fast_float-8.3.1

[Bruce Dubbs]

New point version.

[Joe Locash]

Version 8.3.1

What's Changed

• Keep chars_format::javascript out of the common hot path by @Algunenano in #410
• Restore the simple loop in fastfloat_strncasecmp by @lemire in #412
• Fold the format at compile time under clang, mark cold checks unlikely by @lemire in #413
• Fix std::float16_t rounding of subnormal ties and fast-path overflow by @jadidbourbaki in #408
• Average branch misses over the repetitions in the benchmark by @Algunenano in #415 drop the unreachable tail of parse_mantissa by @caolanm in #409
• CI: pin Alpine to v3.22 to fix setup-alpine failures by @lemire in #417
• Reorder parsed_number_string_t fields to avoid padding (fixes #418) by @lemire in #419
• Add manually triggered release workflow; fix on-release asset upload by @lemire in #420

Fixed at 461501e6a5.