Vietnamese text with tone marks not rendering correctly in input box

Bug Report

Description

When typing Vietnamese text using an IME (e.g., Telex or VNI input method), characters with tone marks (combining diacritical marks) cause the input box display to break — text appears cut off or garbled mid-word.

The same text typed in the same terminal with OpenAI Codex renders correctly, suggesting this is an application-level input rendering issue, not a terminal issue.

Steps to Reproduce

1. Use a Vietnamese IME (Telex/VNI) in a terminal
2. Open Claude Code
3. Type a Vietnamese sentence with tone marks, e.g. giờ tôi nói

Expected Behavior

Full Vietnamese text displayed correctly in the input box.

Actual Behavior

Text appears cut off after certain tone-marked characters (e.g., displays giờ tôi n instead of giờ tôi nói).

Environment

• OS: WSL2 (Linux 6.6.87.2-microsoft-standard-WSL2)
• Shell: zsh
• Claude Code version: 2.1.52

Root Cause (suspected)

Claude Code likely miscalculates cursor position for Vietnamese combining characters (NFD Unicode sequences), causing the display to shift incorrectly. Other CLI tools (e.g., OpenAI Codex) in the same terminal handle this correctly.