Jupyter language server considers cells separately - regression in v1.37.2

[mrgrain]

Consider the following simple notebook:

This notebook works as expected. It defines the variable hello in cell 1 and has no problem printing it from cell 2.
The issue is that the language server incorrectly evaluates the context of each cell separately and fails to find the hello variable in cell 2 (red squiggly line, see screenshot below). This used to work correctly in v1.37.1.

There is a variation of this issue involving globalThis. Also used to work in v1.37.1.

With stuff.ts being:

globalThis.hello2 = 'also world';

Full notebook source:

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "const hello = \"world\";"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "world\n"
     ]
    }
   ],
   "source": [
    "console.log(hello);"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Deno",
   "language": "typescript",
   "name": "deno"
  },
  "language_info": {
   "file_extension": ".ts",
   "mimetype": "text/x.typescript",
   "name": "typescript",
   "nb_converter": "script",
   "pygments_lexer": "typescript",
   "version": "5.2.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

[nayeemrmn]

Tracked in denoland/vscode_deno#932.

I believed it worked in previous versions because the Deno LSP wasn't being used at all, and it wouldn't have worked for cells containing imports. Correct me if I'm wrong.

[mrgrain]

Ah apologies for the duplicate. 🙏🏻

I believed it worked in previous versions because the Deno LSP wasn't being used at all, and it wouldn't have worked for cells containing imports. Correct me if I'm wrong.

This might well be the case, but I don't know for sure 😅