Namespace Serenity.UnityProceduralAnimator.Tests.Infrastructure

Classes

ConditionTests

Tests for the inline Condition hierarchy: AlwaysTrue, Threshold (hysteresis), Range, And/Or/Not composites, plus validation messages. State is supplied via a FakeStateReader and UnityProceduralContext.

GraphicColorSequenceLayerSOTests

Tests for GraphicColorSequenceLayerSO: early-return guards, the cycles-gated hold-last-step behavior it shares with TransformPositionSequenceLayerSO (no wrap-seam unwrap — UnityEngine.Color.Lerp(UnityEngine.Color,UnityEngine.Color,System.Single) has no angular wraparound ambiguity, same as UnityEngine.Vector3.Lerp(UnityEngine.Vector3,UnityEngine.Vector3,System.Single)), and — the deliberate difference from every transform-facet sequence layer — that weight is written to the delta VERBATIM rather than pre-blended toward identity here (see Weight's own remarks: the applier performs the weight-blend, since a Graphic's base color is only known there). Drives the layer directly via its public 6-delta Evaluate override with a crafted UnityProceduralContext — the same seam TransformPositionSequenceLayerSOTests uses.

LayerAndActionSOTests

Tests for representative ProceduralLayerSO / ProceduralActionSO assets that are deterministic without the per-frame animator loop: PeriodicOscillationLayerSO (uses ctx.Time) and FlinchActionSO (driven directly via ActionRuntimeState).

ProceduralActionSchedulerTests

PlayMode tests for ProceduralActionScheduler: starting, cooldown gating, retrigger, per-frame ticking, retirement of finished actions, and stopping everything at once.

ProceduralAnimatorProfileSOIdentityTests

Tests for ProceduralAnimatorProfileSO's IFoundationSettings identity (Id/Guid), separate from its rule/layer/action composition tests.

ProceduralAnimatorTargetBinderTests

PlayMode tests for ProceduralAnimatorTargetBinder. Builds a real hierarchy with a real Renderer and Material, drives registration from a real ProceduralAnimatorProfileSO, and asserts what reached the appliers plus what was reported as skipped.

ProceduralBoneExclusionSetTests

PlayMode tests for ProceduralBoneExclusionSet: which bone offsets survive a strip, and that replacing the set is not additive.

ProceduralLayerDrivenChannelsTests

Tests for DrivenChannels: a table-driven assertion for every concrete layer type in the codebase today, plus a reflection-based consistency check that a type's declared channels always match which base-class Evaluate overload it actually overrides — the contract DrivenChannels's own doc comment states in prose.

ProceduralLayerWeightSchedulerTests

PlayMode tests for ProceduralLayerWeightScheduler: seeding from a resolved configuration, the first-tick snap, blend-in and blend-out ramps driven by a rule, the lazily stamped ON_ACTIVATE layer clock, and auto-deactivation once a repeat window elapses.

RuleSOTests

Tests for RuleSO: default condition behavior, condition evaluation, weight mapping via an injected WeightProvider, and the RuleEffect enum.

SettingsRefTests

Tests for the serializable ID reference structs (BoneIdRef, MaterialIdRef, TransformIdRef) and BoneBinding: conversions to domain IDs, validity, equality, implicit operators.

StateBusComponentTests

PlayMode tests for StateBusComponent (MonoBehaviour implementing IStateReader). Smoothing is left disabled (default) so Set/Get is synchronous and deterministic.

TargetMappingTests

Tests for UnityTransformTargetMapping and UnityMaterialTargetMapping: path-based resolution against real Transform hierarchies, validity, defaults, ToString.

TransformPositionSequenceLayerSOTests

Tests for TransformPositionSequenceLayerSO: early-return guards, effective-weight blending toward identity (Vector3.zero, see Weight's sibling convention on TransformPositionData.Weight), and the cycles-gated hold-last-step behavior it shares with TransformScaleSequenceLayerSO (no wrap-seam unwrap — unlike TransformRotationSequenceLayerSO, UnityEngine.Vector3.Lerp(UnityEngine.Vector3,UnityEngine.Vector3,System.Single) has no angular wraparound ambiguity). Drives the layer directly via its public 3-delta Evaluate override with a crafted UnityProceduralContext — the same seam TransformRotationSequenceLayerSOTests uses.

TransformRigApplierTests

PlayMode tests for TransformRigApplier. Drives real Unity Transforms via a runtime UnityRigDefinition and asserts the resulting localPosition / localRotation after applying engine-agnostic PoseDeltas.

TransformRotationSequenceLayerSOTests

Tests for TransformRotationSequenceLayerSO: early-return guards, effective-weight blending toward identity (0 degrees, see Weight's sibling convention on Weight), and the loop-seam "always increasing" wrap rule that lets a baked sequence read as one continuous spin instead of snapping backward at the wrap (see ResolveStepAngle's own remarks). Drives the layer directly via its public 3-delta Evaluate override with a crafted UnityProceduralContext — the same seam LayerAndActionSOTests/RuleSOTests use for their own deterministic layer coverage — rather than through reflection or an assembly-internal seam, since this assembly carries no InternalsVisibleTo friend for the ProceduralAnimator infrastructure assembly.

TransformScaleSequenceLayerSOTests

Tests for TransformScaleSequenceLayerSO's wrap-seam/clamped-end handling — mirrors TransformRotationSequenceLayerSOTests's coverage of the analogous rotation quirk: ResolveStepScale must not blend a non-looping last step into step 0's scale once time clamps at the end, the same "no next playthrough to blend into" rule ResolveStepAngle documents.

UnityActivationApplierTests

PlayMode tests for UnityActivationApplier. Registers real GameObjects and asserts activeSelf after applying ActivationDeltas.

UnityGraphicColorApplierTests

PlayMode tests for UnityGraphicColorApplier. Registers real GameObjects carrying UGUI UnityEngine.UI.Image components and asserts their live UnityEngine.UI.Graphic.color after applying UiColorDeltas — mirrors UnityTransformApplierTests's self-correcting-applier coverage shape, adapted for this applier's per-target collection-of-slots model instead of one component per target.