#include <ComposableCameraTransitionBase.h>

Per-frame pose snapshot captured for debug visualization.

Deliberately narrower than [FComposableCameraPose](FComposableCameraPose.md#fcomposablecamerapose) — we only keep the fields DrawStandardTransitionDebug actually reads (position, rotation, resolved FOV in degrees). Skipping FPostProcessSettings matters: that struct embeds TObjectPtr<UTexture> / similar UObject references through its color-grading / vignette / bloom sub-structs, and our cache is NOT a UPROPERTY, so those references wouldn't be tracked by the GC. Caching only POD-like fields sidesteps the issue entirely AND shrinks each transition's per-frame debug memory from ~3× sizeof(FPostProcessSettings) to a few dozen bytes.

Public Attributes¶

Position¶

FVector Position { FVector::ZeroVector }

Rotation¶

FRotator Rotation { FRotator::ZeroRotator }

FOVDegrees¶

float FOVDegrees { 90.f }

FTransitionDebugSnapshot¶

#include <ComposableCameraTransitionBase.h>

Per-frame pose snapshot captured for debug visualization.

Deliberately narrower than [FComposableCameraPose](FComposableCameraPose.md#fcomposablecamerapose) — we only keep the fields DrawStandardTransitionDebug actually reads (position, rotation, resolved FOV in degrees). Skipping FPostProcessSettings matters: that struct embeds TObjectPtr<UTexture> / similar UObject references through its color-grading / vignette / bloom sub-structs, and our cache is NOT a UPROPERTY, so those references wouldn't be tracked by the GC. Caching only POD-like fields sidesteps the issue entirely AND shrinks each transition's per-frame debug memory from ~3× sizeof(FPostProcessSettings) to a few dozen bytes.

Public Attributes¶

Position¶

FVector Position { FVector::ZeroVector }

Rotation¶

FRotator Rotation { FRotator::ZeroRotator }

FOVDegrees¶

float FOVDegrees { 90.f }