Fix #7196

Summary by CodeRabbit

• Improvements

  • Wide rollout of UTF‑8/WTF‑8 handling across the VM and standard library for safer, binary‑preserving text processing and improved non‑ASCII rendering.
  • Many object reprs and I/O paths now correctly display surrogate and non‑ASCII content.

• New Features

  • UTF‑8 string types and VM helpers exposed to APIs for explicit UTF‑8 handling.

• Bug Fixes

  • Stricter string extraction yields earlier, clearer errors and reduces lossy conversions.