Dimension: 6 (hazard surface) · Severity: LOW
src/lib/LibCtPop.sol:6-22 — the nine 256-bit masks are hand-typed hex literals. ctpop (uses M1/M2/M4/H01) and ctpopSlow (uses M1/M2/M4/M8..M128) share M1/M2/M4, so the testCtPopReference cross-check (ctpop == ctpopSlow) alone cannot catch a shared-constant bug — both would be corrupted identically.
Nuance (verified during audit): the n-oracle tests (testCtPopUnshuffled/testCtPopShuffled assert count == n) do functionally enforce the masks — a wrong mask fails count == n — and all nine masks were independently recomputed and are correct today. So this is a convention-vs-enforcement observation, not a live gap: the only enforcement is indirect (via the functional tests), with no explicit constant pin.
Proposed fix (optional, belt-and-suspenders) — add a test that recomputes each mask from its intended repeating pattern and asserts equality, so a mistyped literal fails directly rather than only via a downstream count mismatch.
Dimension: 6 (hazard surface) · Severity: LOW
src/lib/LibCtPop.sol:6-22— the nine 256-bit masks are hand-typed hex literals.ctpop(uses M1/M2/M4/H01) andctpopSlow(uses M1/M2/M4/M8..M128) share M1/M2/M4, so thetestCtPopReferencecross-check (ctpop == ctpopSlow) alone cannot catch a shared-constant bug — both would be corrupted identically.Nuance (verified during audit): the
n-oracle tests (testCtPopUnshuffled/testCtPopShuffledassertcount == n) do functionally enforce the masks — a wrong mask failscount == n— and all nine masks were independently recomputed and are correct today. So this is a convention-vs-enforcement observation, not a live gap: the only enforcement is indirect (via the functional tests), with no explicit constant pin.Proposed fix (optional, belt-and-suspenders) — add a test that recomputes each mask from its intended repeating pattern and asserts equality, so a mistyped literal fails directly rather than only via a downstream count mismatch.