• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

stillwater-sc / universal / 39713268894
89%
master: 84%

Build:
Build:
LAST BUILD BRANCH: main
DEFAULT BRANCH: master
Ran 22 Oct 2026 03:06AM UTC
Jobs 1
Files 761
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

22 Oct 2026 02:47AM UTC coverage: 87.95% (+0.02%) from 87.931%
39713268894

push

github

web-flow
fix(efloat): std::numeric_limits, range-bounded arithmetic, isone() (#1787)

efloat had no std::numeric_limits specialization: numeric_limits<efloat<N>> was
the primary template, with is_specialized false and every limit zero (infinity()
was 0). The new efloat/numeric_limits.hpp, included from core.hpp, reports
digits = nlimbs * 32, epsilon() = 2^(1 - digits), min()/max() at the ends of the
exponent range, the real special values, and denorm_absent.

The arithmetic now keeps to that range, efloat::min_scale/max_scale =
+-(2^62 - 1). normalize() and store_rounded() check it and call enforce_range()
(out of line) on failure: overflow to inf, or the largest value at the result's
precision when the mode rounds toward zero; underflow to 0, or min() when the
mode rounds away from zero and, to nearest, above min() / 2; Overflow /
Underflow / Inexact raised. ldexp normalizes; nextafter steps from max() to inf,
from min() to 0 and from 0 to +-min() in every mode; fma scales its terms when
the exact product would leave the range, and returns z for a finite x y with an
infinite z (fma(max, max, -inf) was NaN). isone() compared with 0x8000'000 and
was always false.

Verified: mpfr_check 1 all 8192 0 wrong; elastic/efloat gcc and clang; test
outputs unchanged against main except api.cpp's infinity line; removing either
range check fails 128 / 224 cases; all 35 mpfr_gap rows within 4% of main.

Resolves #1786

BEHAVIOUR-CHANGE: std::numeric_limits<efloat<N>> is specialized (infinity() was 0); results outside [2^(1-2^62), 2^(2^62)) overflow to inf (or max) and underflow to 0 (or min) per rounding mode; nextafter(0, y) is +-2^(1-2^62), was 2^-1074; efloat::isone() is true for 1

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

90 of 90 new or added lines in 4 files covered. (100.0%)

53338 of 60646 relevant lines covered (87.95%)

9164301.17 hits per line

Jobs
ID Job ID Ran Files Coverage
1 39713268894.1 22 Oct 2026 03:06AM UTC 761
87.95
GitHub Action Run
Source Files on build 39713268894
  • Tree
  • List 761
  • Changed 3
  • Source Changed 3
  • Coverage Changed 3
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #39713268894
  • 4def52f7 on github
  • Prev Build on main (#39695261548)
  • Next Build on main (#39720248990)
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc