Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
/* This Source Code Form is subject to the terms of the Mozilla Public
|
|
|
|
|
* License, v. 2.0. If a copy of the MPL was not distributed with this
|
|
|
|
|
* file, You can obtain one at https://mozilla.org/MPL/2.0/. */
|
|
|
|
|
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
using System;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
using System.Buffers;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
using System.Collections.Generic;
|
2026-09-19 13:02:05 +02:00
|
|
|
using System.Linq;
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
using CultMath;
|
|
|
|
|
using static CultMath.math;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
using float2 = CultMath.float2;
|
|
|
|
|
using float3 = CultMath.float3;
|
|
|
|
|
using int2 = CultMath.int2;
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
using Random = CultMath.Random;
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
|
2026-09-19 13:02:05 +02:00
|
|
|
// Cut 1 (docs/fire-control-cut.md): geometry -- whether a weapon bears on a point is decided here and
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
// nowhere else -- no prefab, transform, ArticulationPoint or renderer may influence it (R6, R7). Mount
|
|
|
|
|
// direction and arc are both derived per read from catalog data and the item's authored rotation; neither
|
|
|
|
|
// is ever stored on the entity.
|
2026-09-19 13:02:05 +02:00
|
|
|
// Cut 2 adds the reveal rule (R5) and the resolved targeting-system stats an entity fires with, both
|
|
|
|
|
// likewise derived on every read and never cached.
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
public static class FireControl
|
|
|
|
|
{
|
|
|
|
|
// The planar unit vector a hardpoint's weapon points along, derived solely from the item's authored
|
|
|
|
|
// rotation (R6, the same convention reaction thrusters already use) -- never from a live Unity
|
|
|
|
|
// transform. This is the item-rotation path Behavior.Direction now uses unconditionally, after the
|
|
|
|
|
// HardpointTransforms readback branch is deleted.
|
|
|
|
|
public static float3 MountDirection(EquippedItem item)
|
|
|
|
|
{
|
|
|
|
|
var itemDirection = item.Entity.Direction.Rotate(item.EquippableItem.Rotation);
|
|
|
|
|
return normalize(float3(itemDirection.x, 0, itemDirection.y));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Full arc width in degrees: the per-hardpoint override when authored (nonzero), else
|
|
|
|
|
// GameplaySettings.FiringArc. Zero on the hardpoint means "use the default," not "no arc."
|
|
|
|
|
public static float ArcFor(EquippedItem item)
|
|
|
|
|
{
|
|
|
|
|
var hardpoint = item.Entity.Hardpoints[item.Position.x, item.Position.y];
|
|
|
|
|
var arc = hardpoint?.FiringArc ?? 0f;
|
|
|
|
|
return arc > 0 ? arc : item.ItemManager.GameplaySettings.FiringArc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Planar bearing test (R7: the simulation is 2D, the 3D is set dressing -- target height never enters
|
|
|
|
|
// this). An arc of 360 degrees or more passes unconditionally, which is how a turret hardpoint
|
|
|
|
|
// (authored FiringArc: 360, Q1) tracks all the way around without a second type.
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
// Cut 5, 5.4 (Soul finding 11): a planar bearing shorter than 1e-6 -- point-blank range, target and firer
|
|
|
|
|
// co-located -- returns true rather than falling through to normalize's NaN. This used to live as a
|
|
|
|
|
// separate special case in Weapon.ArcAllowsFire; one bearing test now owns it, so a target at a weapon's
|
|
|
|
|
// own exact position can never fail to bear regardless of caller.
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
public static bool InArc(EquippedItem weapon, float3 toTarget)
|
|
|
|
|
{
|
|
|
|
|
var arc = ArcFor(weapon);
|
|
|
|
|
if (arc >= 360f) return true;
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
var planarToTarget = float3(toTarget.x, 0, toTarget.z);
|
|
|
|
|
if (lengthsq(planarToTarget) < 1e-6f) return true;
|
|
|
|
|
var planarTarget = normalize(planarToTarget);
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
return dot(MountDirection(weapon), planarTarget) >= cos(radians(arc / 2f));
|
|
|
|
|
}
|
2026-09-19 13:02:05 +02:00
|
|
|
|
|
|
|
|
// Cut 2, R5: whether `observer` has resolved `item` (which must belong to some entity in the zone) well
|
|
|
|
|
// enough to aim at it. Derived fresh from EntityInfoGathered on every call -- nothing caches a reveal
|
|
|
|
|
// result, so decay drops it the moment info falls back below tier, with no clearing loop anywhere.
|
|
|
|
|
//
|
|
|
|
|
// Ranking: the target's own non-hull equipment, hardpoint-mounted gear first (the big obvious things),
|
|
|
|
|
// then by size descending, ties broken by equipment order (OrderByDescending is a stable sort, so this
|
|
|
|
|
// needs no explicit tie-break key). Item i of N reveals once info crosses
|
|
|
|
|
// lerp(TargetArmorInfoThreshold, TargetGearInfoThreshold, i / max(1, N - 1)) -- the first item to reveal
|
|
|
|
|
// does so at the armor tier, the last at the gear tier.
|
|
|
|
|
public static bool IsRevealed(Entity observer, EquippedItem item)
|
|
|
|
|
{
|
|
|
|
|
var target = item.Entity;
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
// Cut 5, 5.5 (Soul finding 10): a destroyed item drops out of the ranking entirely rather than merely
|
|
|
|
|
// failing its own tier -- ranked.IndexOf(item) then returns -1 for it (the same "no longer on this
|
|
|
|
|
// target" path below), and the survivors' tiers close up over the gap instead of leaving one.
|
2026-09-19 13:02:05 +02:00
|
|
|
var ranked = target.Equipment
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
.Where(e => e.Data.HardpointType != HardpointType.Hull && e.EquippableItem.Durability >= .01f)
|
2026-09-19 13:02:05 +02:00
|
|
|
.OrderByDescending(e => e.Data.HardpointType != HardpointType.Tool)
|
|
|
|
|
.ThenByDescending(e => e.Data.Shape.Coordinates.Length)
|
|
|
|
|
.ToList();
|
|
|
|
|
var index = ranked.IndexOf(item);
|
|
|
|
|
if (index < 0) return false; // the hull itself, or an item that is no longer on this target
|
|
|
|
|
|
|
|
|
|
var settings = observer.ItemManager.GameplaySettings;
|
|
|
|
|
var info = observer.EntityInfoGathered.TryGetValue(target, out var gathered) ? gathered : 0f;
|
|
|
|
|
var tier = lerp(settings.TargetArmorInfoThreshold, settings.TargetGearInfoThreshold,
|
|
|
|
|
(float) index / max(1, ranked.Count - 1));
|
|
|
|
|
return info >= tier;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 2, Q4: the stats an entity actually fires with. A working targeting system (equipped, online --
|
|
|
|
|
// not destroyed, not shut down) supplies its own resolved stats; anything else falls back to the
|
|
|
|
|
// authored unaided figures, which make aiming without one a last resort rather than an alternative.
|
|
|
|
|
// Cut 3's roll is the only planned reader.
|
|
|
|
|
public static float Accuracy(Entity entity)
|
|
|
|
|
{
|
|
|
|
|
var system = entity.GetBehavior<TargetingSystem>();
|
|
|
|
|
return system != null && system.Item.Active.Value
|
|
|
|
|
? system.Accuracy
|
|
|
|
|
: entity.ItemManager.GameplaySettings.UnaidedAccuracy;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public static float Resolution(Entity entity)
|
|
|
|
|
{
|
|
|
|
|
var system = entity.GetBehavior<TargetingSystem>();
|
|
|
|
|
return system != null && system.Item.Active.Value ? system.Resolution : 1f;
|
|
|
|
|
}
|
|
|
|
|
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
// Cut 6d (docs/fire-control-cut.md): falls back to GameplaySettings.UnaidedPrecision, the same shape
|
|
|
|
|
// Accuracy/Tracking already fall back to their own Unaided* floors. The old coin-flip rule could get away
|
|
|
|
|
// with a bare 0 here (0 meant "never draw the aimed-item branch," a legitimate value under that rule); the
|
|
|
|
|
// dart-throw kernel reads Precision as a grouping tightness, and 0 would ask Sigma for an infinite, division-
|
|
|
|
|
// guarded group -- not the authored "spraying at the silhouette" floor the spec asks for.
|
2026-09-19 13:02:05 +02:00
|
|
|
public static float Precision(Entity entity)
|
|
|
|
|
{
|
|
|
|
|
var system = entity.GetBehavior<TargetingSystem>();
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
return system != null && system.Item.Active.Value ? system.Precision : entity.ItemManager.GameplaySettings.UnaidedPrecision;
|
2026-09-19 13:02:05 +02:00
|
|
|
}
|
|
|
|
|
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
// Cut 5, 5.1 (Soul finding 3): falls back to GameplaySettings.UnaidedTracking, exactly the shape Accuracy
|
|
|
|
|
// falls back to UnaidedAccuracy. Tracking is now always positive -- Commit no longer carries a zero case.
|
2026-09-19 13:02:05 +02:00
|
|
|
public static float Tracking(Entity entity)
|
|
|
|
|
{
|
|
|
|
|
var system = entity.GetBehavior<TargetingSystem>();
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
return system != null && system.Item.Active.Value ? system.Tracking : entity.ItemManager.GameplaySettings.UnaidedTracking;
|
2026-09-19 13:02:05 +02:00
|
|
|
}
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
|
|
|
|
// Cut 3, the risk this map names explicitly: Combat.cs used to run its own first_order_intercept call to
|
|
|
|
|
// aim, while the roll measured deviation against a separately-computed prediction -- two functions
|
|
|
|
|
// claiming to answer the same question will eventually disagree, and an AI would then aim at one point and
|
|
|
|
|
// be judged against another. This is the one function; CombatState (and FireControl.Fire below) both call
|
|
|
|
|
// it instead of touching CultMath.first_order_intercept directly. Planar: source velocity is never fed in
|
|
|
|
|
// (no shooter here leads its own motion into the shot), matching every existing call site this replaces.
|
|
|
|
|
public static float3 PredictedIntercept(Weapon weapon, Entity source, Entity target)
|
|
|
|
|
{
|
|
|
|
|
var targetVelocity = float3(target.Velocity.x, 0, target.Velocity.y);
|
|
|
|
|
return weapon.Velocity > .01f
|
|
|
|
|
? first_order_intercept(source.Position, float3.zero, weapon.Velocity, target.Position, targetVelocity)
|
|
|
|
|
: target.Position;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 3, R1/R3: the targeting solution (0b table) -- pure, no draw, recomputed fresh on every call. AI
|
|
|
|
|
// reads this to decide whether a shot is worth taking; the HUD reads it to draw a number; Fire below reads
|
|
|
|
|
// it once, at the instant the trigger is pulled, and freezes the result into the shot (R10, Q6). Two reads
|
|
|
|
|
// a tick apart may differ; nothing here or anywhere else caches one.
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
// Cut 10 (docs/fire-control-cut.md): the hot path. Every AI and turret calls this per weapon per tick,
|
|
|
|
|
// and the common case is a target out of range or out of arc -- so the cheap gates run first and bail,
|
|
|
|
|
// exactly as they did before the diagnostics refactor. The factors are computed only for a shot that is
|
|
|
|
|
// actually possible, and they come from the same functions Inspect uses: one fire-control model, shared
|
|
|
|
|
// at the level of its factors rather than by routing the hot path through a diagnostic struct.
|
|
|
|
|
// HitProbabilityMatchesInspect pins that the two can never disagree.
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md): the shooter-side share of the probability -- everything Fire
|
|
|
|
|
// freezes (R10), gated exactly as HitProbability gates. Nothing about the target's facing or silhouette
|
|
|
|
|
// enters here; that is Silhouette's and PSpread's job, priced fresh by HitProbability below and again,
|
|
|
|
|
// live, by Commit at the commit tick. 0 the moment any gate closes.
|
2026-09-24 21:25:15 +02:00
|
|
|
private static float PFire(Weapon weapon, Entity source, Entity target) => PFire(weapon, source, target, out _);
|
|
|
|
|
|
|
|
|
|
// The `out range` overload is HitProbability's own gate call -- one computation of `target.Position -
|
|
|
|
|
// source.Position`, not two. HitProbability used to recompute its own copy for PSpread, which is exactly
|
|
|
|
|
// the kind of split authority Cut 3 named as the risk (two places computing the same vector, free to
|
|
|
|
|
// drift or to have their subtraction order flipped in only one of them).
|
|
|
|
|
private static float PFire(Weapon weapon, Entity source, Entity target, out float range)
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
{
|
2026-09-24 21:25:15 +02:00
|
|
|
range = 0f;
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
if (target == null) return 0f;
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// F8 correction (Soul's fix batch, 2026-09-24): range is computed for any non-null target, before any
|
|
|
|
|
// gate -- Fire reads it through this same overload for flightTime, and flightTime must reflect the
|
|
|
|
|
// real distance even when a gate (not visible, out of arc) already prices the shot at PFire 0. The
|
|
|
|
|
// gates below still short-circuit the rest of the shooter-side probability in the same order as before.
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
var toTarget = target.Position - source.Position;
|
2026-09-24 21:25:15 +02:00
|
|
|
range = length(toTarget);
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
|
|
|
|
|
if (!source.VisibleEntities.Contains(target)) return 0f;
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
if (range < weapon.MinRange || range > weapon.Range) return 0f;
|
|
|
|
|
if (weapon is LockWeapon lockWeapon && !lockWeapon.IsLocked) return 0f;
|
|
|
|
|
if (!InArc(weapon.Item, toTarget)) return 0f;
|
|
|
|
|
|
|
|
|
|
var settings = source.ItemManager.GameplaySettings;
|
|
|
|
|
var info = source.EntityInfoGathered.TryGetValue(target, out var gathered) ? gathered : 0f;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
return Accuracy(source) * PSensor(settings, Resolution(source), info);
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// Cut 12.2 fix batch (F6, Soul, 2026-09-24): the one place a bearing, a silhouette and PSpread are
|
|
|
|
|
// assembled into a live forecast -- HitProbability and Inspect both call this instead of each carrying
|
|
|
|
|
// its own copy. They used to each build their own bearing from TravelDirection and the target's facing;
|
|
|
|
|
// a mutant that read the SHOOTER's facing instead of the target's survived in HitProbability's own copy
|
|
|
|
|
// (M11) while the identical mistake in Inspect's copy was caught (M22), because two functions computing
|
|
|
|
|
// the same thing are two places for the same bug to hide and only be found once.
|
|
|
|
|
// Cost fix batch (Self, 2026-09-24): Forecast reads only sil.POnHull and sil.Span, never sil.Intervals, so
|
|
|
|
|
// it rents the Silhouette's own scratch array from ArrayPool<Interval>.Shared and returns it before
|
|
|
|
|
// returning -- the forecast path (HitProbability, Inspect) no longer allocates on a gated-in call.
|
|
|
|
|
// Commit's own path (CommitProbability -> Silhouette, uncalled from here) still needs sil.Intervals alive
|
|
|
|
|
// for the lateral draw that follows it, so it keeps its own owned array; only the forecast pools.
|
|
|
|
|
private static (Silhouette Silhouette, float PSpread) Forecast(Weapon weapon, Entity source, Entity target, HullData targetHull, float precision, float range)
|
|
|
|
|
{
|
|
|
|
|
var bearing = Bearing(target, TravelDirection(weapon, source, target));
|
|
|
|
|
var buffer = ArrayPool<Interval>.Shared.Rent(targetHull.Shape.Coordinates.Length);
|
|
|
|
|
float pSpread;
|
|
|
|
|
Silhouette sil;
|
|
|
|
|
try
|
|
|
|
|
{
|
|
|
|
|
sil = Silhouette(target, targetHull, source.ResolvedTargetItem, bearing, precision, buffer);
|
|
|
|
|
var settings = source.ItemManager.GameplaySettings;
|
|
|
|
|
pSpread = PSpread(weapon.Spread, sil.Span, range, settings.SchematicCellSize);
|
|
|
|
|
}
|
|
|
|
|
finally
|
|
|
|
|
{
|
|
|
|
|
ArrayPool<Interval>.Shared.Return(buffer);
|
|
|
|
|
}
|
|
|
|
|
// sil.Intervals is the returned buffer -- cleared here so nothing outside this function can read it
|
|
|
|
|
// after it goes back to the pool and another caller starts overwriting it.
|
|
|
|
|
sil.Intervals = null;
|
|
|
|
|
return (sil, pSpread);
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
public static float HitProbability(Weapon weapon, Entity source, Entity target)
|
|
|
|
|
{
|
2026-09-24 21:25:15 +02:00
|
|
|
var pFire = PFire(weapon, source, target, out var range);
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
if (pFire <= 0f) return 0f;
|
|
|
|
|
|
|
|
|
|
var targetHull = source.ItemManager.GetData(target.Hull) as HullData;
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
var (sil, pSpread) = Forecast(weapon, source, target, targetHull, Precision(source), range);
|
|
|
|
|
return pFire * pSpread * sil.POnHull;
|
2026-09-20 21:58:44 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
// Presentation only: the debug HUD's view of exactly the factors HitProbability multiplies. It computes
|
|
|
|
|
// every factor unconditionally, gates included, because a HUD wants to see why a shot is impossible --
|
|
|
|
|
// which is precisely why the hot path must not be routed through it. Both read the same factor
|
2026-09-22 16:39:52 +02:00
|
|
|
// functions below, and PBase is HitProbability's own answer, so the gates and the product have one
|
|
|
|
|
// owner and the HUD cannot grow a second, subtly different model.
|
2026-09-20 21:58:44 +02:00
|
|
|
public static FireControlDiagnostic Inspect(Weapon weapon, Entity source, Entity target)
|
|
|
|
|
{
|
|
|
|
|
var settings = source.ItemManager.GameplaySettings;
|
|
|
|
|
var diagnostic = new FireControlDiagnostic
|
|
|
|
|
{
|
|
|
|
|
HasTarget = target != null,
|
|
|
|
|
Accuracy = Accuracy(source),
|
|
|
|
|
Resolution = Resolution(source),
|
|
|
|
|
Precision = Precision(source),
|
|
|
|
|
Tracking = Tracking(source),
|
|
|
|
|
MinRange = weapon.MinRange,
|
|
|
|
|
MaxRange = weapon.Range
|
|
|
|
|
};
|
|
|
|
|
if (target == null) return diagnostic;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
|
|
|
|
var toTarget = target.Position - source.Position;
|
2026-09-20 21:58:44 +02:00
|
|
|
diagnostic.Range = length(toTarget);
|
|
|
|
|
diagnostic.Visible = source.VisibleEntities.Contains(target);
|
|
|
|
|
diagnostic.InRange = diagnostic.Range >= weapon.MinRange && diagnostic.Range <= weapon.Range;
|
|
|
|
|
diagnostic.Locked = !(weapon is LockWeapon lockWeapon) || lockWeapon.IsLocked;
|
|
|
|
|
diagnostic.InArc = InArc(weapon.Item, toTarget);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
var targetHull = source.ItemManager.GetData(target.Hull) as HullData;
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
diagnostic.Info = source.EntityInfoGathered.TryGetValue(target, out var gathered) ? gathered : 0f;
|
|
|
|
|
diagnostic.InfoDemandCeiling = InfoDemandCeiling(settings, diagnostic.Resolution);
|
|
|
|
|
diagnostic.PSensor = PSensor(settings, diagnostic.Resolution, diagnostic.Info);
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
diagnostic.PFire = diagnostic.Accuracy * diagnostic.PSensor;
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
var (sil, pSpread) = Forecast(weapon, source, target, targetHull, diagnostic.Precision, diagnostic.Range);
|
|
|
|
|
diagnostic.PSpread = pSpread;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
diagnostic.POnHull = sil.POnHull;
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
|
2026-09-22 16:39:52 +02:00
|
|
|
diagnostic.PBase = HitProbability(weapon, source, target);
|
2026-09-20 21:58:44 +02:00
|
|
|
return diagnostic;
|
|
|
|
|
}
|
|
|
|
|
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
// Cut 6c, 6c.1 (operator ruling 2026-09-19): Resolution stays a benefit -- higher is better -- so the
|
|
|
|
|
// formula takes its reciprocal to derive the actual info ceiling instead of reading Resolution as that
|
|
|
|
|
// ceiling directly. Resolution 1 (the unaided fallback) yields a ceiling of 1: needs complete
|
|
|
|
|
// information, the floor by construction. The max guards a zero or negative authored Resolution; no
|
|
|
|
|
// authored value should ever reach it.
|
|
|
|
|
private static float InfoDemandCeiling(GameplaySettings settings, float resolution) =>
|
|
|
|
|
settings.TargetDetectionInfoThreshold + (1f - settings.TargetDetectionInfoThreshold) / max(resolution, 1e-3f);
|
|
|
|
|
|
|
|
|
|
private static float PSensor(GameplaySettings settings, float resolution, float info) =>
|
|
|
|
|
saturate(unlerp(settings.TargetDetectionInfoThreshold, InfoDemandCeiling(settings, resolution), info));
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Whether the weapon's own barrel-dispersion cone reaches the target's actual silhouette at this range:
|
|
|
|
|
// weapon hardware, range-dependent, aim-point-blind. 1 for any zero-spread weapon. Cut 12.2: span comes
|
|
|
|
|
// from the same Silhouette pOnHull reads, so the two factors can never disagree about the target's size
|
|
|
|
|
// (the old bounding half-extent, max(Width, Height), is gone).
|
|
|
|
|
private static float PSpread(float spread, float span, float range, float cellSize)
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
if (spread <= 0) return 1f;
|
|
|
|
|
var halfExtent = .5f * span * cellSize;
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
var angularRadius = degrees(atan(halfExtent / range));
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
return saturate(angularRadius / (spread / 2f));
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md, "the rules every sub-cut reads"): the shot's world-planar travel
|
|
|
|
|
// direction -- PredictedIntercept minus the source's own position, planar (R7). Frozen into PendingShot
|
|
|
|
|
// at Fire (R10); HitProbability and Inspect recompute it fresh, at the current source/target positions,
|
|
|
|
|
// for their forecast. Below a length of 1e-6 (point-blank, source and intercept coincide) it falls back
|
|
|
|
|
// to world +z, the same fallback Apply used before this cut (`:463`).
|
|
|
|
|
public static float2 TravelDirection(Weapon weapon, Entity source, Entity target)
|
|
|
|
|
{
|
|
|
|
|
var toIntercept = (PredictedIntercept(weapon, source, target) - source.Position).xz;
|
|
|
|
|
return lengthsq(toIntercept) < 1e-6f ? float2(0, 1) : normalize(toIntercept);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 12.2: the only place a world direction meets a target's facing. One function, so nothing else may
|
|
|
|
|
// compute a bearing (the authority map's forbidden-writer rule) -- Commit and the live forecast both read
|
|
|
|
|
// this, never a second copy of the formula.
|
|
|
|
|
private static float2 Bearing(Entity target, float2 travelDirection) => normalize(target.ToSchematic(travelDirection));
|
|
|
|
|
|
|
|
|
|
// Cut 12.2: the lateral axis a bearing implies, fixed once here. Silhouette and Lane both read it instead
|
|
|
|
|
// of each carrying their own copy of the sign.
|
|
|
|
|
private static float2 Lateral(float2 bearing) => float2(-bearing.y, bearing.x);
|
Cut 10: the hot path stops looking like, and costing like, a diagnostic
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>
2026-09-22 15:18:12 +02:00
|
|
|
|
2026-09-20 21:58:44 +02:00
|
|
|
public static float DeviationProbability(PendingShot shot, float now, out float deviation)
|
|
|
|
|
{
|
|
|
|
|
var elapsed = now - shot.FireTime;
|
|
|
|
|
var predicted = shot.FireTargetPosition + shot.FireTargetVelocity * elapsed;
|
|
|
|
|
deviation = shot.Target == null ? 0f : length((shot.Target.Position - predicted).xz);
|
|
|
|
|
return shot.Target == null ? 1f : saturate(1f - deviation / shot.Tracking);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
}
|
|
|
|
|
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
// Cut 3, R1: called once per burst step from InstantWeapon.Execute. Computes flight time, freezes the
|
|
|
|
|
// payload snapshot (R10, Q6 -- the gun that fired it, not a re-read later), and queues a PendingShot for
|
|
|
|
|
// Zone.Step to age and eventually resolve. Returns the ShotId so the caller's OnFire event can carry it to
|
|
|
|
|
// presentation. Cut 5, 5.7 (Soul finding 12): does not store a predicted intercept -- Commit judges
|
|
|
|
|
// deviation against a straight-line projection from FireTargetPosition/FireTargetVelocity, never the
|
|
|
|
|
// intercept, so a stored copy decided nothing and PendingShot no longer carries one.
|
2026-09-19 15:16:57 +02:00
|
|
|
// Cut 4 (docs/fire-control-cut.md): damageOverride lets a continuous weapon fire a shot for less than its
|
|
|
|
|
// full Damage -- ConstantWeapon.Execute rolls one of these per GameplaySettings.BeamResolveInterval, for
|
|
|
|
|
// Damage * interval, through this exact same freeze-and-queue path (a flight time of zero, since a beam's
|
|
|
|
|
// authored Velocity is 0, commits and resolves in the tick it fires -- R4's short-flight degradation,
|
|
|
|
|
// unchanged). Every other caller keeps reading the frozen weapon.Damage it always did.
|
|
|
|
|
public static int Fire(Weapon weapon, EquippedItem item, Entity source, float? damageOverride = null)
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
{
|
|
|
|
|
var zone = source.Zone;
|
|
|
|
|
var target = source.Target.Value;
|
|
|
|
|
var now = zone.Time;
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// R1's engage gate and the shooter-side probability, evaluated now and frozen: nothing at commit time
|
|
|
|
|
// re-reads a stat, an info level or a range. Cut 12.2: PFire no longer carries the spread/hull share --
|
|
|
|
|
// those are priced live at Commit, against the target's facing then, not now (Bearing timing, R10).
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// F8 (Soul's fix batch, 2026-09-24): the `out range` overload is the one range HitProbability now
|
|
|
|
|
// reads too -- Fire used to recompute its own second copy of `target.Position - source.Position` for
|
|
|
|
|
// fireRange, right beside the copy PFire's gate already computed.
|
|
|
|
|
var pFire = PFire(weapon, source, target, out var fireRange);
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var travelDirection = target != null ? TravelDirection(weapon, source, target) : float2(0, 1);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
|
|
|
|
var targetVelocity = float3.zero;
|
|
|
|
|
var targetPosition = source.Position;
|
|
|
|
|
var flightTime = 0f;
|
|
|
|
|
if (target != null)
|
|
|
|
|
{
|
|
|
|
|
targetVelocity = float3(target.Velocity.x, 0, target.Velocity.y);
|
|
|
|
|
targetPosition = target.Position;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
flightTime = weapon.Velocity > .01f ? fireRange / weapon.Velocity : 0f;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
var commitHorizon = source.ItemManager.GameplaySettings.CommitHorizon;
|
2026-09-19 18:34:04 +02:00
|
|
|
|
|
|
|
|
// Cut 6b, 6.2 (Soul finding 5): airburst is a property of the weapon, resolved and frozen here like
|
|
|
|
|
// every other payload field -- never decided later by a Unity projectile. A weapon without the
|
|
|
|
|
// Airburst flag freezes a zero radius, which Step below reads as "resolve this shot with Apply."
|
|
|
|
|
var weaponItemData = item.Data as WeaponItemData;
|
|
|
|
|
var isAirburst = weaponItemData != null && weaponItemData.WeaponModifiers.HasFlag(WeaponModifiers.Airburst);
|
|
|
|
|
var burstPosition = isAirburst && target != null ? PredictedIntercept(weapon, source, target) : targetPosition;
|
Fix the merge of Cuts 6b and 6c: a new float field made the catalog unreadable
Cut 6b added WeaponItemData.AirburstRange as a non-nullable float at Key(29).
Every WeaponItemData record in GameData/Aetheria.cc was serialized before that
key existed, so the reader is handed nil for the absent slot and a float cannot
take nil -- AetheriaStores.Open threw MessagePackSerializationException and the
game could not read its own catalog. AirburstRange becomes float?, null meaning
no authored burst radius, which is also the more honest reading of a value no
product sets yet.
Neither cut could have caught this alone. 6b verified with unit tests over
synthetic fixtures and never opened the shipped catalog; 6c opened the catalog
constantly but on a branch without the new field. It appears only in the merge,
and only when something actually reads GameData.
The general rule, which will recur: a new non-nullable value-type field on a
persisted CultCache document cannot read records written before it existed.
Nullable, or a migration that rewrites every affected record -- and a migration
cannot run if the read already throws, so tolerant deserialization comes first.
Verified on the merged tree: catalog reads (53 designs, 47 products), 0 dangling
refs, 215 tests pass, and loadout with all three targeting sellers excluded
generates every hull with 1 armed hull falling back to unaided fire.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 18:56:53 +02:00
|
|
|
var burstRadius = isAirburst ? weaponItemData.AirburstRange ?? 0f : 0f;
|
2026-09-19 18:34:04 +02:00
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
var shot = new PendingShot
|
|
|
|
|
{
|
|
|
|
|
ShotId = zone.NextShotId(),
|
|
|
|
|
Source = source,
|
|
|
|
|
Target = target,
|
|
|
|
|
Weapon = item,
|
|
|
|
|
Aimed = source.ResolvedTargetItem,
|
2026-09-19 15:16:57 +02:00
|
|
|
Damage = damageOverride ?? weapon.Damage,
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
Penetration = weapon.Penetration,
|
|
|
|
|
DamageSpread = weapon.DamageSpread,
|
|
|
|
|
DamageType = weapon.WeaponData.DamageType,
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
PFire = pFire,
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
Tracking = Tracking(source),
|
|
|
|
|
Precision = Precision(source),
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
TravelDirection = travelDirection,
|
|
|
|
|
Spread = weapon.Spread,
|
|
|
|
|
FireRange = fireRange,
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
FireTime = now,
|
|
|
|
|
FireTargetPosition = targetPosition,
|
|
|
|
|
FireTargetVelocity = targetVelocity,
|
|
|
|
|
ArrivalTime = now + flightTime,
|
|
|
|
|
CommitTime = now + max(0f, flightTime - commitHorizon),
|
2026-09-19 18:34:04 +02:00
|
|
|
BurstPosition = burstPosition,
|
|
|
|
|
BurstRadius = burstRadius,
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
Committed = false
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
zone.PendingShots.Add(shot);
|
|
|
|
|
return shot.ShotId;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 3, R4: ages every shot in the zone, commits the ones that have reached their horizon and resolves
|
|
|
|
|
// the ones that have arrived. Called from Zone.Update after the entity loop. A shot whose source or target
|
|
|
|
|
// has left the zone resolves as a miss and is removed outright (0b table), whichever stage it is at.
|
|
|
|
|
public static void Step(Zone zone, float dt)
|
|
|
|
|
{
|
|
|
|
|
var shots = zone.PendingShots;
|
|
|
|
|
if (shots.Count == 0) return;
|
|
|
|
|
|
|
|
|
|
var now = zone.Time;
|
|
|
|
|
for (var i = shots.Count - 1; i >= 0; i--)
|
|
|
|
|
{
|
|
|
|
|
var shot = shots[i];
|
|
|
|
|
|
|
|
|
|
var sourceGone = !zone.Entities.Contains(shot.Source);
|
|
|
|
|
var targetGone = shot.Target != null && !zone.Entities.Contains(shot.Target);
|
|
|
|
|
if (sourceGone || targetGone)
|
|
|
|
|
{
|
Cut 11.4: behavioural tests for what Stryker found undefended, and Soul C4
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>
2026-09-22 16:06:33 +02:00
|
|
|
// Cut 11 (Soul C4): a shot whose source or target has left the zone resolves as a miss whichever
|
|
|
|
|
// stage it is at (0b table). The resolution is published as a fresh miss rather than by
|
|
|
|
|
// rewriting shot.Outcome: a committed outcome is immutable (R4) and stays a true record of what
|
|
|
|
|
// the commit decided, while ShotResolved reports what actually happened -- nothing, because the
|
|
|
|
|
// target is gone. Republishing the committed outcome here used to put a hit marker on a corpse.
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var miss = MakeOutcome(shot, false, false, false, int2.zero, float2.zero, 0f, now);
|
Cut 11.4: behavioural tests for what Stryker found undefended, and Soul C4
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>
2026-09-22 16:06:33 +02:00
|
|
|
if (!shot.Committed) zone.ShotCommitted.OnNext(miss);
|
|
|
|
|
zone.ShotResolved.OnNext(miss);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
shots.RemoveAt(i);
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!shot.Committed && now >= shot.CommitTime)
|
|
|
|
|
{
|
|
|
|
|
shot.Outcome = Commit(zone, shot, now);
|
|
|
|
|
shot.Committed = true;
|
|
|
|
|
zone.ShotCommitted.OnNext(shot.Outcome);
|
|
|
|
|
shots[i] = shot;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (shot.Committed && now >= shot.ArrivalTime)
|
|
|
|
|
{
|
2026-09-19 18:34:04 +02:00
|
|
|
// Cut 6b, 6.2: an airburst shot (frozen BurstRadius > 0) resolves as an area effect instead of
|
|
|
|
|
// a discrete hit -- Splash instead of Apply, never both, which is exactly the double-
|
|
|
|
|
// application Soul was told to hunt for.
|
|
|
|
|
if (shot.BurstRadius > 0f) Splash(zone, shot.BurstPosition, shot.BurstRadius, shot.Damage, shot.DamageType);
|
|
|
|
|
else Apply(shot);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
zone.ShotResolved.OnNext(shot.Outcome);
|
|
|
|
|
shots.RemoveAt(i);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// R3/R4: the one roll. p_base was frozen at fire; the only thing measured live is how far the target has
|
|
|
|
|
// actually strayed, by now, from where a straight-line projection of its fire-time velocity said it would
|
Cut 7: retire a test Cut 6b made vacuous, and the invariant it claimed
Cut 3's harness reported OutOfArcConsumesNoDraw as "not yet reached" -- no
mutation could turn it red. Chasing that rather than accepting it found the
reason, and it is not a fixture defect.
The test asserted that a shot gated to zero draws nothing from ItemManager.
Random, so the shared stream advanced exactly as if the shot had never queued.
Cut 6b moved every roll onto a per-shot local generator seeded from (zone, shot
id), so Commit stopped touching that stream on EVERY path. The test then passed
for a reason unrelated to arcs -- it passed identically with the target dead
ahead -- and the rule it named stopped existing along with the stream it was
about. The short-circuit it guarded is now a cheap guard, not an invariant, and
Commit's comment was still asserting otherwise.
Rewritten as CombatNeverDrawsFromTheSharedStream, a Theory over both an
out-of-arc and an in-arc target, pinning the rule that IS live after 6b: no
combat path draws from the shared stream at all. Verified by hand -- restoring
the shared-stream read in Commit turns the in-arc case red while the out-of-arc
case stays green, which is exactly why the old test could not have caught it.
218 tests pass. All six mutation harnesses run clean with tree-clean PASS.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 20:14:16 +02:00
|
|
|
// be -- forgiven by the targeting system's Tracking, likewise frozen at fire.
|
|
|
|
|
//
|
|
|
|
|
// Cut 7: the short-circuit below (`p > 0f &&`) is a guard, NOT an invariant. It used to be one -- a shot
|
|
|
|
|
// gated to zero at fire had to draw nothing so the shared stream advanced exactly as if it had never
|
|
|
|
|
// queued -- but Cut 6b gave every shot its own generator, seeded from (zone, shot id) and discarded here,
|
|
|
|
|
// so whether a gated shot draws is unobservable from anywhere. The claim this comment used to make is
|
|
|
|
|
// gone with the stream it was about, and the test that asserted it became vacuous the same moment; what
|
|
|
|
|
// replaced it pins the live rule instead (combat touches no shared stream, on any path).
|
Cut 9: the die is not a die
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>
2026-09-22 02:33:22 +02:00
|
|
|
// 9.1 (docs/fire-control-cut.md): fmix32, MurmurHash3's 32-bit finalizer -- xor-shift, multiply,
|
|
|
|
|
// xor-shift, multiply, xor-shift. Diffuses a structured seed (here, a zone seed XORed with a small,
|
|
|
|
|
// low-bits-only ShotId) across every bit before the first xorshift draw reads it, so two seeds that
|
|
|
|
|
// differ only in ShotId's low bits still land on uncorrelated first outputs. Caller-side mixing only
|
|
|
|
|
// -- CultMath.Random itself is unchanged (see Commit's own note on why that follow-up is a separate
|
|
|
|
|
// CultLib cut).
|
|
|
|
|
private static uint MixSeed(uint seed)
|
|
|
|
|
{
|
|
|
|
|
seed ^= seed >> 16;
|
|
|
|
|
seed *= 0x85ebca6bu;
|
|
|
|
|
seed ^= seed >> 13;
|
|
|
|
|
seed *= 0xc2b2ae35u;
|
|
|
|
|
seed ^= seed >> 16;
|
|
|
|
|
return seed;
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md, "Bearing timing"): steps 1-5. Everything the shooter decided is
|
|
|
|
|
// frozen in `shot` (R10); everything the target is doing -- position (deviation) and now also facing
|
|
|
|
|
// (bearing) -- is read live, here, at the commit tick. The Cut 7 guard extends to cover the new factors:
|
|
|
|
|
// when PFire x pDeviation is already 0, no Silhouette is built. Shared by Commit's own roll and by the
|
|
|
|
|
// debug HUD's forecast (TheHudEstimateIsTheCommitPrice) -- one function, one commit-time price.
|
|
|
|
|
public static float CommitProbability(PendingShot shot, float now, out Silhouette sil)
|
|
|
|
|
{
|
|
|
|
|
sil = default;
|
|
|
|
|
if (shot.Target == null) return 0f;
|
|
|
|
|
|
|
|
|
|
var pDeviation = DeviationProbability(shot, now, out _);
|
|
|
|
|
var p = shot.PFire * pDeviation;
|
|
|
|
|
if (p <= 0f) return 0f;
|
|
|
|
|
|
|
|
|
|
var bearing = Bearing(shot.Target, shot.TravelDirection);
|
|
|
|
|
var hullData = shot.Source.ItemManager.GetData(shot.Target.Hull) as HullData;
|
|
|
|
|
sil = Silhouette(shot.Target, hullData, shot.Aimed, bearing, shot.Precision);
|
|
|
|
|
var settings = shot.Source.ItemManager.GameplaySettings;
|
|
|
|
|
return p * PSpread(shot.Spread, sil.Span, shot.FireRange, settings.SchematicCellSize) * sil.POnHull;
|
|
|
|
|
}
|
|
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
private static ShotOutcome Commit(Zone zone, PendingShot shot, float now)
|
|
|
|
|
{
|
2026-09-19 18:34:04 +02:00
|
|
|
// Cut 6b, 6.1 (Soul finding 6): a shot's dice belong to the shot, not to whatever else happened to
|
2026-09-19 18:41:38 +02:00
|
|
|
// draw from the engine-wide shared generator first. A pure function of (zone identity, shot id) --
|
|
|
|
|
// the `| 1u` guards
|
2026-09-19 18:34:04 +02:00
|
|
|
// the degenerate zero seed -- so two runs of the same fight from the same galaxy seed roll identically
|
|
|
|
|
// no matter what the UI drew from the shared stream in between. This generator is local and dies with
|
|
|
|
|
// the call: nothing outside Commit may seed or advance a combat draw, and Commit may not read or write
|
|
|
|
|
// a shared stream.
|
Cut 9: the die is not a die
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>
2026-09-22 02:33:22 +02:00
|
|
|
// 9.1 (docs/fire-control-cut.md): the per-shot seed is unchanged -- a pure function of
|
|
|
|
|
// (zone identity, shot id), Cut 6b's own rule -- but CultMath.Random draws only one xorshift
|
|
|
|
|
// round for NextFloat's first call, and one round does not diffuse a seed that differs from its
|
|
|
|
|
// neighbours only in ShotId's low bits. Unmixed, every zone's first draw was a near-constant
|
|
|
|
|
// function of ShotId (measured span ~0.125), making a shot a step function of p fixed at zone
|
|
|
|
|
// creation, not a roll (Soul C1). MixSeed is a standard fmix32 finalizer (MurmurHash3): this is
|
|
|
|
|
// a caller's obligation, not a CultMath defect -- seeding xorshift with a structured value and
|
|
|
|
|
// drawing once is the documented way to get a correlated first output, and a follow-up to give
|
|
|
|
|
// CultMath.Random a mixed-seed entry point of its own is recorded for a separate CultLib cut.
|
|
|
|
|
var random = new Random(MixSeed((zone.CombatSeed * 2654435761u) ^ (uint) shot.ShotId | 1u));
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var p = CommitProbability(shot, now, out var sil);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
|
|
|
|
var hit = p > 0f && random.NextFloat() < p;
|
|
|
|
|
var cell = int2.zero;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var bearing = float2.zero;
|
|
|
|
|
var lateral = 0f;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
var shielded = false;
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
var shieldBroken = false;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
|
|
|
|
if (hit)
|
|
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2: the same silhouette pOnHull drew its mass from -- same bearing (live, at this commit
|
|
|
|
|
// tick), same sigma (the frozen Precision). A shot that passed the roll always lands on metal (R3:
|
|
|
|
|
// one roll decides damage; the off-hull share was already priced into that roll, not resolved here
|
|
|
|
|
// as a second stage). The lateral draw picks where along the shadow the shot lands; Lane converts
|
|
|
|
|
// that (bearing, lateral) pair into the actual cell -- only the first element (the impact cell,
|
|
|
|
|
// nearest the shooter's side) is used here; the rest of the lane is 12.3's armour march.
|
|
|
|
|
bearing = Bearing(shot.Target, shot.TravelDirection);
|
|
|
|
|
lateral = LateralDraw(sil, random.NextFloat());
|
|
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
var hullData = shot.Source.ItemManager.GetData(shot.Target.Hull) as HullData;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var buffer = ArrayPool<LaneCell>.Shared.Rent(hullData.Shape.Coordinates.Length);
|
|
|
|
|
try
|
|
|
|
|
{
|
|
|
|
|
var count = Lane(hullData, bearing, lateral, buffer);
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// R3: a hit that passed its roll always lands on metal. With LateralDraw's half-open clamp,
|
|
|
|
|
// the drawn `lateral` always falls strictly inside the interval sil.POnHull priced, so Lane
|
|
|
|
|
// must find at least one occupied cell -- an empty lane here is not a miss to fall back on
|
|
|
|
|
// silently (that would put a hit marker on (0,0), which may not even be occupied); it is proof
|
|
|
|
|
// the draw and the lane have drifted apart.
|
|
|
|
|
if (count == 0)
|
|
|
|
|
throw new InvalidOperationException(
|
|
|
|
|
$"shot {shot.ShotId}: Lane found no occupied cell at bearing {bearing} lateral {lateral} " +
|
|
|
|
|
"after a hit passed its roll -- LateralDraw and Lane have drifted apart (R3).");
|
|
|
|
|
cell = buffer[0].Cell;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
}
|
|
|
|
|
finally
|
|
|
|
|
{
|
|
|
|
|
ArrayPool<LaneCell>.Shared.Return(buffer);
|
|
|
|
|
}
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
|
|
|
|
|
var shield = shot.Target.Shield;
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
var shieldActive = shield != null && shield.Item.Active.Value;
|
|
|
|
|
if (shieldActive && shield.CanTakeHit(shot.DamageType, shot.Damage)) shielded = true;
|
|
|
|
|
else if (shieldActive) shieldBroken = true;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
return MakeOutcome(shot, hit, shielded, shieldBroken, cell, bearing, lateral, now);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// R4: the commit is authoritative and immutable from here on -- this only performs what Commit already
|
|
|
|
|
// decided. The one owner of the shield-absorbs-or-hull-takes-it branch (it used to exist seven times,
|
|
|
|
|
// once per Unity effect).
|
|
|
|
|
private static void Apply(PendingShot shot)
|
|
|
|
|
{
|
|
|
|
|
if (!shot.Outcome.Hit) return;
|
|
|
|
|
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
// Cut 5, 5.2 (Soul finding 4): the frozen decision, not a re-check -- a shield that recharges during
|
|
|
|
|
// the flight does not retroactively survive a hit that broke it at commit time.
|
|
|
|
|
if (shot.Outcome.ShieldBroken) shot.Target.Shield.Break();
|
|
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
if (shot.Outcome.Shielded)
|
|
|
|
|
{
|
|
|
|
|
shot.Target.Shield.TakeHit(shot.DamageType, shot.Damage);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (R4): the committed geometry, frozen. Apply reads no position or facing of its own --
|
|
|
|
|
// shot.Outcome.Bearing is already the schematic-frame bearing Commit computed at the commit tick, so
|
|
|
|
|
// turning the target after commit changes nothing here (TurningAfterCommitChangesNothing).
|
|
|
|
|
shot.Target.ApplyHit(shot.Source, shot.Outcome.Cell, shot.DamageSpread, shot.Penetration, shot.Damage, shot.Outcome.Bearing);
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
private static ShotOutcome MakeOutcome(PendingShot shot, bool hit, bool shielded, bool shieldBroken, int2 cell, float2 bearing, float lateral, float now)
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
{
|
|
|
|
|
return new ShotOutcome
|
|
|
|
|
{
|
|
|
|
|
ShotId = shot.ShotId,
|
|
|
|
|
Source = shot.Source,
|
|
|
|
|
Target = shot.Target,
|
|
|
|
|
Weapon = shot.Weapon,
|
|
|
|
|
Hit = hit,
|
|
|
|
|
Shielded = shielded,
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
ShieldBroken = shieldBroken,
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
Cell = cell,
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
Bearing = bearing,
|
|
|
|
|
Lateral = lateral,
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
ArrivalIn = max(0f, shot.ArrivalTime - now),
|
|
|
|
|
DamageType = shot.DamageType
|
|
|
|
|
};
|
|
|
|
|
}
|
|
|
|
|
|
2026-09-19 15:16:57 +02:00
|
|
|
// Cut 4 (docs/fire-control-cut.md): the one splash rule -- an unconditional area effect, not a rolled
|
|
|
|
|
// shot. Nothing here draws: a mine or an airburst round always damages everything it catches, the same
|
|
|
|
|
// as the Physics.OverlapSphere queries this replaces always did. Applies the shield-absorb-or-hull-takes-
|
|
|
|
|
// it branch (F5, docs/stats-and-power-cut.md) per target, then -- for a target the shield didn't fully
|
|
|
|
|
// absorb -- DamageSchematic over the directional half of that target's own hull that faces the blast, the
|
|
|
|
|
// rule moved verbatim from the Splash subscription EntityInstance.cs carried before Cut 3 deleted it
|
|
|
|
|
// (git show b7743789^:Assets/Scripts/Gameplay/EntityInstance.cs), made planar (R7): the blast-to-target
|
|
|
|
|
// direction is measured in the zone's (x,z) plane and rotated into each target's own facing by its
|
|
|
|
|
// Direction, the same rotation Entity.ApplyHit's penetration march uses, in place of a Unity
|
|
|
|
|
// InverseTransformDirection.
|
|
|
|
|
public static void Splash(Zone zone, float3 position, float radius, float damage, DamageType damageType)
|
|
|
|
|
{
|
|
|
|
|
foreach (var target in zone.Entities)
|
|
|
|
|
{
|
|
|
|
|
var toTarget = (target.Position - position).xz;
|
|
|
|
|
if (length(toTarget) > radius) continue;
|
|
|
|
|
|
|
|
|
|
var shield = target.Shield;
|
|
|
|
|
var shieldActive = shield != null && shield.Item.Active.Value;
|
|
|
|
|
var shieldAbsorbs = shieldActive && shield.CanTakeHit(damageType, damage);
|
|
|
|
|
if (shieldActive && !shieldAbsorbs) shield.Break();
|
|
|
|
|
if (shieldAbsorbs)
|
|
|
|
|
{
|
|
|
|
|
shield.TakeHit(damageType, damage);
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
var hullData = target.ItemManager.GetData(target.Hull) as HullData;
|
|
|
|
|
var localDirection = lengthsq(toTarget) > 1e-6f
|
2026-09-24 18:55:53 +02:00
|
|
|
? normalize(target.ToSchematic(toTarget))
|
2026-09-19 15:16:57 +02:00
|
|
|
: float2(0, 1);
|
|
|
|
|
|
|
|
|
|
var hitShape = new Shape(hullData.Shape.Width, hullData.Shape.Height);
|
|
|
|
|
foreach (var v in hullData.Shape.Coordinates)
|
|
|
|
|
if (dot(normalize((float2) v - hullData.Shape.CenterOfMass), localDirection) < 0)
|
|
|
|
|
hitShape[v] = true;
|
|
|
|
|
|
|
|
|
|
target.DamageSchematic(damage, hitShape);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md): Sigma keeps its Cut 6d/9.2 rule unchanged -- the frozen Precision's
|
|
|
|
|
// reciprocal, in hull-schematic cell units, floored at half a cell. The 9.2 floor's old rationale ("the
|
|
|
|
|
// discrete sum undershoots the continuous integral") is retired: Silhouette below integrates the Gaussian
|
|
|
|
|
// exactly over the shadow's intervals, so an exact 1D integral does not undershoot. The floor survives as
|
|
|
|
|
// a design minimum on group tightness (F12-3, TheSigmaFloorHoldsAtHalfACell) -- a targeting system is never
|
|
|
|
|
// authored tighter than half a schematic cell's width of spread, regardless of what the exact integral
|
|
|
|
|
// would otherwise allow.
|
Cut 9: the die is not a die
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>
2026-09-22 02:33:22 +02:00
|
|
|
private const float SigmaFloor = 0.5f;
|
|
|
|
|
private static float Sigma(float precision) => max(1f / max(precision, 1e-3f), SigmaFloor);
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// The aim point a silhouette is centred on (R10/Q6: the same frozen Aimed both HitProbability and Fire
|
|
|
|
|
// read) -- the aimed item's own cell centroid when it currently occupies cells on the target, else the
|
|
|
|
|
// hull's own centre of mass. One path; the uniform-random branch Cut 6d replaced stays deleted.
|
|
|
|
|
private static float2 ResolveAimPoint(Entity target, HullData hullData, EquippedItem aimed)
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var aimedCells = aimed != null ? CellsOf(target, aimed) : null;
|
|
|
|
|
if (aimedCells != null && aimedCells.Length > 0)
|
|
|
|
|
return aimedCells.Aggregate(float2.zero, (total, c) => total + (float2) c) / aimedCells.Length;
|
|
|
|
|
return hullData.Shape.CenterOfMass;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 12.2 (docs/fire-control-cut.md, "the rules every sub-cut reads"): the merged lateral shadow the
|
|
|
|
|
// hull casts across the bearing, and the exact 1D Gaussian mass it carries -- the one function HitProbability's
|
|
|
|
|
// forecast, Inspect's HUD factors and Commit's roll and placement all read. Nowhere else may compute a
|
|
|
|
|
// sigma, a projection or a bearing. Allocates one interval array of length N (occupied-cell count, <=128 on
|
|
|
|
|
// shipped hulls) plus a sort -- only on this, the gated-in, path (Cut 10's own gated-out guard is
|
|
|
|
|
// unaffected: HitProbability and CommitProbability both bail on a zero shooter/deviation factor before
|
|
|
|
|
// this is ever called).
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// Cost fix batch (Self, 2026-09-24): `buffer`, when supplied, replaces the `new Interval[coords.Length]`
|
|
|
|
|
// allocation -- the caller owns renting and returning it (ArrayPool<Interval>.Shared, sized to at least
|
|
|
|
|
// coords.Length). Only HitProbability's and Inspect's forecast reads POnHull/Span and is done with the
|
|
|
|
|
// array before returning, so only they pool it; Commit keeps its own array alive across the LateralDraw
|
|
|
|
|
// call that follows, so it still passes null and gets an owned allocation. A caller-supplied buffer may be
|
|
|
|
|
// larger than coords.Length (ArrayPool rents "at least"), so every loop below is bounded by coords.Length,
|
|
|
|
|
// never buffer.Length.
|
|
|
|
|
public static Silhouette Silhouette(Entity target, HullData hull, EquippedItem aimed, float2 bearing, float precision, Interval[] buffer = null)
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
{
|
|
|
|
|
var ell = Lateral(bearing);
|
|
|
|
|
var coords = hull.Shape.Coordinates;
|
|
|
|
|
var h = (abs(ell.x) + abs(ell.y)) / 2f;
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
var intervals = buffer ?? new Interval[coords.Length];
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
for (var i = 0; i < coords.Length; i++)
|
|
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var centre = dot((float2) coords[i], ell);
|
|
|
|
|
intervals[i] = new Interval { Lo = centre - h, Hi = centre + h };
|
|
|
|
|
}
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
Array.Sort(intervals, 0, coords.Length, IntervalByLoComparer.Instance);
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
|
|
|
|
|
// Merge in place: sorted ascending by Lo, so an interval overlaps (or exactly abuts) the last kept
|
|
|
|
|
// interval whenever its own Lo does not exceed that interval's Hi.
|
|
|
|
|
var count = 0;
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
for (var i = 0; i < coords.Length; i++)
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
{
|
|
|
|
|
if (count > 0 && intervals[i].Lo <= intervals[count - 1].Hi)
|
|
|
|
|
intervals[count - 1].Hi = max(intervals[count - 1].Hi, intervals[i].Hi);
|
|
|
|
|
else
|
|
|
|
|
intervals[count++] = intervals[i];
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
}
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
|
|
|
|
|
var sigma = Sigma(precision);
|
|
|
|
|
var aimPoint = ResolveAimPoint(target, hull, aimed);
|
|
|
|
|
var a = dot(aimPoint, ell);
|
|
|
|
|
|
|
|
|
|
var pOnHull = 0f;
|
|
|
|
|
for (var i = 0; i < count; i++)
|
|
|
|
|
pOnHull += Phi((intervals[i].Hi - a) / sigma) - Phi((intervals[i].Lo - a) / sigma);
|
|
|
|
|
|
|
|
|
|
var span = count > 0 ? intervals[count - 1].Hi - intervals[0].Lo : 0f;
|
|
|
|
|
|
|
|
|
|
return new Silhouette
|
|
|
|
|
{
|
|
|
|
|
Intervals = intervals,
|
|
|
|
|
Count = count,
|
|
|
|
|
AimPoint = aimPoint,
|
|
|
|
|
A = a,
|
|
|
|
|
Sigma = sigma,
|
|
|
|
|
Span = span,
|
|
|
|
|
POnHull = saturate(pOnHull)
|
|
|
|
|
};
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Phi(z) = 1/2 (1 + erf(z / sqrt(2))): the standard normal CDF, CultMath's erf the one function underneath
|
|
|
|
|
// every Gaussian-mass sum this cut computes.
|
|
|
|
|
private static float Phi(float z) => .5f * (1f + erf(z * (1f / SQRT2)));
|
|
|
|
|
|
|
|
|
|
private const float SQRT2 = 1.4142135f;
|
|
|
|
|
|
|
|
|
|
// Cut 12.2: the lateral draw. One NextFloat u, taken after the hit roll -- walks the cumulative interval
|
|
|
|
|
// mass to u * POnHull to select interval k, then inverts the Gaussian CDF inside it. `p` here is the
|
|
|
|
|
// absolute CDF value at the drawn point (Phi(lo_k) plus the leftover target mass within interval k), not
|
|
|
|
|
// a value renormalised to [0,1] -- inverting that gives s directly in the same units A and Sigma are in.
|
|
|
|
|
// Clamped into [lo_k, hi_k) at the end to absorb rounding (a guard, not an invariant); erfinv's own
|
|
|
|
|
// handling of the domain edges (+-infinity at p = 0 or 1) means the clamp is what actually keeps s finite
|
|
|
|
|
// for an interval with negligible mass.
|
|
|
|
|
private static float LateralDraw(Silhouette sil, float u)
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var target = u * sil.POnHull;
|
|
|
|
|
var cumulative = 0f;
|
|
|
|
|
for (var k = 0; k < sil.Count; k++)
|
|
|
|
|
{
|
|
|
|
|
var interval = sil.Intervals[k];
|
|
|
|
|
var loValue = Phi((interval.Lo - sil.A) / sil.Sigma);
|
|
|
|
|
var hiValue = Phi((interval.Hi - sil.A) / sil.Sigma);
|
|
|
|
|
var mass = hiValue - loValue;
|
|
|
|
|
|
|
|
|
|
if (target <= cumulative + mass || k == sil.Count - 1)
|
|
|
|
|
{
|
|
|
|
|
var p = clamp(loValue + (target - cumulative), loValue, hiValue);
|
|
|
|
|
var s = sil.A + sil.Sigma * SQRT2 * erfinv(2f * p - 1f);
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// Half-open [lo_k, hi_k), matching the shadow interval itself ("Each occupied cell c projects
|
|
|
|
|
// to [dot(c,ell)-h, dot(c,ell)+h)"). A draw that rounds to exactly Hi is a real, reachable
|
|
|
|
|
// float value (Soul found 3 of them enumerating all 2^24 NextFloat outputs on the
|
|
|
|
|
// TurningArmour geometry) -- clamping to the closed [Lo, Hi] let s == Hi through, which Lane's
|
|
|
|
|
// own half-open filter (`s >= centre + h` continues) then rejects for every cell, returning an
|
|
|
|
|
// empty lane for a shot that already passed its roll (R3 violation). The epsilon matches
|
|
|
|
|
// Lane's own contiguity tolerance; every merged interval is at least one cell's shadow width
|
|
|
|
|
// wide (>= 2h), far above it.
|
|
|
|
|
var hiExclusive = max(interval.Lo, interval.Hi - 1e-4f);
|
|
|
|
|
return clamp(s, interval.Lo, hiExclusive);
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
cumulative += mass;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return sil.A; // sil.Count == 0: no metal in the shadow at all -- unreachable while POnHull > 0 gated Commit's roll
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md, "the rules every sub-cut reads"): exact slab traversal along the
|
|
|
|
|
// bearing at a fixed lateral offset s, replacing the old 0.5-step sampling march. Collects the occupied
|
|
|
|
|
// cells whose shadow interval contains s, each with its own entry/exit parameter along b, ordered by
|
|
|
|
|
// entry (ties broken by dot(cell, b), then cell index for full determinism), then walks forward while
|
|
|
|
|
// consecutive cells are contiguous (`next.entry <= current.exit + 1e-4`) -- the first gap ends the walk,
|
|
|
|
|
// the same rule Entity.cs's old march kept. A direct hit starts outside the hull (t -> -infinity), so the
|
|
|
|
|
// walk's first element is the impact cell. Only that first element is read in 12.2; the rest is 12.3's
|
|
|
|
|
// armour-first absorption march. Writes into the caller-supplied pooled buffer and returns the walked
|
|
|
|
|
// count; never allocates on its own.
|
|
|
|
|
public static int Lane(HullData hull, float2 b, float s, LaneCell[] buffer)
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
var ell = Lateral(b);
|
|
|
|
|
var h = (abs(ell.x) + abs(ell.y)) / 2f;
|
|
|
|
|
var coords = hull.Shape.Coordinates;
|
|
|
|
|
|
|
|
|
|
var n = 0;
|
|
|
|
|
for (var i = 0; i < coords.Length; i++)
|
|
|
|
|
{
|
|
|
|
|
var c = (float2) coords[i];
|
|
|
|
|
var centre = dot(c, ell);
|
|
|
|
|
if (s < centre - h || s >= centre + h) continue; // outside this cell's own lateral shadow
|
|
|
|
|
|
|
|
|
|
if (!SlabAlongB(c, b, ell, s, out var entry, out var exit)) continue; // the shadow test is a conservative bound; the exact slab can still miss
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// F9 (Soul's fix batch, 2026-09-24): Projection (dot(cell, b)) is precomputed here, into the
|
|
|
|
|
// struct, so the sort below needs no closure over `b` -- LaneCellComparer.Instance is a stateless
|
|
|
|
|
// singleton, not a per-call Comparer<T>.Create allocation the "never allocates" comment used to
|
|
|
|
|
// contradict.
|
|
|
|
|
buffer[n++] = new LaneCell { Cell = coords[i], Entry = entry, Exit = exit, Projection = dot(c, b) };
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
}
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
Array.Sort(buffer, 0, n, LaneCellComparer.Instance);
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
|
|
|
|
|
var walked = n > 0 ? 1 : 0;
|
|
|
|
|
for (var i = 1; i < n; i++)
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
{
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
if (buffer[i].Entry > buffer[i - 1].Exit + 1e-4f) break;
|
|
|
|
|
walked++;
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
}
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
return walked;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// The slab (ray-box) intersection of point(t) = t*b + s*ell against the unit square centred on cell c, one
|
|
|
|
|
// axis at a time: solving bAxis*t + val in [cLo, cHi] for each of x and y, then entry = the later of the
|
|
|
|
|
// two lower bounds, exit = the earlier of the two upper bounds. A near-zero bAxis component means the ray
|
|
|
|
|
// does not move along that axis at all, so it either always satisfies that axis's bound (no constraint on
|
|
|
|
|
// t) or never does (no intersection).
|
|
|
|
|
private static bool SlabAlongB(float2 c, float2 b, float2 ell, float s, out float entry, out float exit)
|
|
|
|
|
{
|
|
|
|
|
entry = float.NegativeInfinity;
|
|
|
|
|
exit = float.PositiveInfinity;
|
|
|
|
|
return SlabAxis(b.x, c.x - .5f, c.x + .5f, s * ell.x, ref entry, ref exit)
|
|
|
|
|
&& SlabAxis(b.y, c.y - .5f, c.y + .5f, s * ell.y, ref entry, ref exit);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
private static bool SlabAxis(float bAxis, float lo, float hi, float val, ref float entry, ref float exit)
|
|
|
|
|
{
|
|
|
|
|
if (abs(bAxis) < 1e-9f)
|
|
|
|
|
return val >= lo && val <= hi; // no motion along this axis: in bounds forever, or never
|
|
|
|
|
|
|
|
|
|
var t1 = (lo - val) / bAxis;
|
|
|
|
|
var t2 = (hi - val) / bAxis;
|
|
|
|
|
entry = max(entry, min(t1, t2));
|
|
|
|
|
exit = min(exit, max(t1, t2));
|
|
|
|
|
return entry <= exit;
|
Cut 6d: the dart-throw kernel replaces the coin-flip landing rule
FireControl.Commit used to place a hit with a coin flip between exactly
the aimed item and a uniform-random hull cell -- no spatial term, so a
near-miss on a stern item was as likely to land in the bow as the stern.
One Gaussian kernel, in hull-schematic cell units, now decides both
halves of "where will this shot go": w(cell) = exp(-d^2/2*sigma^2) over
every occupied cell, sigma = 1/Precision. HitProbability folds the
kernel's on-hull mass share (pOnHull) into PBase at fire time (R10/Q6
unaffected); Commit draws the landing cell from the same weights. A shot
that passes the roll always lands on metal -- the off-hull chance is
priced into that one roll, never resolved as a second miss stage. The
uniform-random branch is deleted outright, not demoted.
Precision now feeds every shot's pOnHull, aimed or not, so its unaided
fallback (FireControl.Precision) reads GameplaySettings.UnaidedPrecision
instead of a bare 0 -- the same shape as the existing UnaidedAccuracy/
UnaidedTracking floors. pOnHull is saturated: the discrete sum over-
estimates the continuous kernel's integral once sigma drops below about
one cell, and a mass share cannot exceed the whole kernel.
Existing fixtures across FireAuthorityTests, FireControlCut4/5/6Tests and
TargetingSystemTests authored Precision as 0 (or relied on the old 0
fallback) purely as a don't-care value under the coin-flip rule, where it
only gated the aimed-item branch. Under the kernel it drives every shot's
sigma, so those fixtures are re-authored to keep pOnHull near its
ceiling and stay pinned on what they actually test.
mutation_tests_fire_control_cut3.sh's two anchors into HitProbability's
return line and the deleted coin-flip branch are updated to the
equivalent lines/mutations on the new code.
2026-09-19 21:37:39 +02:00
|
|
|
}
|
|
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
// The cells of the target's hull schematic actually occupied by `item` -- GearOccupancy is the one source
|
|
|
|
|
// of truth for where an equipped item's footprint lands, the same table Entity.DamageSchematic reads.
|
|
|
|
|
private static int2[] CellsOf(Entity target, EquippedItem item)
|
|
|
|
|
{
|
|
|
|
|
var hullData = target.ItemManager.GetData(target.Hull) as HullData;
|
|
|
|
|
var cells = new List<int2>();
|
|
|
|
|
foreach (var v in hullData.Shape.Coordinates)
|
|
|
|
|
if (target.GearOccupancy[v.x, v.y] == item)
|
|
|
|
|
cells.Add(v);
|
|
|
|
|
return cells.ToArray();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 3 (docs/fire-control-cut.md, 0b table): a shot in flight. Born in FireControl.Fire, aged by
|
|
|
|
|
// FireControl.Step (called from Zone.Update after the entity loop), never serialised -- a zone saved mid-
|
|
|
|
|
// flight loses its shots, which is correct, a save is a scene boundary. Zone owns the collection and the
|
|
|
|
|
// ShotId; FireControl owns every transition. Nothing else may add, remove or mutate one.
|
|
|
|
|
public struct PendingShot
|
|
|
|
|
{
|
|
|
|
|
public int ShotId;
|
|
|
|
|
public Entity Source;
|
|
|
|
|
public Entity Target;
|
|
|
|
|
public EquippedItem Weapon;
|
|
|
|
|
public EquippedItem Aimed;
|
|
|
|
|
|
|
|
|
|
// The frozen payload (R10, Q6): the gun's own stats at the instant the trigger was pulled. Nothing
|
|
|
|
|
// downstream re-evaluates any of these.
|
|
|
|
|
public float Damage;
|
|
|
|
|
public float Penetration;
|
|
|
|
|
public float DamageSpread;
|
|
|
|
|
public DamageType DamageType;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md): renamed from PBase -- its meaning changes from "the whole shot's
|
|
|
|
|
// frozen roll price" to "the shooter's own share of it," behind the same gates (Accuracy x PSensor). The
|
|
|
|
|
// spread and hull factors are no longer frozen here; they are priced live at Commit, against the target's
|
|
|
|
|
// facing then (Bearing timing).
|
|
|
|
|
public float PFire;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
public float Tracking;
|
|
|
|
|
public float Precision;
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2: frozen at Fire alongside the rest of the shooter's own decision (R10) -- the shot's
|
|
|
|
|
// world-planar travel direction (FireControl.TravelDirection), and the weapon stats PSpread reads at
|
|
|
|
|
// Commit against the target's live silhouette.
|
|
|
|
|
public float2 TravelDirection;
|
|
|
|
|
public float Spread;
|
|
|
|
|
public float FireRange;
|
|
|
|
|
|
2026-09-19 18:34:04 +02:00
|
|
|
// Cut 6b, 6.2: the airburst payload, frozen at fire alongside everything else above. BurstRadius is zero
|
|
|
|
|
// for a weapon without the Airburst flag (WeaponModifiers, Enums.cs) -- Step reads that zero as "resolve
|
|
|
|
|
// with Apply," not a separate bool.
|
|
|
|
|
public float3 BurstPosition;
|
|
|
|
|
public float BurstRadius;
|
|
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
public float3 FireTargetPosition;
|
|
|
|
|
public float3 FireTargetVelocity;
|
|
|
|
|
public float FireTime;
|
|
|
|
|
public float CommitTime;
|
|
|
|
|
public float ArrivalTime;
|
|
|
|
|
|
|
|
|
|
public bool Committed;
|
|
|
|
|
public ShotOutcome Outcome;
|
|
|
|
|
}
|
|
|
|
|
|
2026-09-20 21:58:44 +02:00
|
|
|
public struct FireControlDiagnostic
|
|
|
|
|
{
|
|
|
|
|
public bool HasTarget;
|
|
|
|
|
public bool Visible;
|
|
|
|
|
public bool InRange;
|
|
|
|
|
public bool Locked;
|
|
|
|
|
public bool InArc;
|
|
|
|
|
public float Range;
|
|
|
|
|
public float MinRange;
|
|
|
|
|
public float MaxRange;
|
|
|
|
|
public float Info;
|
|
|
|
|
public float InfoDemandCeiling;
|
|
|
|
|
public float Accuracy;
|
|
|
|
|
public float Resolution;
|
|
|
|
|
public float Precision;
|
|
|
|
|
public float Tracking;
|
|
|
|
|
public float PSensor;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (optional, presentation): the shooter-side factor alone, Accuracy x PSensor -- the same value
|
|
|
|
|
// Fire freezes into PendingShot.PFire.
|
|
|
|
|
public float PFire;
|
2026-09-20 21:58:44 +02:00
|
|
|
public float PSpread;
|
|
|
|
|
public float POnHull;
|
|
|
|
|
public float PBase;
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (docs/fire-control-cut.md): a merged interval of a hull's lateral shadow, in the schematic frame's
|
|
|
|
|
// lateral (ell) coordinate.
|
|
|
|
|
public struct Interval
|
|
|
|
|
{
|
|
|
|
|
public float Lo;
|
|
|
|
|
public float Hi;
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// Cost fix batch (Self, 2026-09-24): a stateless singleton, not a per-call Comparer<T>.Create -- Silhouette's
|
|
|
|
|
// own sort no longer allocates a wrapper on every call, matching the ArrayPool pooling this batch added on the
|
|
|
|
|
// forecast path.
|
|
|
|
|
internal sealed class IntervalByLoComparer : IComparer<Interval>
|
|
|
|
|
{
|
|
|
|
|
public static readonly IntervalByLoComparer Instance = new IntervalByLoComparer();
|
|
|
|
|
private IntervalByLoComparer() { }
|
|
|
|
|
public int Compare(Interval x, Interval y) => x.Lo.CompareTo(y.Lo);
|
|
|
|
|
}
|
|
|
|
|
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2: the hull's lateral shadow at one bearing, and the exact Gaussian mass it carries. Intervals holds
|
|
|
|
|
// the merged, sorted shadow (only the first Count entries are valid -- the array is sized to the hull's own
|
|
|
|
|
// occupied-cell count, not trimmed, to avoid a second allocation). A, Sigma and POnHull are read together by
|
|
|
|
|
// the lateral draw; Span is read by PSpread so the two factors can never disagree about the target's size.
|
|
|
|
|
public struct Silhouette
|
|
|
|
|
{
|
|
|
|
|
public Interval[] Intervals;
|
|
|
|
|
public int Count;
|
|
|
|
|
public float2 AimPoint;
|
|
|
|
|
public float A;
|
|
|
|
|
public float Sigma;
|
|
|
|
|
public float Span;
|
|
|
|
|
public float POnHull;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Cut 12.2: one occupied cell FireControl.Lane crossed at a fixed lateral offset, with its own slab entry/exit
|
|
|
|
|
// parameter along the bearing. Entry ascending is nearest-to-farthest along the shot's own travel direction.
|
|
|
|
|
public struct LaneCell
|
|
|
|
|
{
|
|
|
|
|
public int2 Cell;
|
|
|
|
|
public float Entry;
|
|
|
|
|
public float Exit;
|
Cut 12.2 Soul fix batch: half-open clamp, split assembly, real rolls
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>
2026-09-24 23:08:19 +02:00
|
|
|
// F9 (Soul's fix batch, 2026-09-24): dot(Cell, b), precomputed once per candidate so Lane's sort tie-break
|
|
|
|
|
// needs no closure over the bearing.
|
|
|
|
|
public float Projection;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// F9: Lane's own sort order (entry ascending, ties by projection along b, then cell index for full
|
|
|
|
|
// determinism) as a stateless singleton -- Array.Sort's IComparer<T> overload needs an instance, and this one
|
|
|
|
|
// is allocated exactly once, ever, not per Lane() call.
|
|
|
|
|
internal sealed class LaneCellComparer : IComparer<LaneCell>
|
|
|
|
|
{
|
|
|
|
|
public static readonly LaneCellComparer Instance = new LaneCellComparer();
|
|
|
|
|
private LaneCellComparer() { }
|
|
|
|
|
|
|
|
|
|
public int Compare(LaneCell x, LaneCell y)
|
|
|
|
|
{
|
|
|
|
|
var byEntry = x.Entry.CompareTo(y.Entry);
|
|
|
|
|
if (byEntry != 0) return byEntry;
|
|
|
|
|
var byProjection = x.Projection.CompareTo(y.Projection);
|
|
|
|
|
if (byProjection != 0) return byProjection;
|
|
|
|
|
var byX = x.Cell.x.CompareTo(y.Cell.x);
|
|
|
|
|
return byX != 0 ? byX : x.Cell.y.CompareTo(y.Cell.y);
|
|
|
|
|
}
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
}
|
|
|
|
|
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
// Cut 3 (docs/fire-control-cut.md, 0b table): a commit. Created once, at ArrivalTime - CommitHorizon (or at
|
|
|
|
|
// fire time when the flight is shorter than the horizon), immutable from then on, published once on
|
|
|
|
|
// Zone.ShotCommitted and again (unchanged) on Zone.ShotResolved at arrival. Presentations are the only
|
|
|
|
|
// readers, and nothing outside presentation may read one to change state.
|
|
|
|
|
public sealed class ShotOutcome
|
|
|
|
|
{
|
|
|
|
|
public int ShotId;
|
|
|
|
|
public Entity Source;
|
|
|
|
|
public Entity Target;
|
|
|
|
|
public EquippedItem Weapon;
|
|
|
|
|
public bool Hit;
|
|
|
|
|
public bool Shielded;
|
Fire control Cut 5: close the rules that landed on one path and not its twin
Soul's pass over the whole campaign confirmed eleven findings. Seven were code
defects of one shape -- a rule implemented on one path and left off the path
beside it -- and this is all seven.
5.1 Tracking falls back to GameplaySettings.UnaidedTracking (new, 10) instead of
zero, and Commit drops the zero branch that turned a Tracking-less shooter
into a hard wall. Unaided fire was meant to be really, really bad; it was
334 shots and no hits against a target drifting at 5 u/s.
5.2 A shot the shield cannot absorb now breaks it. The decision is frozen into
ShotOutcome.ShieldBroken at commit and performed in Apply, so a shield that
recharges mid-flight cannot retroactively survive. Splash already did this;
the discrete roll never did, which made the whole shield-break mechanic
unreachable for every weapon in the catalog.
5.3 ConstantWeapon gates on ArcAllowsFire alongside StanceAllowsFire. A
side-mounted beam was firing forward.
5.4 The point-blank bearing case moves out of Weapon.ArcAllowsFire and into
InArc, which used to normalize the zero vector into NaN. The trigger said
fire and the probability said impossible.
5.5 A destroyed item drops out of IsRevealed's ranking, so Precision stops
concentrating hits onto a dead subsystem's footprint.
5.6 Zone subscribes Entity.Death and removes the corpse itself. That removal
lived only in Unity's loot-drop subscription, so headless nothing ever left
the zone, shots landed on corpses and no fight could end.
5.7 PendingShot drops PredictedIntercept and FlightTime, written and read
nowhere.
208 tests pass (197 plus eleven new).
Known defect, recorded rather than hidden: the committed mutation harness is
unreliable. Across two runs it failed to restore FireControl.cs and Weapon.cs,
leaving live mutations in the working tree -- including its own 5.1 mutation,
which is how a green suite briefly reported two failures. Its verdicts are
therefore evidence only where re-run from a verified-clean tree. Two anchors
(Zone.cs's ObserveAdd line, Weapon.cs's InArc return) fail to match at all, and
5.3's ConstantWeaponObeysArc survives its own mutation, so that rule is written
but not pinned. Cut 5b owns the repair; nothing here is trusted on the harness's
word, and the suite result above was taken by hand from a hash-verified tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 16:56:40 +02:00
|
|
|
// Cut 5, 5.2 (Soul finding 4): frozen alongside the rest of the outcome (R4) -- a shield present, active,
|
|
|
|
|
// and unable to CanTakeHit this shot is decided broken right here, so Apply performs Break() rather than
|
|
|
|
|
// deciding it, and a shield that recharges mid-flight cannot retroactively dodge it.
|
|
|
|
|
public bool ShieldBroken;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
public int2 Cell;
|
Cut 12.2: where a hit lands is priced and drawn from one silhouette
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>
2026-09-24 20:40:50 +02:00
|
|
|
// Cut 12.2 (R4): the committed geometry, frozen alongside the rest of the outcome -- the schematic-frame
|
|
|
|
|
// bearing and the lateral offset the roll's placement drew. Apply reads these, not a live position or
|
|
|
|
|
// facing (TurningAfterCommitChangesNothing).
|
|
|
|
|
public float2 Bearing;
|
|
|
|
|
public float Lateral;
|
Fire control Cut 3: hit authority moves to FireControl
Deletes the Unity physics hit path (HullCollider.Hit/Splash, and the
raycast/shield/SendHit blocks in Projectile, GuidedProjectile, Laser and
Lightning) and replaces it with a single owner in the same cut: FireControl
now decides whether a shot hits, Zone owns the pending-shot lifecycle, and
Entity.DamageSchematic/ApplyHit (moved verbatim from EntityInstance) apply the
result. Presentation subscribes to Zone.ShotCommitted/ShotResolved and applies
nothing.
FireControl.HitProbability is the pure targeting solution (Accuracy * pSensor
* pSpread, gated by detection/range/lock/arc). FireControl.Fire freezes the
weapon's payload and the shooter's resolved targeting stats at the instant the
trigger is pulled (Q6), computes the predicted intercept through the one
shared PredictedIntercept function (CombatState now calls the same function
instead of duplicating first_order_intercept), and queues a PendingShot.
FireControl.Step commits each shot's outcome at ArrivalTime - CommitHorizon --
deviation from the predicted intercept, measured once, live, at that instant
-- and applies it at ArrivalTime. A shot whose flight is shorter than the
horizon commits and resolves in the same tick it fires. Rolls draw from the
shooter's own ItemManager.Random (Q5, seeded and reproducible); an out-of-arc
or otherwise gated shot draws nothing.
Player fire is arc-gated through the same predicate AI and turret fire use
(Weapon.ArcAllowsFire, checked in InstantWeapon.Trigger -- the one point every
shooter's trigger passes through), closing Q2.
Adds GameplaySettings.CommitHorizon, SchematicCellSize and
AgentMinHitProbability (first guesses; the headless suite is the tuning
harness, operator play smoke is the arbiter).
FORK reported, not worked around silently: HullCollider.SendHit/SendSplash
cannot be deleted yet without breaking ConstantLaser.cs, ConstantLightning.cs,
ConstantParticleWeapon.cs, HitscanEffect.cs and Mine.cs, which still call them
from their own live Physics queries -- that's Cut 4's migration. They fire
into the void today: nothing subscribes to Hit/Splash any more for the four
weapon kinds this cut owns.
tests/Aetheria.Shared.Tests/FireAuthorityTests.cs pins the roll's seeding, the
gate-consumes-no-draw rule, resolve-on-arrival, the commit horizon and its
evasion window, short-flight degradation, the frozen-payload snapshot, aimed
hits, the item-then-hull damage order, the planar penetration march and the
shield-absorbs branch. FireControlTests.cs and InputCapacitorTests.cs are
updated for the new AI gate and OnFire(int) signature.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 14:07:07 +02:00
|
|
|
public float ArrivalIn;
|
|
|
|
|
public DamageType DamageType;
|
Delete HardpointTransforms; mount direction and arc become sim data
Entity.HardpointTransforms was written only by Unity (EntityInstance.cs),
so the simulation's notion of where a gun points was a readback from a
renderer that never runs headless -- Combat.cs threw KeyNotFoundException
on the first combat tick outside Unity. Deletes the field, its writer, and
the readback branch in Behavior.Direction, leaving the item-rotation path
(the same convention reaction thrusters use, R6) as the only source of
mount direction.
Adds FireControl (ServerShared), a static class owning whether a weapon
bears on a point: MountDirection (planar, from ItemRotation), ArcFor (the
per-hardpoint HardpointData.FiringArc override, falling back to the new
GameplaySettings.FiringArc default of 120), and InArc (the planar dot-cone
test, R7 -- target height never enters it, and an arc of 360+ passes
unconditionally so a turret hardpoint needs no new type, just a number).
Combat.cs and TurretController.cs now gate AI fire through InArc instead
of the old hardcoded 0.99 dot-product test against a live Unity transform.
Settings.cs drops AgentFiringMinDot, which was declared and read by
nothing.
docs/fire-control-cut.md, Cut 1.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-19 07:37:52 +02:00
|
|
|
}
|