fix: use CDATA sections in XML examples to prevent parser errors (#4852)

[hannesrudolph]

Related GitHub Issue

Fixes #4852

Description

This PR fixes the root cause of the apply_diff tool errors on complex XML by properly escaping content in the XML examples using CDATA sections.

The Real Problem

The XML examples in multi-file-search-replace.ts contained unescaped markdown backticks (```) within the XML content blocks. This caused fast-xml-parser to fail with "Cannot read properties of undefined (reading 'addChild')" errors.

The Solution

Use CDATA sections to properly escape the content that contains special characters. This is the standard XML way to include content that might otherwise be interpreted as markup.

Changes Made

• Updated all XML examples in src/core/diff/strategies/multi-file-search-replace.ts to use CDATA sections
• No fallback parser or workarounds needed - this is the proper fix

Before

<content>
\`\`\`
<<<<<<< SEARCH
...
\`\`\`
</content>

After

<content><![CDATA[
<<<<<<< SEARCH
...
]]></content>

This is a much cleaner solution than the fallback parser approach in PR #6774.

---

Important

Fixes XML parsing errors in multi-file-search-replace.ts by using CDATA sections to escape special characters in XML examples.

• Behavior:
  • Fixes XML parsing errors in multi-file-search-replace.ts by using CDATA sections to escape special characters in XML examples.
  • Eliminates the need for fallback parsers or workarounds.
• Changes Made:
  • Updated all XML examples in multi-file-search-replace.ts to use CDATA sections for content containing special characters.

^{This description was created by for 9240dfb. You can customize this summary. It will automatically update as commits are pushed.}

[Copilot]

Pull Request Overview

This PR fixes XML parser errors in the apply_diff tool by properly escaping content in XML examples using CDATA sections. The issue occurred because unescaped markdown backticks (```) within XML content blocks caused the fast-xml-parser to fail with "Cannot read properties of undefined" errors.

• Replaced escaped backticks with CDATA sections in all XML examples
• Simplified XML structure by removing unnecessary escaping
• Applied standard XML practices for including content with special characters

[Copilot]

[nitpick] The CDATA section should be properly indented to match the XML structure. Consider adding appropriate indentation before the opening CDATA tag to maintain consistent formatting with the rest of the XML.

Suggested change

	<content><![CDATA[
	<content>
	<![CDATA[

[Copilot]

[nitpick] The closing CDATA section should be properly indented to match the XML structure. Consider adding appropriate indentation before the closing tags to maintain consistent formatting with the rest of the XML.

Suggested change

	]]></content>
	]]></content>

[roomote]

Thank you for your contribution! I've reviewed the changes and the fix correctly addresses the root cause identified in issue #4852. Using CDATA sections is the proper XML standard for escaping special characters, and this is a much cleaner solution than the fallback parser approach. I have some suggestions for improvement below.

[roomote]

Consider adding test coverage for this CDATA implementation. Could we add a test case in multi-search-replace.spec.ts that specifically verifies XML content with special characters (like <<<<<<<, =======, >>>>>>>) wrapped in CDATA sections parses correctly?

[roomote]

Is the inline CDATA formatting intentional? For better readability, you might consider:

Suggested change

	]]></content>
	<content>
	<![CDATA[
	<<<<<<< SEARCH
	def calculate_total(items):
	total = 0
	for item in items:
	total += item
	return total
	=======
	def calculate_total(items):
	"""Calculate total with 10% markup"""
	return sum(item * 1.1 for item in items)
	>>>>>>> REPLACE
	]]>
	</content>

This would make the CDATA boundaries more visually distinct.

[roomote]

Consider adding a comment here explaining why CDATA sections are used, something like:

Suggested change

	<diff>
	<diff>
	<!-- Using CDATA to properly escape special characters that would otherwise cause XML parsing errors (see issue #4852) -->
	<content><![CDATA[

This would help future developers understand the rationale.

[daniel-lxs]

LGTM