What's Changed
- Add FASTFLOAT_ROUNDS_TO_NEAREST to skip the rounding-mode check by @lemire in #432
- Benchmark fast_float with and without FASTFLOAT_ROUNDS_TO_NEAREST by @lemire in #433
- Avoid zero-filling bigint buffers in the slow path (C++26 [[indeterminate]]) by @lemire in #440
Full Changelog: v8.3.3...v8.4.0