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

stillwater-sc / universal / 42425064522
89%
master: 84%

Build:
Build:
LAST BUILD BRANCH: fix/fixpnt-quire-resolve-saturates
DEFAULT BRANCH: master
Ran 07 Nov 2026 11:41PM UTC
Jobs 1
Files 868
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

07 Nov 2026 11:23PM UTC coverage: 88.773% (+0.003%) from 88.77%
42425064522

push

github

web-flow
fix(edecimal): one form of zero, NaN/inf conversion to 0, and single-step shifts (#1992)

* fix(edecimal): one form of zero, NaN/inf conversion to 0, and single-step shifts

Negation and the remainder of an exact division produced a negative zero, and
shifting zero left padded it with zero digits. Both compared unequal to
edecimal(0). Unary minus, setdigit and the remainder now leave zero positive,
and << leaves zero alone.

A NaN or infinite float or double converted to a huge garbage value (the bits
of inf read as 2^128 or 2^1024). They now convert to 0, as they already did on
the long double path and for einteger and erational.

<< and >> insert and erase all their digits in one call, and the long division
collects quotient digits and stores them once instead of inserting each at the
front.

parse("2.0") stays rejected: the strict rule from #865 (no fractional digits,
even zeros) is kept and now documented.

Tests: arithmetic/reference.cpp checks + - * / % negate << >> against int64 with
a no-negative-zero, no-padding check; conversion/assignment.cpp (an empty stub)
now checks float/double/long double conversion including NaN and inf. Both fail
against the previous header.

Resolves #1989

BEHAVIOUR-CHANGE: edecimal -x % y with an exact division, -edecimal(0) and edecimal(0) << k now give +0, which compares equal to edecimal(0)
BEHAVIOUR-CHANGE: edecimal from a NaN or infinite float/double is now 0 instead of a large garbage value

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

* test(edecimal): assert parse("2.0") stays rejected

Records the decision on #1989: the strict no-fractional-digits rule from #865 is kept.

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

* fix(edecimal): move the quotient digits instead of assigning from reverse iterators

gcc 13 on aarch64 and MinGW reports a -Wstringop-overflow false positive on the
vector assign() from reverse iterators in decint_divide, which -Werror in the
noexceptions integration ... (continued)

13 of 13 new or added lines in 1 file covered. (100.0%)

4 existing lines in 2 files now uncovered.

59598 of 67135 relevant lines covered (88.77%)

9694318.7 hits per line

Coverage Regressions

Lines Coverage ∆ File
3
97.67
-0.19% include/sw/universal/number/cfloat/cfloat_impl.hpp
1
93.16
-0.85% include/sw/universal/verification/correctly_rounded_reference.hpp
Jobs
ID Job ID Ran Files Coverage
1 42425064522.1 07 Nov 2026 11:41PM UTC 868
88.77
GitHub Action Run
Source Files on build 42425064522
  • Tree
  • List 868
  • Changed 3
  • Source Changed 1
  • Coverage Changed 3
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #42425064522
  • c2c2a12b on github
  • Prev Build on main (#42406034972)
  • Next Build on main (#42429189490)
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