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

stillwater-sc / universal / 42662405004
89%
master: 84%

Build:
Build:
LAST BUILD BRANCH: fix/fixpnt-quire-resolve-saturates
DEFAULT BRANCH: master
Ran 09 Nov 2026 05:41PM UTC
Jobs 1
Files 878
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

09 Nov 2026 05:12PM UTC coverage: 88.99% (+0.02%) from 88.975%
42662405004

push

github

web-flow
fix(dd): keep the sign of zero and atan2's range in dd and qd; fix stale complex CMake gates (#2034)

#2032: dd's and qd's error-free transformations turned every zero result
into +0 (+0 * -1, -0 + -0, -0 / 2), sin/tan/log1p/expm1/sqrt returned +0 at
-0, and a -0 converted to +0 through double/float/long double (hi + lo).
The operators take a zero result from the leading limbs, the functions
return their zero argument, and the conversions return the leading limb of
a zero. atan2 normalized by sqrt(x^2 + y^2), which overflowed or underflowed
(atan2(2e100, -1e200) was -8e15, atan2(1e160, 3e160) 3.32 in dd and 7.39 in
qd, atan2(3e-200, 1e-200) -inf/NaN); it now scales both arguments by the
same power of 2, limb by limb, which is exact.

#2033: applications/performance/complex and the cfloat complex suite were
gated on BUILD_COMPLEX, which nothing sets (the option is
UNIVERSAL_BUILD_COMPLEX), so neither was ever built. compute.cpp is ported
to sw::universal::complex<T>; cfloat's complex addition.cpp, an empty
skeleton, now checks complex addition against component-wise cfloat
addition and against float.

Tests: static/highprecision/{dd,qd}/arithmetic/signed_zero.cpp and
math/{signed_zero,atan2_range}.cpp, all failing on main; the long double
conversion checks are guarded by LONG_DOUBLE_SUPPORT.

Resolves #2032
Resolves #2033

BEHAVIOUR-CHANGE: dd and qd arithmetic, sin/tan/log1p/expm1/sqrt and conversions to native floats now give zero results the IEEE 754 sign (e.g. +0 * -1 is -0, sin(-0) is -0, double(dd(-0.0)) is -0); dd and qd atan2 now return correct values for arguments whose squares leave the double range

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

50 of 51 new or added lines in 10 files covered. (98.04%)

4 existing lines in 3 files now uncovered.

61129 of 68692 relevant lines covered (88.99%)

9929743.86 hits per line

Uncovered Changes

Lines Coverage ∆ File
1
99.02
-0.98% include/sw/universal/number/dd/math/functions/trigonometry.hpp

Coverage Regressions

Lines Coverage ∆ File
2
85.93
0.7% include/sw/universal/number/qd/qd_impl.hpp
1
97.8
-0.06% include/sw/universal/number/cfloat/cfloat_impl.hpp
1
99.02
-0.98% include/sw/universal/number/dd/math/functions/trigonometry.hpp
Jobs
ID Job ID Ran Files Coverage
1 42662405004.1 09 Nov 2026 05:41PM UTC 878
88.99
GitHub Action Run
Source Files on build 42662405004
  • Tree
  • List 878
  • Changed 13
  • Source Changed 11
  • Coverage Changed 6
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #42662405004
  • 47fc1b8e on github
  • Prev Build on main (#42637949454)
  • Next Build on main (#42691428857)
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