scaffold: VSCode/VSCodium extension wrapping testium lsp

A thin LSP client that spawns `testium lsp` and forwards messages to
VSCode. All language intelligence lives in the testium repo, so this
extension only needs republishing for editor-side UX changes.

  - package.json: declares the .tum language, registers the LSP client,
    exposes testium.serverPath and testium.trace.server settings.
  - src/extension.ts: vscode-languageclient setup over stdio, surfaces a
    user-visible error if `testium lsp` can't start.
  - syntaxes/tum.tmLanguage.json: TextMate grammar embedding source.yaml
    plus tum-specific tokens for $(name), <| python |>, {% jinja %},
    {{ jinja }}.
  - language-configuration.json: comment char, bracket pairs, auto-close
    for $(/), <|/|>, {%/%}, {{/}}.
  - README.md: install (incl. testium[lsp] extra), settings, packaging
    via vsce, publish to both VS Marketplace and Open VSX (the latter
    is what VSCodium / Cursor / Gitpod use — same .vsix artifact).

Works identically on VSCode and VSCodium — no Microsoft-proprietary
API surface used.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

syntaxes/tum.tmLanguage.json

@@ -0,0 +1,81 @@
+{
+  "$schema": "https://raw.githubusercontent.com/martinring/tmlanguage/master/tmlanguage.json",
+  "name": "Testium TUM",
+  "scopeName": "source.tum",
+  "patterns": [
+    { "include": "#tum-tokens" },
+    { "include": "source.yaml" }
+  ],
+  "repository": {
+    "tum-tokens": {
+      "patterns": [
+        { "include": "#globaldict-reference" },
+        { "include": "#eval-expression" },
+        { "include": "#jinja-statement" },
+        { "include": "#jinja-expression" }
+      ]
+    },
+    "globaldict-reference": {
+      "name": "variable.other.tum.globaldict",
+      "begin": "\\$\\(",
+      "end": "\\)",
+      "beginCaptures": {
+        "0": { "name": "punctuation.definition.variable.begin.tum" }
+      },
+      "endCaptures": {
+        "0": { "name": "punctuation.definition.variable.end.tum" }
+      },
+      "patterns": [
+        { "name": "entity.name.variable.tum", "match": "[A-Za-z_][A-Za-z0-9_]*" }
+      ]
+    },
+    "eval-expression": {
+      "name": "meta.embedded.expression.tum",
+      "begin": "<\\|",
+      "end": "\\|>",
+      "beginCaptures": {
+        "0": { "name": "punctuation.definition.expression.begin.tum" }
+      },
+      "endCaptures": {
+        "0": { "name": "punctuation.definition.expression.end.tum" }
+      },
+      "patterns": [
+        { "include": "source.python" }
+      ]
+    },
+    "jinja-statement": {
+      "name": "meta.embedded.block.jinja.tum",
+      "begin": "\\{%-?",
+      "end": "-?%\\}",
+      "beginCaptures": {
+        "0": { "name": "punctuation.definition.template-statement.begin.tum" }
+      },
+      "endCaptures": {
+        "0": { "name": "punctuation.definition.template-statement.end.tum" }
+      },
+      "patterns": [
+        { "name": "keyword.control.jinja.tum",
+          "match": "\\b(for|endfor|if|elif|else|endif|set|include|extends|block|endblock|macro|endmacro|in|and|or|not|is|with|without|context)\\b" },
+        { "name": "string.quoted.double.jinja.tum", "match": "\"(?:[^\"\\\\]|\\\\.)*\"" },
+        { "name": "string.quoted.single.jinja.tum", "match": "'(?:[^'\\\\]|\\\\.)*'" },
+        { "name": "constant.numeric.jinja.tum", "match": "\\b[0-9]+(?:\\.[0-9]+)?\\b" }
+      ]
+    },
+    "jinja-expression": {
+      "name": "meta.embedded.expression.jinja.tum",
+      "begin": "\\{\\{-?",
+      "end": "-?\\}\\}",
+      "beginCaptures": {
+        "0": { "name": "punctuation.definition.template-expression.begin.tum" }
+      },
+      "endCaptures": {
+        "0": { "name": "punctuation.definition.template-expression.end.tum" }
+      },
+      "patterns": [
+        { "name": "string.quoted.double.jinja.tum", "match": "\"(?:[^\"\\\\]|\\\\.)*\"" },
+        { "name": "string.quoted.single.jinja.tum", "match": "'(?:[^'\\\\]|\\\\.)*'" },
+        { "name": "variable.other.jinja.tum", "match": "[A-Za-z_][A-Za-z0-9_]*" }
+      ]
+    }
+  }
+}