Language server not starting for custom language type

Hi. Firstly jupyter-lsp is an awesome project.

Description

I’m struggling to get my custom language server to actually start when editing my custom language.

What’s particularly strange is, it starts without an issue if I configure it to work with python. I just can’t get it to start when I’m using my custom kernel.

I wonder if you could help me work around this issue? I’ve looked at all the existing issues and get the impression there might be ambiguity around how languages are discovered (eg. https://github.com/krassowski/jupyterlab-lsp/issues/190, https://github.com/robocorp/robotframework-lsp/issues/13).

Thanks in advance.

Steps to reproduce

$ mkdir test

test/kernel.json:

{
 "argv": ["python", "-m",
          "testkernel", "-f",
          "{connection_file}"],
 "display_name": "test",
 "name": "test",
 "language": "test"
}

$ jupyter kernelspec install --user test

testkernel.py:

from ipykernel.kernelbase import Kernel

class TestKernel(Kernel):
    implementation = 'test'
    implementation_version = '2'
    banner = "Test banner"
    language_version = '2'
    language = 'test'
    language_info = {
                     'mimetype': 'text/x-test',
                     'name': 'test'
    }

    def do_execute(self, code, silent, store_history=True, user_expressions=None, allow_stdin=False):
        if not silent:
            stream_content = {'name': 'stdout', 'text': code}
            self.send_response(self.iopub_socket, 'stream', stream_content)

        return {'status': 'ok',
                'execution_count': self.execution_count,
                'payload': [],
                'user_expressions': {},
               }

if __name__ == '__main__':
    from ipykernel.kernelapp import IPKernelApp
    IPKernelApp.launch_instance(kernel_class=TestKernel)

~/.jupyter/jupyter_notebook_config.py (put any language server in the argv here if desired)

c.LanguageServerManager.language_servers = {
    "test-language-server": {
        "version": 2,
        "argv": ["touch", "language-server-started"],
        "languages": ["test"],
        "mime_types": [
            "text/x-test",
        ]
    }
}

Findings

When a new notebook is made with the Test kernel, the kernel works as expected. However the langauge server doesn’t start up.

Logs

jupyter lab --log-level=DEBUG

[D 09:32:58.078 LabApp] [lsp] The following Language Servers will be available: {
      "test-language-server": {
        "argv": [
          "touch",
          "language-server-started"
        ],
        "languages": [
          "test"
        ],
        "mime_types": [
          "text/x-test"
        ],
        "version": 2
      }
    }

[D 10:17:15.959 LabApp] Received kernel info: {'status': 'ok', 'protocol_version': '5.3', 'implementation': 'test', 'implementation_version': '2', 'language_info': {'mimetype': 'text/x-test', 'name': 'test'}, 'banner': 'Test banner', 'help_links': []}

Enironment

Running in latest scipy-notebook docker. Using conda for package management. jupyter-lsp 0.9.2
jupyterlab 2.2.8

[mike-vogel]

Just after typing the above I realized I been just looking at the config files. I added the file_extension into the language_info returned by the Kernel and things connected.

[mikesligo]

The language server is implemented in scala, using LSP4J as a base (similiar to scala meta https://github.com/krassowski/jupyterlab-lsp/issues/17). At the moment I only have hover implemented but it looks like it works well with jupyter-lsp.