Self's review found that commit 4e6495dd's fix for F1 replaced the per-lane
forward-arithmetic defect with a second, still-independent membership
computation: Lane's own half-open admission (Lo <= s < Hi, from a cell's own
Extent) versus Lanes' ceil((cell.Lo - centreLateral) / laneSpacing) lane
index. The two agree only when a cell's own (Hi - Lo) equals the externally
supplied laneSpacing exactly in float -- the identical two-derivations trap
Cut 12.2's own fix batch already found and removed from Lane itself,
reintroduced one layer up. Confirmed reachable: on commit 4e6495dd, solid5
b=(0,1) s=0.49999997 n=1 leaves the ceil-based centre lane EMPTY while Lane(s)
itself finds all 5 cells of the row -- breaking R3 ("the centre lane always
meets metal", Commit's own guard) and the 12.2 placement contract.
FireControl.Lanes is now built entirely out of FireControl.Lane: it calls
Lane(hull, b, centreLateral + k * laneSpacing, ...) once per lane index, so
Lane(s) is literally this function's n = 0 case rather than a parallel
re-implementation of its rule. Commit's placement and Apply's damage split
now read one function. The narrower question left -- can two individually
correct Lane() calls, at two different lines, still walk the same cell? --
is resolved by ownership (the lane closest to the centre keeps a contested
cell; the centre lane, k = 0, never loses one), not by a competing formula,
so the centre lane's final cell set is always exactly Lane(centreLateral)'s
own output.
Lane() itself is unchanged (still called directly by Commit and by
FireControlCut12Tests' edge-seed and interval-edge tests, all still green:
HitsLandOnTheFacingEdge, LaneSeamIsHalfOpen, LaneNeverEmptiesAtAnIntervalEdge,
Cut12-edge-68442709, Cut12-edge-64078070).
Adds LanesQueryIsDisjointMatchesLaneAtCentreAndNeverEmptyOnHull: now that
Lanes is a public, pure geometric query, the boundary sweep is a behavioural
test of its own contract, no reflection -- over every cell's own shadow edge
and float neighbours, at exact-axis, 45-degree, |bx| > |by| bearings and
Soul's own two reproduction cases, spread 1-2: (a) no cell is walked by two
lanes, (b) the centre lane equals a plain Lane(s) call, (c) the centre lane
is never empty when the centre lateral is inside the hull's own shadow. Hand-
confirmed to fail on 4e6495dd's ceil-based Lanes (property (b), the same
input above) and to pass on this commit (fc123fix2-mut3.log).
Adds CommittedCellIsTheFirstCellTheCentreLaneDamages: an end-to-end check
through Fire -> Commit -> Apply (damageSpread 0, a single lane) that the
committed outcome.Cell is the first cell Apply damages.
Full suite: 286 tests, identical across 3 runs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
F3 (medium, confirmed): NoCellIsStruckByTwoLanesAtAnAngledBearing used
penetration 0 (impact cells only) and one fixed-seed shot per bearing, at
only two bearings, so it survived R1e (one-cell spacing when |bx| > |by|),
R1h (spacing 1/max(|bx|,|by|)), R1c (spacing * 0.9), and R1b (spacing *
0.99). Its own comment also falsely claimed a plain revert "fails this at
both bearings" without ever checking a real revert.
Reworked as NoCellIsStruckByTwoLanesInOneShotAtAnyBearing: seven bearings
(exact axis, both 45-degree diagonals, |bx| > |by| and |bx| < |by| on both
sides, all within the fixture's own FiringArc), 300 hits sampled per bearing
at a wide targeting spread (so the committed Lateral's draw visits many
points along the hull's shadow), and ample penetration.
Hand-verified against fc123fix2-mut.log: reverting to one-cell spacing (R0)
and R1c/R1e/R1h are all caught by the full suite -- but not by this test.
F1's structural fix (this branch's earlier commit, FireControl.Lanes) assigns
each cell to exactly one lane index by a single computation, so a wrong
laneSpacing can misassign or drop a cell but can no longer make two lanes
claim the same one; R0/R1c/R1e/R1h are instead caught by the pre-existing
damage-shape tests (DirectHitAtAnAngledBearingMatchesAnIndependentModel,
FootprintWidensAtAnAngledBearingComparedToAxisAligned, and others). R1b (a
1% narrowing) survives the whole 97-test FireControl suite under any of
those tests, including this one, even at 5000 hits per bearing -- recorded
here as an open survivor for Soul's own mutation pass, not claimed dead.
This test's own job is now a standing regression guard on the disjointness
invariant through the real Fire/Commit/Apply pipeline, not a killer of
spacing-magnitude mutants (that job now belongs to the damage-shape tests).
F4: deletes MirrorImageShotsProduceStatisticallyMirroredHullDamage. Soul
showed it cannot fail on spacing -- it survives the full revert and every
spacing mutant, because no per-shot mirror pair is reachable through Fire
(LateralDraw's distribution reverses under mirroring, so a shot and its
"mirror" are independently drawn samples, not a matched pair). It was meant
to pin lane processing order under a shared-cell race; that question is
folded into the F2 disjointness rule (a shared cell can no longer arise) and
the pending F2 ruling on multi-cell item lane order, which remains on hold
per this fix batch's brief.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Soul's second pass (CONFIRMED): even after the F2 shadow-width fix, Apply's
per-lane loop still called Lane() once per lane index, each call re-deriving
its own lateral s(li) = Lateral + (li - n) * laneSpacing by forward float
addition, then testing that value against a cell's Extent interval
independently of every other lane. Two lane lines are only exactly
laneSpacing apart in real arithmetic; accumulating the offset by repeated
addition can shift a side lane by an ulp or two, letting it land inside a
neighbouring cell's half-open interval that its own unrounded value would
never reach -- the same "two derivations of one quantity" defect Cut 12.2's
own fix batch already removed from Lane itself, recurring one layer up in
Apply.
FireControl.Apply now calls a new Lanes() that decides each candidate cell's
lane index with one computation: k = ceil((cell.Lo - centreLateral) /
laneSpacing) is the unique integer for which centreLateral + k * laneSpacing
falls in the cell's own [Lo, Hi) -- derived once, from values no other lane's
arithmetic has touched. A single Ceiling call returns exactly one integer, so
a cell can be assigned to at most one lane by construction, not by float
luck. When k == 0 this is the same condition Lane() tests for centreLateral
itself, so the centre lane's cell set is unchanged (the 12.2 placement
contract, HitsLandOnTheFacingEdge, still holds). laneSpacing is still used,
unchanged, to place each admitted lane's own line for AlongBearing's slab
solve -- geometry, not membership.
Lane() itself is untouched (still used directly by Commit and by
FireControlCut12Tests' edge-seed tests).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
MirrorImageShotsProduceStatisticallyMirroredDamage compared per-cell armour
damage SUMS between a bearing and its mirror, using heavy (1000) uniform
armour. That armour never depleted, so no cell's absorption ever competed
with another lane's, and the test could not see the one thing the operator's
ruling is about: two lanes racing for one cell's finite armour under the old
one-cell spacing. Renamed to MirrorImageShotsProduceStatisticallyMirroredHullDamage,
it now resets a small, finite armour value before every individual shot (so
the race is live on every trial) and compares aggregate HULL damage, where a
lost race's leftover actually lands.
Finding, reported rather than hidden: even with this fix, the test does not
fail under the reverted one-cell spacing. Reasoning checked against measured
behaviour: FireControl.Apply always processes lanes in a fixed index order
(li ascending, i.e. lateral increasing), never based on the bearing's sign.
Over many INDEPENDENTLY random lateral draws, a bearing and its mirror sample
statistically equivalent distributions of "which side's lane goes first" --
the fixed-index tie-break rule respects the mirror symmetry in aggregate, so
independently-drawn trials wash out the order-dependent bias that Soul's own
LaneOrderDemo probe found (that probe compares ascending vs descending
processing of the SAME committed shot, not independently-drawn mirrored
bearings). A production-level test that reliably fails under the old spacing
for this specific property would need to force an exact, deterministic
lateral for two geometrically mirrored shots, which is not reachable through
Fire/Commit without reflection (Commit's lateral draw is randomised, and
Apply is private). NoCellIsStruckByTwoLanesAtAnAngledBearing and
FootprintWidensAtAnAngledBearingComparedToAxisAligned remain the two tests
that actually discriminate the ruling behaviourally (confirmed failing under
a hand-reverted one-cell spacing). The mirror test stays in the suite as a
true, currently-passing invariant and useful regression coverage, not as a
defense of this specific ruling.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
F1: no existing test through Fire commits at a non-axis bearing.
DirectHitAtAnAngledBearingMatchesAnIndependentModel checks every cell's
armour, every item's durability and the hull's own total against an
independent double-precision ray/box model (ported from the Soul probe that
verified this cut's arithmetic to 14,984/15,000 shots), at 45 degrees, a
shallow angle and a bearing with |bx| > |by|.
DiagonalDirectHitReachesEveryCornerCellInOrderThroughFireCommitApply pins the
spec's own diagonal case (Probe1's corner-crossing lane) through the full
Fire -> Commit -> Apply path rather than a direct Lane call: the lane reaches
(0,0), then (1,1), then (2,2), in that order (a bearing this close to 45
degrees also admits the diagonal's row/column neighbours, so the walked lane
is usually five cells, not three -- the spec only requires the three corner
cells present and in order, the same standard the existing direct-Lane test
already pins).
F2 (operator ruling, 2026-09-25: spread lanes should diffuse sideways rather
than double up): three behavioural tests, deliberately not asserting the
spacing value itself (operator: "test behavior, not shape").
NoCellIsStruckByTwoLanesAtAnAngledBearing fires at 45 degrees and at a bearing
with |bx| > |by|, spread 2, heavy uniform armour, and checks no cell fires
ArmorDamage twice. MirrorImageShotsProduceStatisticallyMirroredDamage compares
300 hits at a bearing against 300 at its mirror image, and asserts the
per-cell damage totals are mirrored within tolerance.
FootprintWidensAtAnAngledBearingComparedToAxisAligned compares the lateral
span of hit cells for the same spread at 45 degrees against axis-aligned, and
asserts the angled footprint is wider.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The previous AlongBearingReturnsTheExactSlabIntersection wrote out SlabAxis's
own threshold (1e-9f) and clamp a second time in its ExactSlab helper, so a
mutant that only moved that threshold (C7b, 1e-9 to .2) agreed with both
copies identically and survived.
AlongBearingReturnsTheExactSlabIntersectionAtSeveralBearings replaces it: the
expected entry/exit now come from an exact ray-box intersection that never
special-cases a near-zero bearing component -- IEEE division by a signed zero
already yields the correctly signed infinity, so an axis with no motion is
unconstrained without a hand-written epsilon to disagree with SlabAxis's own.
Checked at the original fixture's bearing plus two within about 11 degrees of
an axis, inside C7b's mutated .2f threshold but nowhere near the real 1e-9f
one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
F4: AnAPRoundStopsAtTheFirstGapAndLeavesTheFarSideOfTheNotchUntouched, a notched
U-shaped hull, pins Lane's own contiguity break (an AP round with ample
penetration must stop at the first gap, not reach the far side).
F5: EverySideLanesRemainderReachesTheHullNotOnlyTheCentreLanes, an
axis-aligned spread-1 fixture (lanes share no cells, so order cannot matter),
pins that every lane's own leftover reaches the hull, not only the centre
lane's.
F6: AbsorbSpendsArmourBeforeTheItemOnTheSameCell, direct Absorb calls (this
file's existing style for the function's own arithmetic) with armour and an
item on the SAME cell for the first time -- no prior fixture combined them.
F7: ADirectHitFromStarboardDamagesStarboardCellsWhenFacingIsMostlyLateral, a
direct-hit test at a facing with |fx| > |fy|, computing "starboard" as the
independent right-perpendicular of the facing rather than reading it from
Entity.ToSchematic. Cut 11's Splash tests are the only thing guarding this
handedness today, and 12.4 deletes Splash.
Both F4 and F5 track hull damage per-attempt (reset on each retry), not as a
before/after total across every retry: this fixture's off-target lanes carry
no armour, so a stray hit on the wrong column during a retry spends real
damage on the hull and would otherwise contaminate a cross-attempt total.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Operator ruling (2026-09-25): overlapping spread lanes should diffuse sideways
rather than double up. FireControl.Apply now spaces lanes by the shadow width
a cell casts at the committed bearing (Extent's own h, doubled) instead of by
one cell -- 1 at axis-aligned bearings (unchanged), up to sqrt(2) at 45
degrees. Reuses Extent, the one function that already owns h for Silhouette
and Lane, rather than re-deriving it a second way.
F3: PenetrationExcludesACellExactlyAtItsOwnBoundary's 1e6-durability marker
absorbed the whole shot before the walk could reach the boundary cell, so the
test could not fail under the `>=`-to-`>` mutant. The marker now carries 10
durability, leaving a remainder that would visibly damage the boundary cell if
the walk incorrectly included it. Adds
PenetrationExcludesCellsBeyondANonIntegerDepth, pinning the ordinary
"< penetration" comparison at a depth (1.5) that isn't an exact-integer
boundary.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
First Stryker pass over the changed lines (Entity.cs: Absorb/DamageHull;
FireControl.cs: Apply, Splash) killed 584/844 candidate mutants; these
six survivors on the cut's own diff get real behavioural tests instead:
- Entity.Absorb's `d > 0f` and `d > 0.1f` guards, and DamageHull's
`damage > .1f` guard, each had their boundary flipped to `>=` and
survived -- at the exact boundary (0 incoming, or exactly the .1f
carry-forward threshold), real code does nothing while the mutant
fires an extra event or spends an extra fraction. Not a rare float
coincidence: a spent lane routinely reaches a cell with exactly 0
remaining, and an authored penetration/armour value can land exactly
on .1f.
- FireControl.Apply's `target.IncomingHit.OnNext(shot.Source)` (moved
here from the deleted Entity.ApplyHit) had no test that it still fires.
- FireControl.Apply's two `walked[li] > 0` checks (metalLanes counting,
and the absorb-loop entry) both flipped to `>= 0` and survived, since a
lane count can never be negative -- a real bug class (a lane entirely
off the hull would dilute the split and leak a damage share to the
hull) that a hull no wider than the spread now catches.
- The penetration cutoff's `>= shot.Penetration` flipped to `>` survived
at the exact boundary -- an axis-aligned lane's entry values are exact
integers, so an authored integer penetration routinely lands exactly on
a cell's own entry difference.
Recorded, not chased: `shot.Outcome.Lateral - (li - n)` (the lane-offset
sign) is an equivalent mutant -- 2n+1 lanes built symmetrically around n
cover the same SET of offsets either way, just enumerated in reverse, and
the loop's own per-lane work is order-independent. The missing
`ArrayPool<LaneCell>.Return` survivor is a perf-hygiene gap (GC pressure,
not a spec'd invariant -- only the forecast path is zero-allocation), not
yet reached by a dedicated allocation test.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DamageIsAbsorbedInOrderAlongTheRay, ArmourFacesTheShot,
BuriedCockpitNeedsPenetration, SpreadWidensTheImpactAcrossLanes,
ARayThroughACornerCrossesTheCornerCell and LaneRemainderReachesTheHull,
per docs/fire-control-cut.md's Cut 12.3 verification list. Builds its own
fixtures (the file's own Build), matching every other cut's own test file.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Items carried into 12.3's brief from 12.2's third Soul pass, all on code
12.3's own lane walk builds on:
- LaneSeamIsHalfOpen pins the s >= Hi admission test (FireControl.cs) on
the smallest fixture where the seam matters -- the s > Hi mutant that
survived 12.2 would have interleaved two columns at a shared seam once
12.3's walk started reading more than buffer[0].
- AlongBearingReturnsTheExactSlabIntersection gives AlongBearing/SlabAxis
their first test of actual VALUES (every mutant there survived in 12.2,
equivalent while only buffer[0].Cell was read) -- a diagonal fixture
checks Lane's own Entry/Exit against an independent slab solve, cell for
cell, in order.
- Corrected the allocation comments on Forecast and
LaneAndSilhouetteAllocateNothingGivenAPooledBuffer: the ~120 B/call on a
gated-in HitProbability is three GetBehavior<TargetingSystem> calls (not
EquippedItem.Behaviors, which is a plain array), plus CellsOf's ~944 B
when an item is aimed at; noted that the loose <500_000 bound doesn't
catch a reintroduced comparer wrapper but the buffered zero-allocation
asserts do.
- ZeroProbabilityShotCommitsAMissWithoutBuildingASilhouette is now
behavioural: PFire 0 comes from closing VisibleEntities (the real gate),
and "no Silhouette was built" is checked by asserting zero allocation
rather than reading CommitProbability's own struct default.
- Dropped the unused System.Reflection import.
Also replaces ExpectedMarchCells, which reproduced Entity.ApplyHit's own
deleted 0.5-step march, with ExpectedLaneCells, built on the
already-verified FireControl.Lane plus an independent penetration cutoff --
used by TurningAfterCommitChangesNothing, whose fixture also gets real
per-cell armour (the flat 1000 armour every Build() authors would let the
new sequential Absorb model's first cell alone swallow the whole 5-damage
shot, leaving nothing to show on the rest of a several-cell lane).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
HardpointHitDamagesItemThenHull and PenetrationMarchIsPlanar both called
Entity.ApplyHit directly, which Cut 12.3 deletes. The rules they pinned:
"damage above armour plus item durability zeroes the item and passes the
remainder to the hull" is now FireControlCut123Tests.DamageIsAbsorbedInOrderAlongTheRay,
through the real Fire/Commit/Apply path. "The penetration march is planar,
in the target's own schematic frame" is subsumed by 12.2's own
HitsLandOnTheFacingEdge and TurningArmourIntoTheShotTakesItOnTheArmour
(which already exercise FireControl.Apply's committed, schematic-frame
bearing end to end) and by FireControlCut123Tests.ArmourFacesTheShot,
which pins planarity across several non-axis-aligned facings.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Deletes Entity.DamageSchematic (the even split over the whole hit shape)
and Entity.ApplyHit (the 0.5-step penetration march), replacing both with
Entity.Absorb (armour then item, per cell) and Entity.DamageHull, moved
verbatim from DamageSchematic's own body. FireControl.Apply now builds
2n+1 parallel lanes from the committed Bearing/Lateral (Q12-2 = A),
shares damage evenly over the lanes that meet metal, walks each lane in
order via the already-verified FireControl.Lane, absorbing cell by cell
up to the penetration depth, and sends every lane's own remainder to the
hull where it ends (Q12-3 = A). Splash is rewritten onto Absorb/DamageHull
(arithmetically identical to before) since DamageSchematic is gone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Measured (Release, 7x9 solid hull -- the LonginusX shipped-hull fixture the
brief named is not available to this test file):
- Lane: 0 B / 1000 calls (was: 64 B/call from Array.Sort's comparer wrapper;
now confirmed 0 after the earlier IComparable<T> change in this batch).
- Silhouette (caller-supplied buffer): 0 B / 1000 calls.
- HitProbability, gated in (a real, non-gated-out call): ~120 B/call. This is
NOT from Lane, Silhouette or the comparer -- it comes from Accuracy/
Resolution's own GetBehavior<T>, which walks Entity.Equipment and each
item's Behaviors (both ReactiveCollection<T>); that type's enumerator is
heap-allocated per foreach. This predates Cut 12.2 and is out of this cut's
scope to fix.
- Inspect, gated in: ~280 B/call (same cause, more GetBehavior<T> calls).
The Forecast comment claiming "the forecast path (HitProbability, Inspect) no
longer allocates on a gated-in call" was not true and is corrected to say what
was actually measured: Silhouette itself is allocation-free given a pooled
buffer, and the residual cost belongs to Entity's own collection choice.
LaneAndSilhouetteAllocateNothingGivenAPooledBuffer asserts the true zero-alloc
claims exactly and records (with a loose regression bound, not a false zero)
the unrelated Entity-level cost on the full HitProbability path.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CommitProbability's `if (p <= 0f) return 0f;` guards against building a
Silhouette for a shot that cannot hit. Stryker's `p < 0` boundary mutant
(excluding only p == 0 exactly) survived because nothing exercised the guard
at that exact boundary. ZeroProbabilityShotCommitsAMissWithoutBuildingA
Silhouette sets a real, fired shot's PFire to 0, calls CommitProbability
directly and asserts sil stays the zeroed default (Intervals null, Count 0 --
no Silhouette was built), then resolves the shot through Zone.Update and
checks it lands as a clean miss.
Manually verified against the exact `p < 0` patch: it now fails (sil.Intervals
comes back non-null, a real Silhouette array).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Only a 2x2 (BigMarker) aimed item was ever tested -- a square's centroid and
its bounding-box centre are the same point, so N11 (aim point = bounding-box
centre instead of the true cell-average centroid) survived. AimingAtAnLShaped
ItemCentresOnItsCentroidNotItsBoundingBox equips a 3-cell L-tromino (Build's
new equipLMarker fixture), fires from a diagonal facing so the shape's true
asymmetry projects fully onto the lateral axis, and asserts the empirical mean
lateral offset tracks the item's own true centroid (computed by the test) and
not its bounding-box centre.
Manually verified against N11's own patch: it now fails (empirical -2.139,
tracking the wrong bounding-box prediction -2.121, against the true centroid
-1.886); the existing BigMarker test is unaffected since a square's two
candidate points coincide.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TheHudEstimateIsTheCommitPrice used to fire each shot, then reach into
Zone.PendingShots and overwrite the frozen Tracking field after the fact. Build
now takes a `tracking` parameter (default 100000, matching the old fixed
value) alongside Precision and Accuracy, so this test authors Tracking = 6 up
front like every other frozen shooter stat, with no post-hoc patch loop.
Manually verified M15 (Commit applies PSpread twice) still fails this test
(empirical 0.070 vs forecast 0.126) after the rewrite.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
F8 (Soul's earlier fix batch, 2026-09-24) moved range computation ahead of the
visibility gate in FireControl.PFire so Fire's own flight time stays real even
when the shot prices at PFire 0, but no test observed that behaviour --
N4 (Soul's own mutation set, reverting F8) survived every existing test.
InvisibleTargetStillResolvesAtItsRealFlightTime fires at a target removed from
VisibleEntities and checks the shot does not resolve before its real 5s
arrival (N4's bug would resolve it instantly, at range 0) and resolves as a
miss once it does.
Manually verified against the exact N4 patch: it now fails (arrival at ~1s
instead of ~6s, i.e. range treated as 0).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TwoProngHullSplitsTheScatterAcrossBothIntervals's own within-prong check
(atNearColumn > atFarColumn) held under M12/M12b (an entire second silhouette
built at the wrong sigma) and the sigma/sqrt(2) Stryker mutants, because every
wrong sigma tried still puts more mass on the nearer column -- direction alone
cannot separate a sigma of 1.5 from 1.5*sqrt(2). Their analytic near/far mass
ratios are well apart (~2.9 true vs ~1.73 at sigma*sqrt(2)), so the test now
computes the true ratio from the same per-cell Gaussian integral it already
uses and asserts the empirical ratio (8000 shots) tracks it within a tolerance
that rejects the wrong-sigma ratio.
Manually verified against the exact M12 and M12b patches from Soul's mutation
script: both now fail this test (empirical ratios 581 and 1.68, matching a
collapsed and a sqrt(2)-shifted sigma respectively, against the analytic ~2.9).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Replaces LateralDrawIsHalfOpenAtTheShadowsTopEdge (reflection into the private
LateralDraw) with:
- ShotAtBearingLoTouchingOnlyACellCornerHitsMetal and
ShotAtBearingLoOnTurningArmourHitsMetal, Soul's two named-seed crash repros
(Cut12-edge-68442709 and Cut12-edge-64078070 -- StableHash(zone name) seeds
Zone.CombatSeed, so the exact name reproduces the exact die roll) run
through the real public Fire/Zone.Update path, asserting the shot commits a
hit whose cell's own projected lateral interval contains the drawn Lateral,
instead of throwing.
- LaneNeverEmptiesAtAnIntervalEdge, the general sweep: every interval's Lo and
the point just below its own half-open Hi, across several bearings,
precisions and hulls (including a non-convex, holed one), must never return
an empty lane from FireControl.Lane.
Build gained a zoneName parameter (Cut 6.1's one-fixed-name convention doesn't
serve a repro that needs an exact CombatSeed) and a tracking parameter and an
equipLMarker fixture, both used by later findings in this batch.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
S1 (HIGH, crash in play): Lane's own shadow prefilter (a re-derivation of
Silhouette's centre/h) and the independent SlabAlongB 2-axis intersection test
could disagree at a float edge, rejecting a cell the shadow sum had already
counted metal for -- an empty lane after a hit passed its roll (R3 violation),
reproducible through Zone.Update. Both are replaced by one shared admission
test (Extent(c, axis), the exact support of a unit square along an axis) that
Silhouette and Lane now both call; SlabAlongB is deleted, and its geometry is
kept only as a non-rejecting AlongBearing() that computes entry/exit for
already-admitted cells (the doc's own equivalence proof -- "s in the shadow" is
exactly "the line intersects" -- makes rejection there unreachable by
construction, not merely unlikely).
S3 (allocation): Interval and LaneCell now implement IComparable<T> (same
order as the deleted IntervalByLoComparer/LaneCellComparer) so Silhouette's and
Lane's sorts use Array.Sort(T[], int, int), which resolves T's own comparison
at JIT time, instead of the (T[], int, int, IComparer<T>) overload, which
wraps even a cached singleton comparer in a fresh delegate per call.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
F1 (correctness): LateralDraw clamped the drawn lateral to the closed
[Lo, Hi], not the spec's half-open [Lo, Hi). Soul enumerated all 2^24
NextFloat outputs against TurningArmourIntoTheShotTakesItOnTheArmour's own
geometry and found 3 that draw exactly Hi, which Lane's own half-open filter
then admits no cell for -- a hit that passed its roll landing nowhere (R3).
Fixed the clamp; Lane's silent int2.zero fallback on an empty lane is now a
thrown InvalidOperationException, since an empty lane is provably impossible
after the fix. Committed Soul's edge probe as
LateralDrawIsHalfOpenAtTheShadowsTopEdge.
F6/F8: HitProbability, Inspect and Fire each assembled bearing -> Silhouette
-> PSpread (or recomputed range) in their own copy; a shooter-frame bearing
bug in HitProbability's copy survived while the identical bug in Inspect's
copy was caught. Collapsed into one Forecast function and one PFire(out range)
overload; both callers, and Fire, read the one computation.
F2: FireControlCut12Tests' zone names carried a Guid, breaking every other
cut's fixed-seed convention (FireControlCut6Tests.cs:62-63) and reseeding the
die every run -- the actual source of the flake in
TurningArmourIntoTheShotTakesItOnTheArmour. Fixed to a fixed zone name.
F3/F4/F5/F7: added a two-prong hull fixture (LateralDraw's walk past the first
interval, and a wrong-sigma placement, were unreachable when every fixture
projects to one interval), a real-roll empirical test for CommitProbability
(the old one called the pure function twice and compared it to itself),
enough Penetration in TurningAfterCommitChangesNothing for ApplyHit to
actually read the bearing, and a 2x2 item fixture for ResolveAimPoint's
average (every "Marker" elsewhere is 1x1, so sum/length and sum*length agree).
F9: Lane's own comment claimed no allocation while its sort used
Comparer<T>.Create over a closure; LaneCell now carries its own precomputed
projection and Lane sorts with a stateless singleton comparer. Corrected an
overclaiming comment in FireAuthorityTests (SigmaFloor makes the aimed-lane
draw ~68% per shot, not deterministic; the fixture's fixed zone name is what
actually makes it reproduce).
Cost: Silhouette takes an optional caller-rented Interval[] buffer; the
forecast path (HitProbability/Inspect, via Forecast) rents from
ArrayPool<Interval>.Shared and returns it before returning, since neither
reads sil.Intervals after the call. Commit/CommitProbability still get an
owned array, since LateralDraw needs it alive past the call. Measured on
LonginusX (66 cells, Release): unaimed 672 B -> 184 B/call (1512-1637ns before,
1482-1560ns after); aimed 1160 B -> 672 B/call (2123-2140ns before,
2094-2162ns after) -- the remaining ~488B/call gap between aimed and unaimed
is ResolveAimPoint's own CellsOf allocation (a List<int2>+ToArray per call),
not in this batch's scope.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The first Stryker pass (375 mutants, 69 survivors on this cut's own changed
lines) found one real split-authority bug and several untested edges:
- HitProbability recomputed target.Position - source.Position a second time
for PSpread's range, a second place the same subtraction could flip sign.
PFire now returns its own gate-computed range by out parameter; HitProbability
reads that one value instead of a second copy.
- TravelDirection's point-blank fallback and Fire's null-target guard around it
were only ever exercised at shooter/target pairs where the real direction
happened to equal the (0,1) fallback by coincidence.
- diagnostic.PFire (presentation) was only ever checked at PSensor == 1, where
Accuracy * PSensor and Accuracy / PSensor agree.
- CommitProbability's four-factor product was only ever checked with
pDeviation == 1 (no fixture moved a target between fire and commit), where
* and / by 1 agree.
- ResolveAimPoint's zero-cells fallback (an aimed item that no longer occupies
any cell) was never reached; every fixture's marker survives to be hit.
New tests: TravelDirectionMatchesThePredictedInterceptAndFallsBackAtPointBlank,
DiagnosticPFireIsAccuracyTimesPSensor, CommitProbabilityIsTheExactProductOfItsFourFactors,
AimPointFallsBackWhenTheAimedItemNoLongerOccupiesAnyCell.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
FireControl.Silhouette replaces HullKernel/WeightedPick/POnHull: the hull's
lateral shadow at a bearing, merged into intervals, integrated exactly with
CultMath.erf/erfinv instead of a discrete per-cell Gaussian sum. FireControl.Lane
does exact slab traversal to pick the impact cell from a lateral offset; only
its first element (the entry-nearest occupied cell) is used here, per the cut
map -- the rest is 12.3's armour-first absorption march.
Bearing timing: PendingShot.PBase is renamed PFire and now carries only the
shooter-side factor (Accuracy x PSensor) behind the gates; TravelDirection,
Spread and FireRange are frozen alongside it. The target's facing is read live
at Commit through FireControl.CommitProbability (steps 1-5 of the spec's
"Bearing timing"), not frozen at Fire, so a target that turns its armour into
an incoming shot takes the hit there. ShotOutcome gains Bearing/Lateral,
frozen at commit; Apply reads only those, never a live position or facing.
Entity.ApplyHit takes an already-schematic bearing instead of a world
direction, and the ToSchematic call 12.1 put inside it is gone.
Superseded tests: PlacementAndProbabilityShareOneKernel, EveryHitLandsOnMetal
and AimingAtTheSternHitsTheStern are replaced by
FireControlCut12Tests.PlacementAndProbabilityShareOneSilhouette,
HitsLandOnTheFacingEdge and AimingAtAnItemCentresTheScatterOnItsLane.
FireAuthorityTests.AimedHitLandsOnSelectedItem and PenetrationMarchIsPlanar are
updated for ApplyHit's new bearing parameter and Lane's entry-nearest-cell
placement (a solid hull's Tool item is never itself the entry cell; reaching
it now needs the same penetration march the new tests exercise).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The headless build pinned one CultLib revision for every package. CultMath
0.2.4 needs a newer revision than CultCache's, and bumping the single pin
would have compiled six commits of newer CultCache under the headless tests
while Unity stays on caching-unity-v1.4.0. CultMathRoot/CultMathRevision now
pin CultMath at the 0.2.4 tag; CultLibRoot/CultLibRevision keep CultCache at
45c2f40. Verified: 244 pass with both roots right, and each wrong root is
refused by name.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Lock hash is the tagged merge 6d5e2096; batchmode resolve and compile clean.
Nothing calls erf yet; Cut 12.2 will.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Entity.ToSchematic(float2) is the one schematic-frame transform;
Entity.ApplyHit's penetration march and FireControl.Splash's
directional half both call it instead of each carrying an inline
forward/right copy. ShotOutcome.Aimed and its plumbing (Q12-4 = A) are
deleted: the aimed local/flag in Commit, MakeOutcome's aimed
parameter, and the AimedCells half of ResolveAimPoint's tuple, which
now returns float2 directly.
No behaviour change. All 244 tests pass unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Records what Soul found (the wrong-signed erfinv edge that would have put hits
on the opposite side of a hull), corrects three stale verification lines, and
sets a cost measurement 12.2 owes rather than assumes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The blast radius, not penetration, picks the model: no radius means the lane
walk with cell-by-cell absorption; a radius means the shot reaches fuse depth
and detonates as an area. AP round snipes a component, explosive penetrator
guts a section.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Operator: distribute the damage over the affected area. A blast covers the
cells within its radius and each takes the share of the disc inside it;
damage over open grid is lost, so an airburst wastes most of itself and an
internal detonation delivers nearly all of it. A shot has a direction and
keeps the lane model; a blast does not. Q12-6 ruled out, Q12-7 dissolved,
Q12-8 ruled.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Records the operator rulings of 2026-09-22 (Q12-1..4, penetration retune as a
follow-up, WeaponModifiers as labels) and adds sub-cut 12.4, which replaces
Splash with one detonation owner.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The Splash sign flip Cut 11 pinned was an undefended mutant, not a shipped
bug: the line is unchanged since 482c47bf. The operator's wrong-side bug
was legacy Unity code.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Inspect.PBase is HitProbability's own answer instead of a second product of
the factors. New behavioural tests pin a beam shot carrying its override
damage, a fitted targeting system forgiving a jink, a shot orphaned before
commit committing as a plain miss, and the HUD's arrival countdown. The
tautological damage assertion in BeamRollsPerInterval is deleted.
Stryker on FireControl.cs: 51 survivors to 43, all triaged in the map.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Most of Stryker's surviving mutants in FireControl.cs traced to a handful of
fixture constants that make a mutant compute exactly the same number as the
original -- the same defect class as the zero-velocity target and the damage
that happened to equal an item's durability, found here as a pattern.
- The kernel fixture's shooter sat at float3.zero, where `target - source` and
`target + source` are the same vector. HitProbabilityMatchesInspect gains a
configuration with the shooter far from the origin: the correct separation is
500, in range; the mutated sum is 1300, out of range.
- The kernel fixture's Resolution was 1000, which saturates pSensor to exactly 1
above the detection threshold, so multiplying and dividing by it agreed. Now 2.
- EvasionCountsUntilCommitAndNotAfter fired at zone time 0, so `now - FireTime`
equalled `now + FireTime`. Cut 5 gave this test a moving target to defend its
own declared mutation and it still survived; the clock now advances first.
- Every fixture used zero spread and a cell size of 1, leaving the whole spread
term untested. PSpreadIsTheSilhouettesShareOfTheSpreadCone pins what pSpread
means -- a silhouette filling the cone scores 1, half the cone scores .5, and
a bigger one saturates -- rather than how it is computed.
Triaged as equivalent, not chased: the float-threshold boundary flips, and the
three MixSeed mutants that turn a right shift left -- measured, the die stays
fair under all three (worst deviation from p .017-.030 against .022 correct), so
killing them would mean pinning the mixer's bit arithmetic, not its behaviour.
Recorded, not fixed: FireControl.Apply's derivation of hitDirection from the two
positions has no geometry test through the FireControl path, and
StatResolverTests.AModifierThatTargetsItsOwnMagnitudeStatIsRefusedAtEquip fails
intermittently in the full suite -- likely cross-class pollution through
CultCache's process-wide document registry.
241 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Nine tests, each stating as behaviour a rule a real Stryker mutant survived:
- A shield that absorbs must eventually run dry, on both the discrete path and
Splash's (whose absorb branch had no coverage at all). Deleting Shield.TakeHit
survived because earlier tests asserted only that the hull was untouched.
- Splash damages the side the blast came from, across five facings and both
sides. Flipping the sign of `right` survived every suite, and it survives at
the default facing (0, 1) too: the flip only changes the damaged side when a
ship faces more along world x than z. So the test spins the target through
facings where it bites. The current code is correct; it can no longer regress.
- Different zones roll differently. Replacing `CombatSeed *` with `/` dropped the
zone from the seed and survived, because the die test checks one zone only.
- A shot whose target left the zone resolves as a miss.
That last one is a real defect (Soul C4), fixed here. Step rewrote the outcome
only for uncommitted shots, so a hit committed inside the horizon and then
orphaned by its target dying was republished as a hit -- a hit marker on a
corpse. The resolution is now a fresh miss; the committed outcome is left
untouched, since R4 makes it immutable.
The fixture puts the shooter off the origin on purpose: at float3.zero,
`target - source` equals `target + source`, which is how four position mutants
survived.
240 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Twenty-three harnesses, 5,816 lines, across the fire control, stats and power,
shield reserve and condition ratio campaigns. Every one targeted code under
Assets/Scripts/ServerShared, which Stryker's pinned scope now covers, and every
one carried the same liabilities: literal-text anchors coupled to code shape, a
restore that could leak a mutant into the tree, a tree-clean verdict that
restored before it measured, and a red result that could not tell a killed
mutant from a compile failure. The behavioural tests they measured all stay.
Kept and recorded, not yet answered: the addressables harness and the three
shield panel harnesses. They target Unity Editor probes that Unity compiles
outside Aetheria.Shared, where Stryker cannot reach.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Pinned in dotnet-tools.json (dotnet-stryker 5.0.0) so any checkout reproduces
it with `dotnet tool restore`. Reports land in StrykerOutput/, now ignored.
Spike against FireControl.cs alone: 276 mutants in 7m18s, 208 killed, 67
survived, 4 uncovered, score 74.6%. It killed every mutation the hand-written
anchor harnesses defended, compiles all mutants into one assembly and switches
between them at runtime -- so it never edits a source file, has nothing to
re-anchor after a refactor, and cannot leak a mutant into the tree.
Among the survivors: Splash's handedness is untested (flipping the sign of
`right` passes every test -- the operator's "damage on the wrong side of the
ship"), deleting the discrete shield's TakeHit survives, and most of the rest
trace to degenerate fixture values -- shooter at the origin, fire time zero,
spread zero -- which make whole families of mutants arithmetically identical.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The diagnostics refactor (72c0109c) kept the debug HUD and the simulation on one
fire-control model by routing HitProbability THROUGH Inspect. That was the right
invariant implemented at the wrong level: every AI and turret evaluation, per
weapon per tick, built the full diagnostic struct and ran IsRevealed's sort, the
aim-point cell walk and HullKernel's allocations and per-cell exp -- and only
then checked whether the target was even in range. A Soul pass measured 9.87 us
and 448 B per call on paths that used to bail in nanoseconds. Operator: it made
the hot path look like a diagnostic one.
One model is now shared at the level of its factors. HitProbability runs its
cheap gates first and bails, exactly as it did before the refactor, and computes
PSensor, PSpread and POnHull only for a shot that is actually possible. Inspect
calls the same four factor functions and fills the struct unconditionally,
because a HUD wants to see why a shot is impossible. The HUD panel and its one
test consumer are unchanged.
Two tests make the design enforced rather than conventional:
- HitProbabilityMatchesInspect: exact equality across 21 configurations covering
every gate the hot path bails on plus Precision and info sweeps. Dies when a
factor changes in only one caller.
- GatedOutHitProbabilityAllocatesNothing: 1000 out-of-range calls allocate zero
bytes. Dies when HitProbability is routed back through Inspect.
Both mutations verified RED and distinguished from a compile failure, with
restores checked against a hash taken before mutating.
231 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The combat roll was not a roll. FireControl seeded CultMath's xorshift32 per
shot and drew once; NextFloat returns the top 24 bits of the first round, and
one round cannot carry ShotId's low-bit variation into the high bits. Every
zone's hit roll spanned about an eighth with its threshold fixed at zone
creation, so a shot was a step function of p -- below the band it never hit,
above it always did. An unaided shooter hit 0 of 18,926 shots, which is the
operator's "I can target and fire with no hits registering," and being stable
per zone is why it came and went.
Measured across six zone seeds: hit fraction at p=.3 went from 0.0% in five of
six zones to 29-31%, at p=.6 to 58-61%, spanning the full range. The seed stays
a pure function of (zone, shot id); only its diffusion changes.
Three cuts shipped over this because nothing in the campaign measured a
distribution -- every fixture uses accuracy 1 or 0, and a constant satisfies a
determinism test perfectly. The missing test, not the seeding line, was the
defect.
Also: pOnHull collapsed to zero above Precision ~4 because no shipped hull's
centre of mass is integral, so a better targeting system made a weapon unable to
hit anything; sigma is now floored at half a cell. And the LockWeapon crash Cut
8 did not actually fix -- that code only runs on active entities -- is closed at
its real route, dock/target dies/undock.
229 tests. All nine harnesses clean, run sequentially in an isolated worktree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
9.1 changed the line those two harnesses pin, stranding five anchors including
cut6's own no-op control -- so cut6 was defending nothing until now. Same thing
the diagnostics refactor did to four harnesses earlier, and the second time in a
day that a change to Commit has stranded anchors elsewhere. The rules are
unchanged; only the text moved.
That maintenance cost is now a known property of pinning mutations to literal
source text. It buys a stale anchor failing loudly within one run, which is how
both instances were caught immediately, and it belongs in Cut 10's list beside
Soul's harness audit.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
9.1, blocking. FireControl seeded CultMath's xorshift32 per shot and drew once.
NextFloat returns the top 24 bits of the FIRST round, and one round cannot carry
ShotId's low-bit variation into the high bits -- so every zone's hit roll spanned
about an eighth, with the threshold fixed at zone creation. A shot was a step
function of p: below the band it never hit, above it always did. An unaided
shooter hit 0 of 18,926 shots. That is the operator's "no hits registering," and
being stable per zone is why it came and went.
The seed derivation is unchanged -- still a pure function of (zone, shot id),
Cut 6b's rule -- and only its diffusion into the first output changes, via an
fmix32 finalizer. Measured across six zone seeds: hit fraction at p=.3 went from
0.0% in five of six zones to 29-31%, and at p=.6 to 58-61%, with the roll
spanning the full range.
The defect survived three cuts because nothing in the campaign measured a
distribution -- every fixture uses accuracy 1 or 0, and a constant satisfies a
determinism test perfectly. TheDieIsUniform now asserts the observed hit
fraction is within .08 of p across several zone seeds, which is the test whose
absence was the real defect.
9.2. Sigma is floored at half a cell. HullKernel approximates a continuous
Gaussian with a discrete sum over occupied cells, and that sum collapses once
sigma falls well under one cell against an aim point that is not exactly on a
cell centre -- true of every shipped hull. Past Precision ~4, pOnHull went to
zero and a better targeting system made the weapon unable to hit anything.
9.3. An active entity's Target is now always a live entity it has info on.
Cut 8's _active guard could never have fixed the operator's LockWeapon crash --
that code only runs on active entities, and cut8's own harness header says so.
The live route was dock, target dies, undock; Deactivate disposed the
subscription that nulls a stale Target and Activate reseeded EntityInfoGathered
without reconciling it. Still not a TryGetValue in LockWeapon.
229 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
