends on a well-written Textmate grammar and is a band aid of sorts.

If there is native vscode support for the language, find a Tree-sitter syntax online then suggest it in an Anycode issue. Otherwise, please open an issue on the community-maintained Treesitter syntax highlighter extension and someone might deal with it.


  • Language contribution and grammar contribution defined via contributes in the extension manifest (or textmate-languageservice-contributes).
  • Your grammar is bundled in the extension source code and is consumable by vscode-textmate (which can load PList XML, JSON or YAML grammars).
  • A configuration file is available in the extension, defaulting to ./textmate-configuration.json. You can also use textmate-languageservices property of package.json to map language ID to relative path.

Example language extension manifest - ./package.json:

	"name": "lua",
	"displayName": "Textmate language service for Lua",
	"description": "Lua enhanced support for Visual Studio Code",
	"version": "0.0.1",
	"publisher": "",
	"license": "",
	"engines": {
		"vscode": "^1.51.1"
	"categories": [
		"Programming Languages"
	"contributes": {
		"languages": [{
			"id": "lua",
			"aliases": ["Lua"],
			"extensions": [".lua", ".moon", ".luau"],
			"configuration": "./language-configuration.json"
		"grammars": [{
			"language": "lua",
			"scopeName": "source.lua",
			"path": "./syntaxes/Lua.tmLanguage.json"


Create a JSON file named textmate-configuration.json in the extension directory. The file accepts comments and trailing commas.

If you only want to use the document and/or tokenization services, this file can be as simple as {}.

Textmate configuration fields:

  • assignment - optional (object)
    Collection of Textmate scope selectors for variable assignment scopes when including variable symbols:
    • separator: Token to separate multiple assignments (string)
    • single: Token to match single variable assignment. (string)
    • multiple: Token to match multiple variable assignment. (string)
  • declarations - optional (array)
    List of Textmate scope selectors for declaration token scopes.
  • dedentation - optional (array)
    List of Textmate tokens for dedented code block declarations (e.g. ELSE, ELSEIF).
    Tokens still need to be listed in indentation with the decrementing value -1.
  • exclude (string) VS Code glob pattern for files to exclude from workspace symbol search.
  • indentation - optional (object)
    Indentation level offset for Textmate token types (used to implement folding).
  • punctuation - optional (object)
    Collection of punctuation tokens with a significant effect on syntax providers. Properties:
    • continuation: Token scope selector for line continuation (to use in region matching). (string)
  • markers - optional (object)
    Stringified regular expression patterns for folding region comments.
    • start: Escaped regular expression for start region marker. (string)
    • end: Escaped regular expression for end region marker. (string) Properties:
  • symbols - optional (object)
    Map of document symbol tokens to their symbol kind (vscode.SymbolKind value).

Configuration examples

Template for textmate-configuration.json file:

  "assignment": {
    "single": "",
    "multiple": "",
    "separator": ""
  "declarations": [],
  "dedentation": [
  "exclude": "{.modules,.includes}/**",
  "indentation": {
    "punctuation.definition.comment.begin.custom": 1,
    "punctuation.definition.comment.end.custom": -1,
    "keyword.control.begin.custom": 1,
    "keyword.control.end.custom": -1
  "punctuation": {
    "continuation": "punctuation.separator.continuation.line.custom"
  "markers": {
    "start": "^\\s*#?region\\b",
    "end": "^\\s*#?end\\s?region\\b"
  "symbols": {
    "keyword.control.custom": 2,
    "entity.name.function.custom": 11

An example configuration file that targets Lua:

  "assignment": {
    "single": "meta.assignment.variable.single.lua",
    "multiple": "meta.assignment.variable.group.lua",
    "separator": "punctuation.separator.comma.lua"
  "declarations": [
    "meta.declaration.lua entity.name",
    "meta.assignment.definition.lua entity.name"
  "dedentation": [
  "exclude": "{.luarocks,lua_modules}/**",
  "indentation": {
    "punctuation.definition.comment.begin.lua": 1,
    "punctuation.definition.comment.end.lua": -1,
    "keyword.control.begin.lua": 1,
    "keyword.control.end.lua": -1
  "markers": {
    "start": "^\\s*#?region\\b",
    "end": "^\\s*#?end\\s?region\\b"
  "symbols": {
    "keyword.control.lua": 2,
    "entity.name.function.lua": 11


Language extension

Extension code sample - ./src/extension.ts:

import TextmateLanguageService from 'vscode-textmate-languageservice';

export async function activate(context: vscode.ExtensionContext) {
	const selector: vscode.DocumentSelector = 'lua';
	const textmateService = new TextmateLanguageService(selector, context);

	const foldingRangeProvider = await textmateService.createFoldingRangeProvider();
	const documentSymbolProvider = await textmateService.createDocumentSymbolProvider();
	const workspaceSymbolProvider = await textmateService.createWorkspaceSymbolProvider();
	const definitionProvider = await textmateService.createDefinitionProvider();

	context.subscriptions.push(vscode.languages.registerDocumentSymbolProvider(selector, documentSymbolProvider));
	context.subscriptions.push(vscode.languages.registerFoldingRangeProvider(selector, foldingRangeProvider));
	context.subscriptions.push(vscode.languages.registerDefinitionProvider(selector, peekDefinitionProvider));


Extension code sample - ./src/extension.ts:

import TextmateLanguageService from 'vscode-textmate-languageservice';

export async function activate(context: vscode.ExtensionContext) {
    const selector: vscode.DocumentSelector = 'custom';
    const textmateService = new TextmateLanguageService('custom', context);
    const textmateTokenService = await textmateService.initTokenService();
    const textDocument = vscode.window.activeTextEditor!.document;
    const tokens = textmateTokenService.fetch(textDocument);

NB: If you would like to:

You can use the custom "textmate-languageservice-contributes" property in package.json:

	"textmate-languageservice-contributes": {
		"languages": [{
			"id": "typescript",
			"aliases": ["TypeScript"],
			"extensions": [".ts", ".tsx", ".cts", ".mts"]
		"grammars": [{
			"language": "typescript",
			"scopeName": "source.ts",
			"path": "./syntaxes/TypeScript.tmLanguage.json"

Use Oniguruma WASM buffer

This is the vscode-oniguruma build of Oniguruma written in C, compiled to WASM format with memory hooks to V8.

This is not streaming 🙁 but vscode libs must bundle WebAssembly deps so as to support web ecosystem.

import TextmateLanguageService from 'vscode-textmate-languageservice';
const onigurumaPromise = TextmateLanguageService.utils.getOniguruma();


DownloadsWeekly Downloads






Unpacked Size

1.18 MB

Total Files


Last publish


  • zm-sdet