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

stillwater-sc / universal / 38065842850
89%
master: 84%

Build:
Build:
LAST BUILD BRANCH: main
DEFAULT BRANCH: master
Ran 10 Oct 2026 04:25PM UTC
Jobs 1
Files 694
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

10 Oct 2026 03:59PM UTC coverage: 86.333% (-0.03%) from 86.359%
38065842850

push

github

web-flow
feat(ucalc): areal and poxel tile types, single tiles and tile intervals (#1657)

Phase 1 of #1654: the ubit tile number systems in ucalc.

- areal8..64 and poxel8..64 are single tiles, with the library's
  sticky-flag arithmetic. Each value records whether it encloses its true
  value: literals and arithmetic on exact tiles do, open operands and
  functions computed through double do not.
- areal8i..poxel64i are tile_interval enclosures: sqrt, cos, abs and
  integer powers enclose, and functions without an enclosing
  implementation return the entire line.
- Literals are read exactly from their decimal text. poxel goes through
  a blocktriple with a sticky bit, so 0.1 is one tile even in poxel64.
- New syntax: [a, b] / hull(a, b), and x~ / above(x) for ULP-wide inputs.
- show reports tiles, sign verdict and enclosure, and types groups the
  families.
- Values crossing types: a plain double is exact; another tile type's
  value enters by its bounds, or as the entire line when it does not
  enclose; a rounded value is bracketed.
- The regression reproduces quadratic_roots.cpp (#1649) tile for tile at
  16/32/64 bits, exact and ULP-wide. It also reads every lattice point
  and open tile of areal8/16 and poxel8/16 back from exact decimal text.

areal fix (#1655): conversion from double to areal<64,11>, and from
float to areal<32,8>, dropped the source's last fraction bit without
setting the ubit. The guards become shiftRight >= 0, and VerifyLastSourceBit
covers the case. Both conversion suites now accumulate their failure
counts. Float subnormals are a separate defect, #1656.

Fixes #1655
Refs #1654

BEHAVIOUR-CHANGE: areal<64,11> from double and areal<32,8> from float now set the ubit when the source's last fraction bit is dropped, instead of returning a truncated value marked exact

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

4 of 5 new or added lines in 1 file covered. (80.0%)

14 existing lines in 2 files now uncovered.

47257 of 54738 relevant lines covered (86.33%)

9972394.25 hits per line

Uncovered Changes

Lines Coverage ∆ File
1
91.74
-0.4% include/sw/universal/number/areal/areal_impl.hpp

Coverage Regressions

Lines Coverage ∆ File
10
96.17
-0.65% include/sw/universal/number/cfloat/cfloat_impl.hpp
4
91.74
-0.4% include/sw/universal/number/areal/areal_impl.hpp
Jobs
ID Job ID Ran Files Coverage
1 38065842850.1 10 Oct 2026 04:25PM UTC 694
86.33
GitHub Action Run
Source Files on build 38065842850
  • Tree
  • List 694
  • Changed 2
  • Source Changed 1
  • Coverage Changed 2
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #38065842850
  • b4cc4c39 on github
  • Prev Build on main (#37946807971)
  • Next Build on main (#38068257786)
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