@codingame/languageserver-mutualized

@codingame/languageserver-mutualized ·

NPM module to plug multiple language clients on a single language server.

What you need to know:

• Only initialization request of the first client will be used, so it's a good idea for all client to have the exact same capabilities and initialize params
• Some features won't work (configuration requests, execute command server request...)

What it allows you to do:

• Plug clients on different independant files
• Plug clients on the same files, then you need to synchronize the content of the editor between clients

How it works:

• If a file is open by at least one client, it's open on the server
• Every time a client change the content of a file, we stack their changes and flush it after 500ms of inactivity
• As soon a there is a request from this client, we flush their changes and then forward the request to the server
• Every time the server sends a diagnostics notification, it's forwarded to the clients having this file open

Installation

npm install @codingame/languageserver-mutualized 

Usage

2 functions are exported:

• createLanguageClient which create a LanguageClient from a language server connection
• bindLanguageClient which bind a client connection on an existing LanguageClient and return a promise resolved when the client leaves or the server is shutdown

Examples

There is an example that demonstrate how the languageserver-mutualized can be used. It uses monaco-editor and monaco-languageclient.

The important file is json-server-launcher.ts

To run it:

npm ci
npm start

It will open a page in your browser with 2 editors to illustrate the 2 use-cases:

• The first editor is synchronized between all clients and use the same file uri
• The second editor use a random file uri and is not synchronized

Advanced users

Cache

After a change, most clients send some requests (SemanticTokens, CodeLens...). If a file is used by multiple clients, every client will send the request and the server will need to answer it multiple times.

To prevent this, a cache can be provided when creating the LanguageClient

Configuration

The configuration coming from the client cannot be used. However, some language servers requires some configuration to work.

The configuration can be provided using the getConfiguration LanguageClient option.

It will be called when:

• The language server sends a workspace/configuration request (the scope will be ignored)
• When providing the synchronizeConfigurationSections option (old deprecated way but still used by many language servers)