fix: bounds check Substring calls in MarkdownConverter

[rorychatt]

Fixes ArgumentOutOfRangeException in MarkdownConverter that was causing documentation extraction failures for files like IVYCHILD001.md. Also includes a new test suite to verify the fix.

[rorychatt]

@claude review this

[ivy-interactive-claude-code]

Claude finished @rorychatt's task in 1m 16s —— View job

---

Code Review

• Gather context and read changed files
• Review code changes in MarkdownConverter.cs
• Review new tests in MarkdownConverterTests.cs
• Provide feedback

---

Overall: Looks good! ✓ The fix is solid and the approach is consistent.

Summary

The PR adds defensive bounds checking to 4 Substring calls that could throw ArgumentOutOfRangeException when Markdig span positions don't align perfectly with the markdown content length.

What's Good

1. Consistent fix pattern - All 4 locations use the same robust pattern:

   int start = Math.Max(0, mBlock.Span.Start);
   int length = Math.Min(mBlock.Span.Length, markdownContent.Length - start);

2. Tests cover edge cases - The emoji test is particularly good since surrogate pairs can cause offset misalignment between character and byte positions.

3. No over-engineering - The fix is minimal and focused.

Suggestions (Optional)

1. One more Substring to protect - Line 85 in ConvertAsync has an unprotected call:

   string yamlContent = markdownContent.Substring(yamlBlock.Span.Start, yamlBlock.Span.Length);

   Consider applying the same bounds check for consistency. Fix this →

2. Test cleanup - Minor: the using System.Text; and using Markdig; imports in the test file appear unused. Not a blocker.

Verdict

Approve - The core fix is correct and addresses the reported issue. The optional suggestion about line 85 can be addressed in a follow-up if desired.

---

|