./math/fast_float, Number parsing library

[ Image CVSweb ] [ Image Homepage ] [ Image RSS ] [ Image Required by ]

Branch: CURRENT, Version: 8.3.3, Package name: fast_float-8.3.3, Maintainer: pkgsrc-users

The fast_float library provides fast header-only implementations for the C++
from_chars functions for float and double types as well as integer types. These
functions convert ASCII strings representing decimal values (e.g., 1.3e10) into
binary types. We provide exact rounding (including round to even). In our
experience, these fast_float functions many times faster than comparable
number-parsing functions from existing C++ standard libraries.

Master sites:

Filesize: 153.17 KB

Version history: (Expand)

CVS history: (Expand)


   2026-10-31 18:19:29 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.3.3

8.3.3

extend fuzzer to wide char types, integer parsing and all formats
Leave std::float32_t/std::float64_t unmodified on invalid input
Avoid right-shifting a negative value in detail::power()
Reject integer_times_pow10 arguments wider than 64 bits
Silence -Wc++20-extensions at the remaining [[unlikely]] sites
   2026-10-11 18:06:50 by Thomas Klausner | Files touched by this commit (2) | Package updated
Log message:
fast_float: update to 8.3.2.

avoid signed negation overflow in integer_times_pow10 by @sahvx655-wq in #421
avoid signed overflow in is_integer on negative wchar_t units by @sahvx655-wq in #422
   2026-10-05 06:58:04 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.3.1

8.3.1

Keep chars_format::javascript out of the common hot path
Restore the simple loop in fastfloat_strncasecmp
Fold the format at compile time under clang, mark cold checks unlikely
Fix std::float16_t rounding of subnormal ties and fast-path overflow
Average branch misses over the repetitions in the benchmark
drop the unreachable tail of parse_mantissa
CI: pin Alpine to v3.22 to fix setup-alpine failures
Reorder parsed_number_string_t fields to avoid padding
Add manually triggered release workflow; fix on-release asset upload
   2026-09-22 09:22:39 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.3.0

8.3.0

New chars_format::javascript format (ECMAScript DecimalLiteral): like JSON, but \ 
the integer part may be empty (.5) and the fractional part may be empty (5., \ 
5.e3); leading zeros, inf and nan are rejected.
New chars_format::javascript_sloppy format: additionally accepts sloppy-mode \ 
NonOctalDecimalIntegerLiteral (08.5 is 8.5). A legacy octal literal (0775) is \ 
rejected with the new parse_error::legacy_octal_integer_part so it can be parsed \ 
in base 8.
Faster integer-part scanning (4-digit SWAR on GCC) and skipping the \ 
rounding-mode probe for long mantissas
Avoid old-style casts
Fix compatibility with min()/max() macros
Guard is_space against negative signed code units
docs: fix strtod function name in README
   2026-06-22 12:50:20 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.2.10

8.2.10
Int overflow check with a faster approach
   2026-06-13 07:30:25 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.2.9

8.2.9

Remove an else if statement that is always false
Remove an unreachable return statement
Fix compile error with gcc 9: use of [[unlikely]]
reject non-digit wide code units in uint8/uint16 integer fast path
   2026-06-09 06:33:58 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.2.8

8.2.8
Fix compile error in clang<10: fails on pragma -Wc++20-extensions
   2026-06-08 05:44:18 by Adam Ciarcinski | Files touched by this commit (2) | Package updated
Log message:
fast_float: updated to 8.2.7

8.2.7

Fix spelling
Skip materializing parsed_number_string_t spans on the hot path