fix(typst): add newlines to @skip so multi-line code args parse cleanly

'\n' inside CodeArgs was an invalid token, triggering Lezer error recovery
and resetting parser state before codeIdentTokenizer could fire.  Heading
detection is unaffected — headingTokenizer uses raw input.peek(-1) char
reads which see the '\n' byte regardless of what @skip consumes.

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

services/web/frontend/js/features/source-editor/lezer-typst/typst.grammar

@@ -27,7 +27,6 @@ item {
   Label |
   Ref |
   Escape |
-  Newline |
   MarkupContent
 }
 
@@ -184,10 +183,12 @@ Escape { "\\" EscapeChar }
 
 // ── Regular tokens ────────────────────────────────────────────────────────
 @tokens {
-  // Horizontal whitespace only.  Newlines are kept as explicit Newline items
+  // All whitespace including newlines.  Heading detection still works because
-  // so that HeadingMark (which checks start-of-line via input.peek(-1)) can
+  // headingTokenizer uses input.peek(-1) on the raw character stream — it sees
-  // reliably detect newlines in the raw input stream.
+  // the '\n' byte regardless of what @skip consumes at the token level.
-  spaces { $[ \t]+ }
+  // Including '\n' here lets multi-line code expressions (e.g. #figure(\n  ...\n))
+  // parse without error instead of triggering Lezer error recovery.
+  spaces { $[ \t\n\r]+ }
 
   // Boolean / null literals — distinct from keywords for highlighting.
   CodeBool { "true" | "false" | "none" | "auto" }
@@ -213,9 +214,6 @@ Escape { "\\" EscapeChar }
   // Escape: any single character after backslash.
   EscapeChar { _ }
 
-  // Newline item — kept out of @skip so heading detection works.
-  Newline { "\n" }
-
   // Resolve ambiguities in merged states:
   // EscapeChar > spaces: after '\', EscapeChar must win over the skip token.
   // "(" > "." > "]": callSuffix delimiters must win over MarkupContent after