perf: optimize newline preservation using memchr
Replace byte-by-byte loops with efficient chunked filling:
- Use memchr::memchr2 to find newline positions
- Fill spaces in chunks between newlines using optimized fill()
- Preserves correctness while restoring performance
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>