unreal-engine-cpp-pro（unreal 引擎 CPP 专业版）

Unreal Engine C++ Pro

This skill provides expert-level guidelines for developing with Unreal Engine 5 using C++. It focuses on writing robust, performant, and standard-compliant code.

When to Use

Use this skill when:

• Developing C++ code for Unreal Engine 5.x projects
• Writing Actors, Components, or UObject-derived classes
• Optimizing performance-critical code in Unreal Engine
• Debugging memory leaks or garbage collection issues
• Implementing Blueprint-exposed functionality
• Following Epic Games' coding standards and conventions
• Working with Unreal's reflection system (UCLASS, USTRUCT, UFUNCTION)
• Managing asset loading and soft references

Do not use this skill when:

• Working with Blueprint-only projects (no C++ code)
• Developing for Unreal Engine versions prior to 5.x
• Working on non-Unreal game engines
• The task is unrelated to Unreal Engine development

Core Principles

1. UObject & Garbage Collection:

- Always use UPROPERTY() for UObject* member variables to ensure they are tracked by the Garbage Collector (GC). - Use TStrongObjectPtr<> if you need to keep a root reference outside of a UObject graph, but prefer addToRoot() generally. - Understand the IsValid() check vs nullptr. IsValid() handles pending kill state safely.

1. Unreal Reflection System:

- Use UCLASS(), USTRUCT(), UENUM(), UFUNCTION() to expose types to the reflection system and Blueprints. - Minimize BlueprintReadWrite when possible; prefer BlueprintReadOnly for state that shouldn't be trampled by logic in UI/Level BPs.

1. Performance First:

- Tick: Disable Ticking (bCanEverTick = false) by default. Only enable it if absolutely necessary. Prefer timers (GetWorldTimerManager()) or event-driven logic. - Casting: Avoid Cast<T>() in hot loops. Cache references in BeginPlay. - Structs vs Classes: Use F structs for data-heavy, non-UObject types to reduce overhead.

Naming Conventions (Strict)

Follow Epic Games' coding standard:

• Templates: Prefix with T (e.g., TArray, TMap).
• UObject: Prefix with U (e.g., UCharacterMovementComponent).
• AActor: Prefix with A (e.g., AMyGameMode).
• SWidget: Prefix with S (Slate widgets).
• Structs: Prefix with F (e.g., FVector).
• Enums: Prefix with E (e.g., EWeaponState).
• Interfaces: Prefix with I (e.g., IInteractable).
• Booleans: Prefix with b (e.g., bIsDead).

Common Patterns

1. Robust Component Lookup

Avoid GetComponentByClass in Tick. Do it in PostInitializeComponents or BeginPlay.

void AMyCharacter::PostInitializeComponents() {
    Super::PostInitializeComponents();
    HealthComp = FindComponentByClass<UHealthComponent>();
    check(HealthComp); // Fail hard in dev if missing
}

2. Interface Implementation

Use interfaces to decouple systems (e.g., Interaction system).

// Interface call check
if (TargetActor->Implements<UInteractable>()) {
    IInteractable::Execute_OnInteract(TargetActor, this);
}

3. Async Loading (Soft References)

Avoid hard references (UPROPERTY(EditDefaultsOnly) TSubclassOf<AActor>) for massive assets which force load orders. Use TSoftClassPtr or TSoftObjectPtr.

UPROPERTY(EditAnywhere, BlueprintReadWrite)
TSoftClassPtr<AWeapon> WeaponClassToLoad;

void AMyCharacter::Equip() {
    if (WeaponClassToLoad.IsPending()) {
        WeaponClassToLoad.LoadSynchronous(); // Or use StreamableManager for async
    }
}

Debugging

• Logging: Use UE_LOG with custom categories. DEFINE_LOG_CATEGORY_STATIC(LogMyGame, Log, All); UE_LOG(LogMyGame, Warning, TEXT("Health is low: %f"), CurrentHealth);
• Screen Messages: if (GEngine) GEngine->AddOnScreenDebugMessage(-1, 5.f, FColor::Red, TEXT("Died!"));
• Visual Logger: extremely useful for AI debugging. Implement IVisualLoggerDebugSnapshotInterface.

Checklist before PR

• Does this Actor need to Tick? Can it be a Timer?
• Are all UObject* members wrapped in UPROPERTY?
• Are hard references (TSubclassOf) causing load chains? Can they be Soft Ptrs?
• Did you clean up verified delegates in EndPlay?

Limitations

• Use this skill only when the task clearly matches the scope described above.
• Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
• Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.