fix(git-sync): subpages round-trips (was {{SUBPAGES}} literal) + exhaustive all-node round-trip test

subpages exported to the literal `{{SUBPAGES}}`, which has no markdown/HTML
inverse, so on re-import it came back as a plain paragraph holding the visible
text "{{SUBPAGES}}" — the embed rendered as that literal string on the page
after a sync (round-trip data loss, seen live). It now emits the schema-matching
`<div data-type="subpages">` like every other embed node, so the schema's
parseHTML rebuilds the subpages node. Also dropped the leaf-atom content-hole
in the subpages renderHTML.

New committed regression coverage:
- packages/git-sync/test/roundtrip-all-nodes.test.ts — exhaustive serialize ->
  deserialize round trip for ALL 40 node/mark types; each asserts the node/mark
  survives and no `{{...}}` literal leaks. This is the test that caught subpages.
- §13.1 gate (git-sync-converter-gate.spec.ts): subpages added to the green
  corpus (round-trips through the REAL server schema).
- Corrected two PR-authored tests that asserted the old {{SUBPAGES}} loss as
  "by design" — they now assert the fixed round trip.

Also folds in review #1679 coverage-gap tests (no prod change): orchestrator
pollTick/enabledSpaces, datasource 3-way merge dispatch, page.repo
last_updated_source provenance SQL.

git-sync vitest 659 (+1 expected-fail), server tsc clean, server specs green.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

packages/git-sync/test/markdown-converter-gaps.test.ts

@@ -68,28 +68,27 @@ describe('pageBreak data loss (no converter case — SPEC §11 divergence)', ()
 });
 
 // ---------------------------------------------------------------------------
-// 2. subpages LOSSY round-trip (`case "subpages"` emits `{{SUBPAGES}}`).
+// 2. subpages round-trip (`case "subpages"` emits the schema-matching div).
 //
-// The golden test only pins the EMISSION string. The token has no markdown or
-// HTML meaning, so on re-import marked treats `{{SUBPAGES}}` as ordinary text:
-// the subpages BLOCK comes back as a plain PARAGRAPH carrying that literal
-// string, NOT a `subpages` node. The export is "lossy but legible" by design;
-// this test pins the actual lossy round-trip behavior.
+// It used to emit the literal `{{SUBPAGES}}`, which has no markdown/HTML meaning,
+// so on re-import the subpages BLOCK came back as a plain PARAGRAPH carrying the
+// literal string (the embed rendered as visible "{{SUBPAGES}}" text on the page
+// after a sync — data loss). It now emits `<div data-type="subpages">` like the
+// other embed nodes, so the schema's parseHTML rebuilds the subpages node.
 // ---------------------------------------------------------------------------
-describe('subpages lossy round-trip ({{SUBPAGES}} placeholder)', () => {
-  it('emits {{SUBPAGES}} which re-imports as a paragraph, not a subpages node', async () => {
+describe('subpages round-trip (schema-matching div)', () => {
+  it('emits the subpages div and re-imports as a subpages node (no literal leak)', async () => {
     const { md1, doc2 } = await roundTrip({ type: 'subpages' });
-    expect(md1).toBe('{{SUBPAGES}}');
+    expect(md1).toBe('<div data-type="subpages"></div>');
 
-    // The re-imported doc has a single paragraph holding the literal token.
-    const top = doc2.content || [];
-    expect(top).toHaveLength(1);
-    expect(top[0].type).toBe('paragraph');
-    expect(top[0].content?.[0]).toMatchObject({ type: 'text', text: '{{SUBPAGES}}' });
-
-    // The subpages node itself is gone: nothing in the doc is a subpages node.
-    const allTypes = top.map((n: any) => n.type);
-    expect(allTypes).not.toContain('subpages');
+    const collect = (n: any): string[] => [
+      n.type,
+      ...((n.content || []) as any[]).flatMap(collect),
+    ];
+    const allTypes = (doc2.content || []).flatMap(collect);
+    // The subpages node survives, and no literal {{SUBPAGES}} text leaked back.
+    expect(allTypes).toContain('subpages');
+    expect(JSON.stringify(doc2)).not.toContain('{{SUBPAGES}}');
   });
 });
 

packages/git-sync/test/markdown-converter-golden.test.ts

@@ -142,8 +142,8 @@ describe('paragraph.textAlign -> <div align>', () => {
 });
 
 describe('subpages token + unknown-in-container fallback', () => {
-  it('subpages emits the {{SUBPAGES}} placeholder token', () => {
-    expect(c({ type: 'subpages' })).toBe('{{SUBPAGES}}');
+  it('subpages emits the schema-matching div (round-trips, unlike the old {{SUBPAGES}} literal)', () => {
+    expect(c({ type: 'subpages' })).toBe('<div data-type="subpages"></div>');
   });
 
   it('an unknown block inside a raw-HTML container is wrapped in <div> (never markdown)', () => {

packages/git-sync/test/roundtrip-all-nodes.test.ts

@@ -0,0 +1,109 @@
+import { describe, expect, it } from 'vitest';
+import { convertProseMirrorToMarkdown } from '../src/lib/markdown-converter.js';
+import { markdownToProseMirror } from '../src/lib/markdown-to-prosemirror.js';
+
+/**
+ * Exhaustive serialize -> deserialize round trip for EVERY node and mark type the
+ * Docmost document schema supports. The git-sync converter exports a page body to
+ * Markdown and imports it back; any node type that has no parseHTML inverse (or is
+ * serialized to a literal that never re-parses) silently degrades to plain text on
+ * a round trip — e.g. `subpages` used to export as the literal `{{SUBPAGES}}` and
+ * came back as the visible text "{{SUBPAGES}}" instead of the embed.
+ *
+ * This guards the whole class: for one representative fixture per type, the node
+ * (or mark) MUST still be present after convert -> import, and the exported
+ * Markdown must not contain a `{{...}}` template literal (the old lossy form).
+ */
+
+const T = (t: string, marks?: any[]) =>
+  marks ? { type: 'text', text: t, marks } : { type: 'text', text: t };
+const P = (...c: any[]) => ({ type: 'paragraph', content: c });
+const doc = (...c: any[]) => ({ type: 'doc', content: c });
+
+// `primary` is the node/mark type that must survive the round trip.
+const FIXTURES: Record<string, { doc: any; primary: string }> = {
+  paragraph: { doc: doc(P(T('hello'))), primary: 'paragraph' },
+  heading: { doc: doc({ type: 'heading', attrs: { level: 2 }, content: [T('H2')] }), primary: 'heading' },
+  blockquote: { doc: doc({ type: 'blockquote', content: [P(T('q'))] }), primary: 'blockquote' },
+  codeBlock: { doc: doc({ type: 'codeBlock', attrs: { language: 'js' }, content: [T('foo()')] }), primary: 'codeBlock' },
+  bulletList: { doc: doc({ type: 'bulletList', content: [{ type: 'listItem', content: [P(T('a'))] }] }), primary: 'bulletList' },
+  orderedList: { doc: doc({ type: 'orderedList', attrs: { start: 1 }, content: [{ type: 'listItem', content: [P(T('a'))] }] }), primary: 'orderedList' },
+  taskList: { doc: doc({ type: 'taskList', content: [{ type: 'taskItem', attrs: { checked: true }, content: [P(T('done'))] }] }), primary: 'taskList' },
+  horizontalRule: { doc: doc({ type: 'horizontalRule' }), primary: 'horizontalRule' },
+  image: { doc: doc({ type: 'image', attrs: { src: '/f/x.png', width: '320', align: 'center' } }), primary: 'image' },
+  hardBreak: { doc: doc(P(T('a'), { type: 'hardBreak' }, T('b'))), primary: 'hardBreak' },
+  callout: { doc: doc({ type: 'callout', attrs: { type: 'info' }, content: [P(T('note'))] }), primary: 'callout' },
+  columns: {
+    doc: doc({ type: 'columns', content: [
+      { type: 'column', attrs: { width: '50%' }, content: [P(T('L'))] },
+      { type: 'column', attrs: { width: '50%' }, content: [P(T('R'))] }] }),
+    primary: 'column',
+  },
+  details: {
+    doc: doc({ type: 'details', content: [
+      { type: 'detailsSummary', content: [T('Sum')] },
+      { type: 'detailsContent', content: [P(T('body'))] }] }),
+    primary: 'details',
+  },
+  table: {
+    doc: doc({ type: 'table', content: [
+      { type: 'tableRow', content: [{ type: 'tableHeader', content: [P(T('H1'))] }, { type: 'tableHeader', content: [P(T('H2'))] }] },
+      { type: 'tableRow', content: [{ type: 'tableCell', content: [P(T('C1'))] }, { type: 'tableCell', content: [P(T('C2'))] }] }] }),
+    primary: 'tableCell',
+  },
+  mathBlock: { doc: doc({ type: 'mathBlock', attrs: { math: 'x^2' } }), primary: 'mathBlock' },
+  mathInline: { doc: doc(P({ type: 'mathInline', attrs: { math: 'x^2' } })), primary: 'mathInline' },
+  mention: { doc: doc(P({ type: 'mention', attrs: { id: 'u1', label: 'Bob', entityType: 'user', entityId: 'u1' } })), primary: 'mention' },
+  drawio: { doc: doc({ type: 'drawio', attrs: { src: '/f/d.drawio', attachmentId: 'a1' } }), primary: 'drawio' },
+  excalidraw: { doc: doc({ type: 'excalidraw', attrs: { src: '/f/e.excalidraw', attachmentId: 'a1' } }), primary: 'excalidraw' },
+  embed: { doc: doc({ type: 'embed', attrs: { src: 'https://youtube.com/x', provider: 'iframe' } }), primary: 'embed' },
+  pdf: { doc: doc({ type: 'pdf', attrs: { src: '/f/x.pdf', attachmentId: 'a1' } }), primary: 'pdf' },
+  video: { doc: doc({ type: 'video', attrs: { src: '/f/v.mp4', width: '640' } }), primary: 'video' },
+  audio: { doc: doc({ type: 'audio', attrs: { src: '/f/a.mp3' } }), primary: 'audio' },
+  attachment: { doc: doc({ type: 'attachment', attrs: { url: '/f/x.zip', name: 'x.zip', attachmentId: 'a1' } }), primary: 'attachment' },
+  youtube: { doc: doc({ type: 'youtube', attrs: { src: 'https://youtube.com/watch?v=x' } }), primary: 'youtube' },
+  subpages: { doc: doc({ type: 'subpages' }), primary: 'subpages' },
+  pageBreak: { doc: doc({ type: 'pageBreak' }), primary: 'pageBreak' },
+  htmlEmbed: { doc: doc({ type: 'htmlEmbed', attrs: { source: '<b>hi</b>' } }), primary: 'htmlEmbed' },
+  pageEmbed: { doc: doc({ type: 'pageEmbed', attrs: { pageId: 'p1' } }), primary: 'pageEmbed' },
+  transclusion: { doc: doc({ type: 'transclusionSource', attrs: { pageId: 'p1' } }), primary: 'transclusionSource' },
+  footnote: {
+    doc: doc(
+      P(T('x'), { type: 'footnoteReference', attrs: { id: 'fn1' } }),
+      { type: 'footnotesList', content: [{ type: 'footnoteDefinition', attrs: { id: 'fn1' }, content: [P(T('note'))] }] }),
+    primary: 'footnoteReference',
+  },
+  status: { doc: doc(P({ type: 'status', attrs: { text: 'Done', color: 'green' } })), primary: 'status' },
+  // marks
+  bold: { doc: doc(P(T('b', [{ type: 'bold' }]))), primary: 'bold' },
+  italic: { doc: doc(P(T('i', [{ type: 'italic' }]))), primary: 'italic' },
+  strike: { doc: doc(P(T('s', [{ type: 'strike' }]))), primary: 'strike' },
+  code: { doc: doc(P(T('c', [{ type: 'code' }]))), primary: 'code' },
+  underline: { doc: doc(P(T('u', [{ type: 'underline' }]))), primary: 'underline' },
+  superscript: { doc: doc(P(T('x', [{ type: 'superscript' }]))), primary: 'superscript' },
+  subscript: { doc: doc(P(T('x', [{ type: 'subscript' }]))), primary: 'subscript' },
+  highlight: { doc: doc(P(T('h', [{ type: 'highlight', attrs: { color: 'yellow' } }]))), primary: 'highlight' },
+  link: { doc: doc(P(T('l', [{ type: 'link', attrs: { href: 'https://x.com' } }]))), primary: 'link' },
+};
+
+function collectTypes(n: any, set = new Set<string>()): Set<string> {
+  if (!n || typeof n !== 'object') return set;
+  if (n.type) set.add(n.type);
+  if (Array.isArray(n.content)) n.content.forEach((c: any) => collectTypes(c, set));
+  if (Array.isArray(n.marks)) n.marks.forEach((m: any) => m?.type && set.add(m.type));
+  return set;
+}
+
+describe('git-sync converter: every node/mark type survives a Markdown round trip', () => {
+  for (const [name, { doc: original, primary }] of Object.entries(FIXTURES)) {
+    it(`round-trips ${name} (keeps the ${primary} node/mark, no literal leak)`, async () => {
+      const md = convertProseMirrorToMarkdown(original);
+      // The lossy old form serialized embeds to `{{...}}` literals that never
+      // re-parsed; no node may export to one.
+      expect(md).not.toMatch(/\{\{.*\}\}/);
+      const back = await markdownToProseMirror(md);
+      const types = collectTypes(back);
+      expect(types.has(primary)).toBe(true);
+    });
+  }
+});