Markdoc language server
This is the official Visual Studio Code extension and language server for the Markdoc authoring framework.
When the language server and extension are used together and configured to load a Markdoc schema, they support the following features:
- Syntax highlighting for Markdoc tags inside of Markdown content
- Autocompletion of Markdoc tags and attribute values declared in the schema
- Autocompletion of links based on routes declared in document frontmatter
- Clicking through partial tags and document links to open the corresponding file
- "Peeking" at the underlying file for a partial tag inside of a document
- Identifying all of the references to a partial to show where it is used
- Code folding for content ranges within block-level Markdoc tags
- Formatting ranges and whole documents using Markdoc's built-in formatter
- Validating the document against the user's schema and displaying inline error indicators
- Displaying error messages for malformed Markdoc tag syntax and structural errors like unmatched opening and closing tags
- Creating new Markdoc files from user-defined templates via the new file menu
- Linked editing for matched opening and closing tag names
Configuration quickstart
After installing the Markdoc extension in Visual Studio Code, create a Markdoc language server configuration file. The extension looks for a file called markdoc.config.json
in your workspace root, but you can customize this in the extension's settings.
The JSON configuration file consists of an array of server instance descriptions. The following example demonstrates how to create a basic Markdoc language server configuration:
[
{
"id": "my-site",
"path": "docs/content",
"schema": {
"path": "docs/dist/schema.js",
"type": "node",
"property": "default",
"watch": true
},
"routing": {
"frontmatter": "route"
}
}
]
- The
id
property is a string that uniquely identifies the server configuration. In Visual Studio Code, the extension displays this in the status bar when a file from the configuration has focus
- The
path
property is a string that contains the path to the base directory where your Markdoc content is stored, relative to your workspace root directory
- The
schema
property is an object that describes the location of your Markdoc schema and how to load it from the filesystem
- The
path
property is string that contains the path to the schema module, relative to the workspace root directory
- The
type
property is a string that must be either "node"
or "esm"
. This property indicates whether the schema module uses node's module.exports
convention or the standards-based ECMAScript modules export
syntax
- The
property
property is the name of the property exported by the module that contains the schema.
- When this property is omitted and the type is set to
"node"
, the extension automatically assumes that the schema itself is the top-level value of module.exports
.
- When the property is omitted and the type is set to
"esm"
, the extension automatically assumes the schema itself is the default export (e.g. export default {}
)
- The
watch
property is a boolean value that, when set to true
, configures the extension to monitor the schema file for changes and reload it automatically
- The
routing
property is an optional object that describes your project's routing configuration
- The
frontmatter
property is a string that tells the extension which property in the Markdoc file's YAML frontmatter contains the URL route associated with the file
Standalone Server Configuration
When using the server standalone without the client VS Code extension you can pass the configuration object to your LSP client like so:
{
"root": "/path/to/project/root",
"path": "relative/path/to/markdoc/files",
"config": {
"root": "/path/to/project/root",
"path": "relative/path/to/markdoc/files"
}
}
Invoke the server with markdoc-ls --stdio
from within your LSP client.
File extensions
In order to distinguish Markdoc files from Markdown files, the Visual Studio Code extension expects Markdoc files to use one of the following file extensions: .markdoc
, .markdoc.md
, or .mdoc
.
It does not recognize .md
files by default. If you would like to still use a .md
extension, you can instead modify your Visual Studio Code workspace configuration with a custom file association. Add the following content to your .vscode/settings.json
file:
{
"files.associations": {
"*.md": "markdoc"
}
}
Advanced configuration
It is possible to have multiple Markdoc configurations for the same workspace by adding additional configuration objects to the top-level array. This is useful in cases where you have multiple websites with different schemas under different subdirectories of the same workspace. For example, you might want separate configurations for narrative documentation and an API reference.
In multi-root workspaces, a Markdoc configuration file is specific to an individual workspace root. You can have separate Markdoc configuration files for each root. If you need to override the location of the Markdoc language server configuration file in a multi-root workspace, you can use a folder setting to customize this behavior per root.
Extending the language server with custom functionality
The language server is published as a package on npm in order to support extensibility and customization. You can tailor the functionality to your needs by adding plugins or creating subclasses that substitute built-in plugins. Support for this is somewhat experimental and the APIs exposed by the package are still subject to change. We will not guarantee API stability or backwards compatibility for language server plugins until the 1.0 release.
Contributing
Contributions and feedback are welcomed and encouraged. Feel free to open PRs here, or open issues and discussion threads in the Markdoc core repo.
Building from source
$ npm install
$ (cd server && npm install) && (cd client && npm install)
$ npm run build
$ npm run build:types
$ npm run build:extension
Running unit tests
The test suite relies on the 'node:test' module that is only included in Node.js 18.x or higher.
$ npm run test
License
This project uses the MIT license.