Releases: fastfloat/fast_float
Releases · fastfloat/fast_float
Release list
Version 8.3.0
What's Changed
- New
chars_format::javascriptformat (ECMAScript DecimalLiteral): like JSON, but the integer part may be empty (.5) and the fractional part may be empty (5.,5.e3); leading zeros,infandnanare rejected. #407 - New
chars_format::javascript_sloppyformat: additionally accepts sloppy-modeNonOctalDecimalIntegerLiteral(08.5is 8.5). A legacy octal literal (0775) is rejected with the newparse_error::legacy_octal_integer_partso 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 by @fcostaoliveira in #398
- Avoid old-style casts by @lemire in #405
- Fix compatibility with
min()/max()macros by @dg0yt in #399 - Guard
is_spaceagainst negative signed code units by @sahvx655-wq in #395 - docs: fix strtod function name in README by @latent-9 in #404
Full Changelog: v8.2.10...v8.3.0
Version 8.2.10
What's Changed
Full Changelog: v8.2.9...v8.2.10
Version 8.2.9
What's Changed
- Remove an else if statement that is always false by @correctmost in #390
- Remove an unreachable return statement by @correctmost in #389
- Fix compile error with gcc 9: use of [[unlikely]] by @biojppm in #392
- reject non-digit wide code units in uint8/uint16 integer fast path by @sahvx655-wq in #391
New Contributors
- @correctmost made their first contribution in #390
- @sahvx655-wq made their first contribution in #391
Full Changelog: v8.2.8...v8.2.9
Version 8.2.8
What's Changed
Full Changelog: v8.2.7...v8.2.8
Version 8.2.7
What's Changed
- Fix spelling by @jwakely in #385
- Skip materializing parsed_number_string_t spans on the hot path by @fcostaoliveira in #387
Full Changelog: v8.2.6...v8.2.7
Version 8.2.6
What's Changed
- Unroll the integer-part digit scan (straight-line for the common 1-5 digit case) by @fcostaoliveira in #381
- Add a 4-digit SWAR follow-up to loop_parse_if_eight_digits (clang) by @fcostaoliveira in #382
- Parallelize the exhaustive float32 sweeps across hardware threads (~75-88x) by @fcostaoliveira in #383
New Contributors
- @fcostaoliveira made their first contribution in #381
Full Changelog: v8.2.5...v8.2.6
Version 8.2.5
What's Changed
- Fix Bazel build with bzlmod by @BYVoid in #377
- Replace std::min with ternary operators to avoid dependency by @mlippautz in #379
New Contributors
- @mlippautz made their first contribution in #379
Full Changelog: v8.2.4...v8.2.5
Version 8.2.4
What's Changed
This release can boost the performance by about 10% in some cases thanks to @kolemannix
- Add more boundary test cases for double and float parsing by @976520 in #366
- fix warning C4702: unreachable code by @x1sc0 in #368
- Mention C under other languages by @kolemannix in #372
- A few inlines by @kolemannix in #374
New Contributors
- @976520 made their first contribution in #366
- @x1sc0 made their first contribution in #368
- @kolemannix made their first contribution in #372
Full Changelog: v8.2.3...v8.2.4
Version 8.2.3
What's Changed
- add base check for uint8 by @shikharish in #358
- fix issue with s390x. by @lemire in #360
- optimize uint16 parsing by @shikharish in #359
- optimize fastfloat_strncasecmp by @sleepingeight in #356
- Bump the github-actions group across 1 directory with 4 updates by @dependabot[bot] in #344
- fix early return error in fastfloat_strncasecmp by @sleepingeight in #362
- Clarify behavior for underflow by @pitrou in #363
- Fix overzealous build-time
/permissive-propagating to compile-time (cross-compilation bug) by @nyoma-diamond in #365
New Contributors
- @sleepingeight made their first contribution in #356
- @nyoma-diamond made their first contribution in #365
Full Changelog: v8.2.2...v8.2.3
Version 8.2.2
What's Changed
- fix endianess bug in uint8 parsing by @shikharish in #357
Full Changelog: v8.2.1...v8.2.2