stillwater-sc / universal / 26375706711
84%
master: 84%

LAST BUILD BRANCH: fix/issue-1002-ereal-highprec-constants
DEFAULT BRANCH: master
Ran
Jobs 1
Files 658
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
coverage: 84.008% (+0.03%) from 83.98%
26375706711

push

github

web-flow 
feat(ereal): structural stress test + Priest-normal fuzz retrofit (#954 D3) (#985)

* feat(ereal): structural stress test + Priest-normal fuzz retrofit (#954 D3)

Final deliverable of the #954 first-principles structural oracle: assert
that every result of +, -, *, / on an ereal expansion is in Priest's
normal form (decreasing magnitude, non-overlapping limbs, no interior
zeros), with no external reference.

New elastic/ereal/arithmetic/structural_stress.cpp stresses the boundaries
that the random fuzz inputs (sampled near 1e6) never reach:
  - magnitude sweep:    leading magnitudes 2^-900 .. 2^900 in 2^50 steps
  - subnormal boundary: trailing limbs near DBL_MIN / DBL_TRUE_MIN
  - round-to-even ties: pairs summing onto representable midpoints
  - near extremes:      operands near maxpos and minpos, where products
                        span the widest exponent range (the #981 surface)

Retrofits an is_priest_normal assertion into the four per-operator fuzz
loops (addition/subtraction/multiplication/division), so the structural
invariant is checked on every fuzzed result -- a permanent regression
guard for #981 (unrenormalized expansion_product).

All generators pass at L1+L2 on both gcc-13 and clang-18, 0 violations.

Resolves #954

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix(ereal): honor reps in VerifyRoundToEvenTies (CodeRabbit PR #985)

Wrap the tie generator in the same outer reps loop as the other three
structural-stress generators so regression levels scale it uniformly.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

47362 of 56378 relevant lines covered (84.01%)

8605776.15 hits per line

Coverage Regressions

Lines Coverage ∆ File
2
89.78
 -0.55% include/sw/universal/number/posito/posito_impl.hpp
Jobs
ID Job ID Ran Files Coverage
1 26375706711.1 658
84.01
 GitHub Action Run
Source Files on build 26375706711