Develop Reference

Unity - persistence with inspector-added references

I have a persistent game object that I use to initialize basically everything. The script Persistence as I call it has some public references I just dragged on them via the inspector.
I'm trying to make it persistent like I found online:
public static Persistence instance;
void Awake()
{
if (instance != null)
{
DestroyImmediate(gameObject);
}
else
{
DontDestroyOnLoad(gameObject);
instance = this;
}
}
The thing is, when I load a level from main menu, it's fine. When I load back the main menu from that level, it says
MissingReferenceException: The object of type 'Persistence' has been destroyed but you are still trying to access it.
So I decided to let it create other instances when loading the main menu, but that messes up all the scripts on the game levels that rely on this data.
My question is, how to correctly implement this ~singleton persistent pattern in Unity, given I have inspector-added references?
NOTE that I initialize only from my Awake function in my persistent class and from nowhere else. Literally, my InitializeMe scripts are called from the Persistence class, one after the other.
What should I do differently to make this work? Initialize from a different, non-persistent gameobject? Forget dragging to the inspector? Any advice to make this work is appreciated.

I got it to work. So here's the thing one should know (I should have known) before messing around with persistence in Unity:
You don't drag references to a persistent GO via the inspector as they are gone the instant another scene is loaded except for additive loading. Also the persistent GO may hold data relevant to multiple scenes, hold functionality that is built for general purpose, as reusable as it goes, but it never initializes anything itself or interfere with non-persistent GOs other than being called - used as a tool - or providing the data.
This arcane wisdom is mine though, so anyone more skilled reading this, please do correct me if I'm wrong before others take this to heart.
So to be exact, I made a non-persistent master INITIALIZER that does the exact thing. I get all the scene GOs into a collection once and no GameObject.Find again. I used LINQ queries to conveniently filter my results from that collection. Another thing was my InitializeMe abstract class (or to be more precise, its descendants).
public abstract class InitializeMe : MonoBehaviour
{
public int orderNumber;
public abstract void INIT(INITIALIZER init);
}
here orderNumber is used to determine the order of the initializations, should one object depend on the other. It also worths mentioning that doing like so can result in a very predictable way to setting things up, as it is done one after the other. Note that this doesn't create a bottleneck, because Unity's scripting API only be executed in the main thread - dividing the code into multiple Awake or Start methods wouldn't perform better to my best knowledge as of 2017. It is used by a LINQ query in the INITIALIZER.
The good thing in INITIALIZER is that - again - it holds itself everything, including a reference to the persistent object and the save data as well, so via referencing itself to the InitializeMe methods they can do everything they need - as DataBank provides general purpose tools with persistent data, and INITIALIZER provides volatile (hope I use that right, I mean relevant to only main menu) data with some additional functionality that is used only for the main menu.
Now, persistence was ensured not by the persistent class itself, but by this INITIALIZER in Awake like so:
//find by tag or rename upon instantiation as by doing so will result in a "(Clone)" added to its name.
var d = GameObject.FindGameObjectWithTag("DataBank");
if (d == null)
{
//instantiating it
DATA = GameObject.Instantiate<DataBank>(DATA_blueprint);
//intializing the databank itself. The parameters are needed for me,
// but serve as a good example - all of those are dragged from the inspector to the _INITIALIZER_ but will remain in the whole game as long as databank exists.
DATA.LoadDataFromINIT(_baseColor, _baseColor2, _outlineColor,
new MMtoolBundle(DATA));
//to make it persistent.
DontDestroyOnLoad(DATA);
}else
{
//in this case the main menu is loaded from another scene.
//so we only find that loaded object and get its reference, no initialization as it was already setup.
this.DATA = d.GetComponent<DataBank>();
}
where DATA is my persistent DataBank. Note that by default I have NO DataBank object on the scene. DATA_blueprint is the prefab of DataBank dragged by the inspector (as INITIALIZER is not persistent). It could be loaded via AssetDatabase as well but this was a bit more convenient.
it worths mentioning that while DataBank is a MonoBehaviour itself so it can appear in Unity scenes, none of its members are MonoBehaviours so it is possible to harness the power of inheritance. Should one wish to start a coroutine in a non-mono toolkit, it would have to be started with a reference from DataBank itself.
An obvious drawback of this method is that the existence of my persistent GO depends on INITIALIZER, so that one has to be on the first scene. But again, as most games start with a main menu, this shouldn't pose a big issue (talking about single player games).
YET AGAIN I would recommend someone more skilled to correct me in case I lead others astray, but this solution worked for me - got the same and only persistent GO in all my scenes with no other ever created.

Passing more parameters to Action Efficient?

I have an Action callback called Action<GameObject,GameState,GameObject,Vector3> onCutCB;
I have 5 Scripts that references a function called
void OnCut(GameObject survivalObj,GameState state,GameObject
destroyObj,Vector3 startPos){
////
}
but the problem is that out of these 5 scripts ,first scripts need 3 parameter,second script need all parameters ,rest of the scripts need one parameter
So the question is
1) should i use callback or call the functions manually from by using the reference of the respective classes
2) As i am passing 4 parameters out of which some parameters are useful in some scripts will there be more garbage collection ? than calling the scripts method manually ?
Out of the four parameters i passed 2 are reference type and 2 are value types.

1) should i use callback or call the functions manually from by using
the reference of the respective classes
Calling functions directly is faster but this shouldn't be a problem at all. It doesn't make that much difference.
2) As i am passing 4 parameters out of which some parameters are
useful in some scripts will there be more garbage collection ? than
calling the scripts method manually ? Out of the four parameters i
passed 2 are reference type and 2 are value types.
No, there will be no memory allocation when you pass more parameter to a delegate. The only time memory will be allocated is when you initialize the onCutCB variable. For example, onCutCB = OnCut;. That's why onCutCB must be initialized in the Start or Awake function once. When onCutCB in invoked, memory allocation does not happen.
Read the comments in the code below to see where memory is allocation or not:
Action<GameObject, GameState, GameObject, Vector3> onCutCB;
void OnCut(GameObject survivalObj, GameState state, GameObject
destroyObj, Vector3 startPos = default(Vector3))
{
}
void Start()
{
//MEMORY IS ALLOCATION
onCutCB = OnCut;
}
void Update()
{
//NO MEMORY ALLOCATION HERE
onCutCB(objToUse, gameState, anotherObj, thePos);
}
though when i was playing my game on mobile devices til to minitues it
works fine but after that some times its lags a bit I am using object
pooling also when i restart the game my scene doesnt load i just use
pooling because the gameobject which were instantiated were required.
This problem you described has nothing to do with the gradual slow-down of your game. You are either instantiating many Objects non-stop, starting a coroutine every frame from the Update function and coroutine does not end/stop or something that adds up to the execution time.
Without your all scripts, it is impossible to tell what's causing your problem. Please, do not edit your question with all your code. Find and fix the problem yourself.
Use the Profiler!
Go to Windows --> Profiler then enable Deep Profiler and click on "Record" and clicl the Play button. You will see which script is allocating memory and how much time it takes for each function to complete. Look into these and fix it.
Here is a full Unity tutorial on how to use the Profiler.

Access static class directly or cache it?

I know that it's better to cache a component if you use it often during runtime. How about if I cache a component as a static variable and access it during runtime? Let me give you an example:
public class AudioManager : MonoBehaviour
{
public static AudioManager audioManager;
void Awake()
{
audioManager = this;
}
}
Then if I access it from another class, should I use AudioManager.audioManager or cache it beforehand (speaking in the terms of performance)?
P.S. Let me know if something isn't clear.
Edit: I guess I don't know what the hell I'm doing regarding Singletons, but to make it clear, this is a newbie question regarding performance:
1) _audioManager = AudioManager.audioManager in Awake and then use _audioManager.someVariable in Update
VS
2) AudioManager.audioManager.someVariable in Update

You don't need to cache your own component for speed in any case. A component won't pay a performance penalty for accessing itself.
The use case you may be thinking of is caching access to other components or objects so that you don't have to do lookup with GetComponent or GetComponentWithChildren. That does save time, but it's not a complex pattern: just do the lookup the first time you encounter the other component or gameobject, store it in a field inside this component, and you're good:
class Example: MonoBehavior
{
public GameObject _TargetCharacter;
private Transform _cachedTransform;
private Animator _cachedAnimator;
void Awake()
{
if (_TargetCharacter != null)
{
_cachedTransform = _TargetCharacter.transform;
_cachedAnimator = _TargetCharacter.GetComponent<Animator>();
}
}
void Update()
{
DoSomething (_cachedTransform, _cachedAnimator);
}
}
GetComponent and GetComponentInChildren are comparatively expensive you do don't want to do them every frame. However direct access to something set as a field is not expensive there is no need to cache it.
And making it s Singleton: in this context, you would need to make a separate class which is NOT a Component and have that be the singleton - you can't prevent somebody from attaching lots of a given MonoBehavior to things in the Unity Editor so you don't want to assume there's only one of the component floating around. In the AudioManager example, you'd have to assume that many different components might all be looking at the same AudioManager.
Here a good example comparing shared behavior using Singletons vs shared behavior in a static class. Two means to the same end, with slightly different strengths and weaknesses.
Update In response to OP's comment below:
If the point of the original code is to gain speed using the common Unity caching trick, it's not needed. If the point is, as OP suggests in comment below, to make the manager into a 'service provider' that other classes can call there are two routes to go:
The standard unity trick would be to add the caching behavior to the other classes that need to get to the AudioManager. For example:
class SomeOtherComponent: MonoBehavior
{
AudioManager _manager;
void Awake()
{
AudioManager[] AllAudioManagers = GetComponents<AudioManager>();
if (AllAudioManagers.Count == 1)
{
_manager = AllAudioManagers[0];
}
else
{
throw new RuntimeError ("Expecting one (and only one) AudioManager in scene");
}
}
}
Switching to a proper Singleton implementation more or less amounts to just taking the boilerplate in the Awake method above and centralizing into the singleton class, allowing you to replace it.
For an extended discussion of the singleton pattern in unity try this book

I live by a simple rule, if I am going to be calling a component a lot of times, more than 5, then I cache it. The same rule I apply to Singletons. If it is only one call then I just use the Singleton.getInstance().
What you should be really caching
Unity Components such as Transform.
The logic behind caching things like Transform comes from looking at the documentation. Transform is not a variable but a property, which happens to contain getters and setters. These are hidden from you, but that's the way it is implemented. This property happens to be written in C or C++. Therefore this will cost you on performance.

c# Should I store a copy of a field for convenience?

I have developed a habit of sometimes doing this particular thing and I'm wondering why am I doing it, is there any advantage?
Heres an example from a Unity3d game..
In my class I want to do various calculations and so forth with a float ThingYposition which is a field stored somewhere in Thing.transform.position.y. Rather than be writing Thing.transform.position.y so many times I just make a copy of the float I want at the beginning of the program.
public GameObject Thing;
private float ThingYposition;
public Start()
{
ThingYposition = Thing.transform.position.y
}
public Update()
{
//Do stuff every frame with ThingYposition
}
So this way means my lines of code will be a little less cluttered but the program will use a little bit more memory as I'm storing that float twice now. But will it be any faster? Does accessing a deeply embedded field like Thing.transform.position.y actually use any more processing power than accessing my float field?
Do you think this is harmless habit or should I stop?
Also please note in this example I dont care if the original changes at all and I dont want to change it.

You already stated you don't care if the original changes, so I'll skip that part. The only advantage I can see is in a multi-threaded environment. You don't have to worry about another thread mucking with Thing, since you have a private copy of ThingYposition.
In terms of efficiency, you're well into micro optimizing here. If you're having a problem, profile it and experiment with alternatives. But I can't imagine this is something you really need to worry about.

Since you don't care whether or not the original position changes and will not change it yourself, then this is probably the best approach for the use-case you described.
The answer to the other part of your question, is it faster to access a local vs a "deeply embedded field" depends on how Thing.transform.position.y is implemented. If it just a member field, then the access times would be essentially the same for a local copy or the "deeply embedded field". If Thing.transform.position.y is calculated on every access then the local copy would be faster.

Pacman Game Organization (XNA) - Where/how should I check for collisions with ghosts/food?

I have my Pacman game working, and i'm trying to go back through it and "organize" my code better. Before, I had this "TopObject" which referenced the object in game1.cs which basically referenced every object in my game (pacman, ghosts, map, etc). I made it globally available from ANYWHERE so that I could access my Pacman object from my ghost objects, or my map/tile objects from my ghost objects, or even my ContentMananger from anywhere to load content.
I know this is bad practice, so i'm trying to eliminate it. For instance, I created LoadContent() methods on all my objects which take a "ContentManager" type, which eliminated the need for me to call ContentManager using my TopObject.
Proceeding further, i'm having a really hard time doing stuff without this global reference to all my objects. For instance:
- From within my Pacman Update()..
I need to know if i'm going to run into a wall. I need references to my map/tiles.
I need to know if i've collided with a ghost, so i need references to my ghost objects.
I need to know if I collided with some food and then update the scoreboard, so i need have reference to my scoreboard object.
So essentially, it feels like i'm going to be making a giant mess passing so many object references around to all my objects, that it feels like doing a "global" topobject is much cleaner.
Does anyone know how to organize this better, so that i'm doing it cleanly? Should I even be checking for collisions inside my Pacman Update(), or should I be creating seperate "PacmanObject.CheckCollisionWith()" and calling that from my Game1 Update()? Should collision logic also be seperated out into a new class?
Thanks!

Having a central point that contain reference toward your objects isn't a bad thing per se. However, a good design will put the actions that are best related to each object in the correct class. For example, your player doesn't need to have reference towards the tiles and walls. While doing an Move method, the player could call a static method in the Map class that return true/false if the next move is valid. This way, the player doesn't have any reference toward the map itself.
public class Player
{
public void Update()
{
//stuff to find nextTile position
//Call static function toward a class that contains the map data.
if (Map.TileIsWalkable(nextTile))
Move();
}
}
public class Map
{
public static Map loadedMap; // Handling it as singleton, normally you only load one map at a time.
public static bool TileIsWalkable(Vector2 position)
{
if (loadedMap == null) // Throw assert, shouldn't happen.
return //whatever test needed in loadedMap structure
}
}
You can also have a Ghost manager that keeps reference toward the ghost. The player would only need to call a method in that manager that would loop over all the ghost to see if he collides. Even better, that method could take a position and not a reference. That way, it can be reused to see if ghosts collide between each other. (random idea)
There's plenty of way to do something. The hardest in any code design is to put the stuff at the place that make the more sense and makes it the easiest to update and modify later.