3.0.1

• Hotfix for type definitions missing in 3.0.0.
• Smoke test types to ensure package build always includes type declarations.

3.0.0

• [BREAKING] Rename api.getLanguageConfiguration to api.getLanguageContribution.
• [BREAKING] Rename api.getGrammarConfiguration to api.getGrammarContribution.
• Add getLanguageConfiguration API method to load vscode.LanguageConfiguration.
• Add plaintext language tokenization and grammar resolution.
• Hotfix for "unrecognized language" error for plaintext documents in API token methods.

2.0.0-recovery.1

Document API additions made in 2.0.0

2.0.0

• The VSCE Toolroom open-source collective has adopted the Textmate language service project!
• Redesigned the logo, inspired by the V8 engine and the Textmate osteopermum flower.
• Added support for tokenization and Language providers from built-in grammars as well as service-only grammars.
• Marked TextmateLanguageService~context parameter as optional in the API types.
• Marked the API from 1.0.0 as compatible with 1.55.0, not 1.51.0.
• Provided community resolution to microsoft/vscode#109919 & microsoft/vscode#99356.
• Implemented API methods in an api namespace for developer-friendly logic:
  • Add getTokenInformationAtPosition method for fast positional token polyfill: vscode.TokenInformation.
  • Add getScopeInformationAtPosition method to get Textmate token data: TextmateToken.
  • Add getScopeRangeAtPosition method to get token range: vscode.Range.
  • Add getLanguageConfiguration method for language configuration: LanguageDefinition.
  • Add getGrammarConfiguration method to get language grammar wiring: GrammarLanguageDefinition.
  • Add getContributorExtension method to get extension source of language ID: vscode.Extension.
• Linted the Textmate scope parser correctly & automatically in the test pipeline.
• Added getOniguruma to API utilities, a browser-ready non-streaming build of vscode-oniguruma.

1.2.0

• Add support for creation of tokenization or light document service.
  • TL:DR; swap the "contributes" key with a 'fake' "textmate-language-contributes" key in package.json.
  • Now possible to wire up a fake language and grammar "contribution" to a package service.
• Add service-only tests for TypeScript in the test suite.
• Use TextmateLanguageService as global key instead of LSP in service workers.
  • "LSP" is an cross-process and IDE-agnostic message format/standard for language feature data.
  • This library's just a factory for language feature services in VS Code.
  • This change is not breaking thanks to Webpack.
• Improve diff generation for error logging in the sample output validator that's used to test feature providers.
• Add keywords to the NPM package's metadata for better search engine discovery.
• Skip web testing of vscode.DefinitionProvider and vscode.WorkspaceSymbolProvider factory methods.

1.1.0

• Bundle files using Webpack for performance boost.
• Add browser production support (bundle onig.wasm using encoded-uint8array-loader & prevent reliance on fetch).
• Restore typing declaration files for dependent extension consumers.
• Fix broken Gitlab pipeline so we have CI testing again.
• Fix line number collisions between entry symbols in the outline service.
• Fix container name in symbol information output for the document symbol provider.
• Upgrade vscode-textmate from 7.0.4 to 9.0.0 (microsoft/vscode-textmate#198).
• Ignore test files from package before npm publish to reduce size by ~20%.
• Add a test suite runner for testing compatibility with dependent web extensions.
• Improve test suite performance by 20% by removing dependencies & bundling.

1.0.0 (stable)

• Achieved browser readiness by handling hashing better. We use native hashing of file text contents to keep it fast.
• Upgrade from SHA-1 algorithm (a famous collision-attack vector) and adopt stable SHA-256 alternatives.
• Remove last external dependency (git-sha1) so we don't need a bundler.

1.0.0-rc-2

• Fix the line number in the folding provider for top-level declaration folds after the first declaration.
• Add browser production support with a cost-benefit tradeoff... we now load onig.wasm (Textmate grammar regex parser) without streaming.
• Remove any system dependencies in the test scripts. Plus the scripts use the CLIs better & are much cleaner.
• Migrate previous CI workflow pipeline to Gitlab.

1.0.0-rc-1

• Switch API to async factory class format that generates providers and exposes services.
• Introduce optional top-level "textmate-languageservices" map of language ID to path, supporting extension manifests with multiple configured languages.
  We default to ./textmate-configuration.json for one language.
• Rewrite configuration loading and service caching to occur asynchronously using native VS Code APIs.
• Rewrite folding provider to remove performance overheads in header & block folding.
• Fix external file search matching in the definition provider (which was searching files matching ./Symbol.ext before).
• Fix line token incrementation for decremented lines in the tokenizer service (FKA TextmateEngine).
• Add performance layer to Textmate scope selector parser to bypass the need for a WASM parser.