From 551e08290ad1f2308d3be2445fdaf2b27140d155 Mon Sep 17 00:00:00 2001
From: fantomc0der <michael.w.weisz@gmail.com>
Date: Mon, 27 Apr 2026 18:44:02 -0400
Subject: [PATCH] Fix colspan cells producing misaligned markdown table columns

When a table cell has colspan>1, the table-cell Turndown rule now emits
extra empty pipe-delimited columns (colspan-1 of them) so the rendered
row always has the same column count as the separator and data rows.

Previously a colspan="2" header like "mikethecoder" produced "| mikethecoder |"
(1 column) while the separator and data rows had 2 columns, causing most
markdown renderers to reject the table entirely.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---
 src/converters/html-to-markdown.ts        | 5 ++++-
 tests/converters/html-to-markdown.test.ts | 6 +++---
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/src/converters/html-to-markdown.ts b/src/converters/html-to-markdown.ts
index 61e1d9a..7d919b5 100644
--- a/src/converters/html-to-markdown.ts
+++ b/src/converters/html-to-markdown.ts
@@ -194,7 +194,10 @@ function createTurndownService(): TurndownService {
       const text = content.replace(/[\r\n]+/g, " ").replace(/\|/g, "\\|").trim();
       const index = directCellIndex(node);
       const prefix = index === 0 ? "| " : " ";
-      return prefix + text + " |";
+      const rawSpan = parseInt((node as HTMLElement).getAttribute?.("colspan") ?? "1", 10);
+      const span = Number.isFinite(rawSpan) && rawSpan > 1 ? rawSpan : 1;
+      const extra = span > 1 ? " |".repeat(span - 1) : "";
+      return prefix + text + " |" + extra;
     },
   });
 
diff --git a/tests/converters/html-to-markdown.test.ts b/tests/converters/html-to-markdown.test.ts
index e4865f4..48f57ab 100644
--- a/tests/converters/html-to-markdown.test.ts
+++ b/tests/converters/html-to-markdown.test.ts
@@ -288,7 +288,7 @@ describe("htmlToMarkdown", () => {
     expect(result).toContain("| Magic Number | 17 |");
   });
 
-  it("separator column count respects colspan on first-row cells", () => {
+  it("colspan cells emit extra empty columns to keep table width consistent", () => {
     const html = `
       <table>
         <tr><td colspan="2">Header</td></tr>
@@ -296,7 +296,7 @@ describe("htmlToMarkdown", () => {
       </table>
     `;
     const result = htmlToMarkdown(html);
-    expect(result).toContain("| Header |");
+    expect(result).toContain("| Header | |");
     expect(result).toContain("| --- | --- |");
     expect(result).toContain("| Left | Right |");
   });
@@ -351,7 +351,7 @@ describe("htmlToMarkdown", () => {
       </div>
     `;
     const result = htmlToMarkdown(html);
-    expect(result).toContain("| someuser |");
+    expect(result).toContain("| someuser | |");
     expect(result).toContain("| --- | --- |");
     expect(result).toContain("| Magic Number | 17 |");
     expect(result).toContain("| Job | Celebrity Nobody |");