fix(docx-core): keep paragraph sectPr inside diffmatch paragraphs

coverage/emitted-schema-known-failures.json

@@ -1,10 +1,4 @@
 [
-  {
-    "id": "ilpa-orphaned-ppr-close",
-    "issue": "#450",
-    "match": "not well-formed: Opening and ending tag mismatch: \"w:body\" != \"w:pPr\"",
-    "reason": "ILPA-pair comparison drops the opening <w:p><w:pPr> of a paragraph-mark sectPr paragraph, leaving orphaned close tags at body level"
-  },
   {
     "id": "duplicate-rprchange-in-rpr",
     "issue": "#451",

packages/docx-core/src/baselines/diffmatch/xmlParser.test.ts

@@ -122,8 +122,36 @@ describe('diffmatch xml parser (xmldom)', () => {
     await then('the body parts and sectPr are correctly separated', () => {
       expect(parts.beforeBody).toContain('<w:body>');
       expect(parts.afterBody).toContain('</w:body>');
-      expect(extracted.content).toContain('<w:p>');
-      expect(extracted.sectPr).toContain('<w:sectPr>');
+      expect(extracted.content).toContain('<w:p');
+      expect(extracted.content).toContain('<w:t>A</w:t>');
+      expect(extracted.sectPr).toContain('<w:sectPr');
+      expect(extracted.sectPr).toContain('<w:pgSz');
+    });
+  });
+
+  test('does not extract a paragraph-mark sectPr as a body-level sectPr', async ({ given, when, then }: AllureBddContext) => {
+    let parts: ReturnType<typeof getBodyContent>;
+    let extracted: ReturnType<typeof extractSectPr>;
+
+    await given('a body ending with a paragraph whose pPr contains sectPr and run content', () => {
+      const xml = `<?xml version="1.0" encoding="UTF-8"?>
+<w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
+  <w:body>
+    <w:p><w:r><w:t>A</w:t></w:r></w:p>
+    <w:p><w:pPr><w:sectPr><w:pgSz w:w="12240" w:h="15840"/></w:sectPr></w:pPr><w:r><w:br w:type="page"/></w:r></w:p>
+  </w:body>
+</w:document>`;
+      parts = getBodyContent(xml);
+    });
+
+    await when('sectPr is extracted for document reconstruction', () => {
+      extracted = extractSectPr(parts.bodyContent);
+    });
+
+    await then('the paragraph remains intact and no body-level sectPr is returned', () => {
+      expect(extracted.sectPr).toBeNull();
+      expect(extracted.content).toContain('<w:pPr><w:sectPr>');
+      expect(extracted.content).toContain('</w:pPr><w:r>');
     });
   });
 });

packages/docx-core/src/baselines/diffmatch/xmlParser.ts

@@ -210,17 +210,21 @@ export function extractSectPr(bodyContent: string): {
   content: string;
   sectPr: string | null;
 } {
-  // Find the last w:sectPr in the body
-  const sectPrMatch = bodyContent.match(/<w:sectPr[^>]*>[\s\S]*?<\/w:sectPr>\s*$/);
+  const trimmed = bodyContent.trimEnd();
+  if (!trimmed.endsWith('</w:sectPr>')) return { content: bodyContent, sectPr: null };
 
-  if (!sectPrMatch) {
-    return { content: bodyContent, sectPr: null };
-  }
+  const sectPrStart = trimmed.lastIndexOf('<w:sectPr');
+  if (sectPrStart < 0) return { content: bodyContent, sectPr: null };
+
+  const beforeSectPr = trimmed.slice(0, sectPrStart);
+  const lastParagraphOpen = beforeSectPr.lastIndexOf('<w:p');
+  const lastParagraphClose = beforeSectPr.lastIndexOf('</w:p>');
+  if (lastParagraphOpen > lastParagraphClose) return { content: bodyContent, sectPr: null };
 
-  const sectPrStart = bodyContent.lastIndexOf(sectPrMatch[0]);
+  const trailingWhitespace = bodyContent.slice(trimmed.length);
   return {
-    content: bodyContent.slice(0, sectPrStart),
-    sectPr: sectPrMatch[0],
+    content: bodyContent.slice(0, sectPrStart) + trailingWhitespace,
+    sectPr: trimmed.slice(sectPrStart),
   };
 }