claude-code - 💡(How to fix) Fix Edit tool corrupts multi-byte UTF-8 characters at replacement string boundary [2 comments, 2 participants]

• Silently corrupts file content — no error is raised

TL;DR

The issue can likely be resolved by ensuring proper handling of UTF-8 encoding in the Edit tool, specifically when processing multi-byte characters in the new_string parameter.

Guidance

• Verify that the Edit tool is configured to support UTF-8 encoding and that the new_string parameter is being processed as a UTF-8 encoded string.
• Check the Edit tool's documentation for any specific requirements or limitations related to handling multi-byte characters.
• Test the Edit tool with different input strings containing multi-byte characters to determine if the issue is specific to certain characters or sequences.
• Consider adding error handling or logging to detect and report any encoding-related issues during the editing process.

Example

No specific code snippet can be provided without more information about the Edit tool's implementation, but ensuring that strings are handled as UTF-8 encoded can be crucial:

# Hypothetical example of ensuring UTF-8 encoding in Python
new_string = "**進行中**"
new_string_encoded = new_string.encode('utf-8')

Notes

The provided information suggests a potential issue with the Edit tool's handling of UTF-8 encoding, but without more details about the tool's implementation, it's difficult to provide a definitive solution. The issue may be specific to the macOS environment or the claude-opus-4-6 model.

Recommendation

Apply a workaround by manually verifying and correcting the output of the Edit tool, especially when working with CJK characters, until a more permanent solution can be implemented to ensure proper UTF-8 encoding handling.