fix(stdlib): use UTF-8 boundaries for string assignment 🧵

[timfennis]

Summary

• translate character indexes to UTF-8 byte boundaries before mutating strings
• fix both single-character and slice assignment for multibyte text
• cover positive indexes, negative indexes, and Unicode slices with functional tests

Testing

• cargo fmt --all -- --check
• cargo test -p functional_tests --no-fail-fast
• cargo test --workspace

The workspace test run passes with one pre-existing ndc_lsp test warning about an unnecessary qualification.

AI disclosure

This change was implemented with OpenAI Codex assistance.