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

stillwater-sc / universal / 24248240237
84%
master: 84%

Build:
Build:
LAST BUILD BRANCH: fix/issue-1027-cfloat-frexp-semantics
DEFAULT BRANCH: master
Ran 10 Apr 2026 03:02PM UTC
Jobs 1
Files 633
Run time 2min
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

10 Apr 2026 02:35PM UTC coverage: 84.371% (+0.02%) from 84.353%
24248240237

push

github

web-flow
feat(bisection): promote storage to blockbinary for nbits > 64 (#706)

* feat(bisection): promote storage to blockbinary for nbits > 64

Replace int64_t storage with a dual-path design:
  using bits_type = conditional_t<(nbits <= 64), int64_t,
      blockbinary<nbits, bt, BinaryNumberType::Signed>>;

For nbits <= 64 the compiled code is identical to before (int64_t).
For nbits > 64, blockbinary handles multi-limb two's complement
arithmetic natively. The nbits <= 64 static_assert is removed.

Key changes to bisection_impl.hpp:
- Add bisection_detail namespace with bit_mask, bit_test, sign_extend,
  nan_encoding, and negate_bits helpers that abstract int64_t vs
  blockbinary differences
- Template bisection_encode/bisection_decode over <R, B> where B is
  the storage type (was hardcoded int64_t)
- Move NaN check in decode BEFORE the XOR sign-flip -- the post-XOR
  check was incorrect for blockbinary because the XOR of the zero
  encoding produces the same bit pattern as the NaN sentinel in
  fixed-width representations
- All class methods (setbits, getbits, setbit, at, maxpos, minpos,
  maxneg, minneg, setnan, isnan, sign, setzero, negation) use
  if constexpr to dispatch between int64_t and blockbinary paths

manipulators.hpp: to_binary() now uses v.at(i) instead of v.getbits()
so it works for nbits > 64.

New test: static/bisection/conversion/wide.cpp validates 80-bit and
128-bit bisection types with dd AuxReal: round-trip, monotonicity, and
basic arithmetic.

Resolves #704

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(bisection): address CodeRabbit review on PR #706

Three fixes to bisection_impl.hpp:

1. bit_mask/bit_test: use uint64_t for intermediate bit manipulation
   in the int64_t path to avoid UB from left-shifting into the sign
   bit (pos==63) and implementation-defined right-shift of signed values.

2. setbit(): route the int64_t path through getbits()/setbits() to
   re-normalize sign extension after modify... (continued)

54 of 58 new or added lines in 2 files covered. (93.1%)

3 existing lines in 2 files now uncovered.

44835 of 53140 relevant lines covered (84.37%)

6522668.86 hits per line

Uncovered Changes

Lines Coverage ∆ File
4
96.95
-2.31% include/sw/universal/number/bisection/bisection_impl.hpp

Coverage Regressions

Lines Coverage ∆ File
2
89.8
-0.57% include/sw/universal/number/posito/posito_impl.hpp
1
96.95
-2.31% include/sw/universal/number/bisection/bisection_impl.hpp
Jobs
ID Job ID Ran Files Coverage
1 24248240237.1 10 Apr 2026 03:02PM UTC 633
84.37
GitHub Action Run
Source Files on build 24248240237
  • Tree
  • List 633
  • Changed 5
  • Source Changed 2
  • Coverage Changed 5
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #24248240237
  • 12691dac on github
  • Prev Build on main (#24223726910)
  • Next Build on main (#24350825956)
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc