My game base consists of a series of modules, organized as classes, that are created, updated and interact when needed.
A few examples could be: CWindowManager, CGraphicsManager, CPhysicsManager, and so on.
I'm ashamed to have to say that I currently use global pointers for them (extern CWindowManager* g_WindowManager;), and I know that this is probably a bad thing to do.
In any case, the thing is that these modules need to be created and deleted dynamically, and that in the right order of course. Theres also the problem that modules like CPhysicsManager are scene-dependent, therefore they are deleted when the scene is switched and then created again.
Now, I'd like to switch away from using globals for working with modules in my game.
I'm not afraid of the refactoring, but I can't really think of what would be the best alternative to globals.
I have thought about creating a CModuleManager class and storing instances of the modules in there as members, which are then derived from a CModule base class. Although I can't really think of how this would work in detail.
This seems like a common problem in software development and especially game development, so:
- What is the best option for managing modules, compared to simply using global pointers?
Some things to consider when using global data:
multi-threading. You need to be careful if the global data can be accessed by different threads concurrently.
testability. If you are writing unit tests, then the global data needs to be initialized before any code accesses the global data.
An alternative to using global data is to pass the object instances that each class/method requires as a parameter. This has the benefit that all the dependencies of a class are visibly from the API. It also makes writing unit tests much easier. The disadvantage is that the code can get messy if you are passing objects all over the place.
Your idea of having a ModuleManager sounds like the Service Locator pattern - which I think is a feasible solution. This link may help: http://gameprogrammingpatterns.com/service-locator.html.
If you choose to carry on using global pointers, then dynamically deleting the global data can be achieved in C++ using smart pointers (auto_ptr).
Related
I hate to beat a dead horse, that said, I've gone over so many conflicting articles over the past few days in regards to the use of the singleton pattern.
This question isn't be about which is the better choice in general, rather what makes sense for my use case.
The pet project I'm working on is a game. Some of the code that I'm currently working on, I'm leaning towards using a singleton pattern.
The use cases are as follows:
a globally accessible logger.
an OpenGL rendering manager.
file system access.
network access.
etc.
Now for clarification, more than a couple of the above require shared state between accesses. For instance, the logger is wrapping a logging library and requires a pointer to the output log, the network requires an established open connection, etc.
Now from what I can tell it's more suggested that singletons be avoided, so lets look at how we may do that. A lot of the articles simply say to create the instance at the top and pass it down as a parameter to anywhere that is needed. While I agree that this is technically doable, my question then becomes, how does one manage the potentially massive number of parameters? Well what comes to mind is wrapping the different instances in a sort of "context" object and passing that, then doing something like context->log("Hello World"). Now sure that isn't to bad, but what if you have a sort of framework like so:
game_loop(ctx)
->update_entities(ctx)
->on_preupdate(ctx)
->run_something(ctx)
->only use ctx->log() in some freak edge case in this function.
->on_update(ctx)
->whatever(ctx)
->ctx->networksend(stuff)
->update_physics(ctx)
->ctx->networksend(stuff)
//maybe ctx never uses log here.
You get the point... in some areas, some aspects of the "ctx" aren't ever used but you're still stuck passing it literally everywhere in case you may want to debug something down the line using logger, or maybe later in development, you actually want networking or whatever in that section of code.
I feel like the above example would much rather be suited to a globally accessible singleton, but I must admit, I'm coming from a C#/Java/JS background which may color my view. I want to adopt the mindset/best practices of a C++ programmer, yet like I said, I can't seem to find a straight answer. I also noticed that the articles that suggest just passing the "singleton" as a parameter only give very simplistic use cases that anyone would agree a parameter would be the better way to go.
In this game example, you probably wan't to access logging everywhere even if you don't plan on using it immediately. File system stuff may be all over but until you build out the project, it's really hard to say when/where it will be most useful.
So do I:
Stick with using singletons for these use cases regardless of how "evil/bad" people say it is.
Wrap everything in a context object, and pass it literally everywhere. (seems kinda gross IMO, but if that's the "more accepted/better" way of doing it, so be it.)
Something completely else. (Really lost as to what that might be.)
If option 1, from a performance standpoint, should I switch to using namespace functions, and hiding the "private" variables / functions in anonymous namespaces like most people do in C? (I'm guessing there will be a small boost in performance, but then I'll be stuck having to call an "init" and "destroy" method on a few of these rather than being able to just allow the constructor/destructor to do that for me, still might be worth while?)
Now I realize this may be a bit opinion based, but I'm hoping I can still get a relatively good answer when a more complicated/nested code base is in question.
Edit:
After much more deliberation I've decided to use the "Service Locator" pattern instead. To prevent a global/singleton of the Service Locator I'm making anything that may use the services inherit from a abstract base class that requires the Service Locator be passed when constructed.
I haven't implemented everything yet so I'm still unsure if I'll run into any problems with this approach, and would still love feedback on if this is a reasonable alternative to the singleton / global scope dilemma.
I had read that Service Locator is also somewhat of an anti-pattern, that said, many of the example I found implemented it with statics and/or as a singleton, perhaps using it as I've described removes the aspects that cause it to be an anti-pattern?
Whenever you think you want to use a Singleton, ask yourself the following question: Why is it that it must be ensured at all cost that there never exists more than one instance of this class at any point in time? Because the whole point of the Singleton pattern is to make sure that there can never be more than one instance of the Singleton. That's what the term "singleton" is all about: there only being one. That's why it's called the Singleton pattern. That's why the pattern calls for the constructor to be private. The point of the Singleton pattern is not and never was to give you a globally-accessible instance of something. The fact that there is a global access point to the sole instance is just a consequence of the Singleton pattern. It is not the objective the Singleton pattern is meant to achieve. If all you want is a globally accessible instance of something, then use a global variable. That's exactly what global variables are for…
The Singleton pattern is probably the one design pattern that's singularly more often misunderstood than not. Is it an intrinsic aspect of the very concept of a network connection that there can only ever be one network connection at a time, and the world would come to an end if that constraint was ever to be violated? If the answer is no, then there is no justification for a network connection to ever be modeled as a Singleton. But don't take my word for it, convince yourself by checking out page 127 of Design Patterns: Elements of Reusable Object-Oriented Software where the Singleton pattern was originally described…😉
Concerning your example: If you're ending up having to pass a massive number of parameters into some place then that first and foremost tells you one thing: there are too many responsibilities in that place. This fact is not changed by the use of Singletons. The use of Singletons simply obfuscates this fact because you're not forced to pass all stuff in through one door in the form of parameters but rather just access whatever you want directly all over the place. But you're still accessing these things. So the dependencies of your piece of code are the same. These dependencies are just not expressed explicitly anymore at some interface level but creep around in the mists. And you never know upfront what stuff a certain piece of code depends on until the moment your build breaks after trying to take away one thing that something else happened to depend upon. Note that this issue is not specific to the Singleton pattern. This is a concern with any kind of global entity in general…
So rather than ask the question of how to best pass a massive number of parameters, you should ask the question of why the hell does this one piece of code need access to that many things? For example, do you really need to explicitly pass the network connection to the game loop? Should the game loop not maybe just know the physics world object and that physics world object is at the moment of creation given some object that handles the network communication. And that object in turn is upon initialization told the network connection it is supposed to use? The log could just be a global variable (or is there really anything about the very idea of a log itself that prohibits there ever being more than one log?). Or maybe it would actually make sense for each thread to have its own log (could be a thread-local variable) so that you get a log from each thread in the order of the control flow that thread happened to take rather than some (at best) interleaved mess that would be the output from multiple threads for which you'd probably want to write some tool so that you'd at least have some hope of making sense of it at all…
Concerning performance, consider that, in a game, you'll typically have some parent objects that each manage collections of small child objects. Performance-critical stuff would generally be happening in places where something has to be done to all child objects in such a collection. The relative overhead of first getting to the parent object itself should generally be negligible…
PS: You might wanna have a look at the Entity Component System pattern…
I'm building a plugin which will be used in a host. This plugin is using a singleton for services I would like to easily access anywhere. The problem comes when I instance several times the same plugin, the same (static) singleton, being specific to the runnable, will be shard among all the instanced plugins. Is there, generally speaking, a way to reduce the scope of the singleton (c++) ?
As each plugin is an instance in itself, I could obviously pass the root class of the plugin to all of it's subclasses but I would like to keep the same global singleton design as possible.
Is there a reason for having a singleton? The rationale is when you need to enforce that there is only one, and need to provide a single point of access. If these aren't really requirements, then just create one and pass it around where needed.
I would gradually get rid of the singleton.
Does the singleton do a lot, or not much?
You might need to divide it up into parts.
If it doesn't do much, just pass it where is is needed, and get rid of its singleton-ness.
If it provides lots of services, create interfaces for each service and pass those around where they are needed. Your design will improve and become more testable and easier to comprehend.
At first, the implementations of the interfaces could delegate to the original singleton, but you want to make them self contained eventually.
A singleton do internally make use of a static variable.
The scope of this static variable is specified by the source file where it is defined and partitioned by its current runnable. For those reason, while running under the same host (and then the same runnable) both plugins (which are the same code) do share the same static variable (and by extension the same singleton).
As we assume in this question the code to be the same for each plugin, the only way to split those singletons would then be to run a new executable. This could be done using the fork unix command for example where both process will then hold their own memory range.
Obviously (as most of you commented) it is a much better approach to avoid using singletons in this case as forking a process is just adding useless complexity.
When you have a set of functions that have no interaction between them, you place them in a namespace. (Example, a math namespace.)
When you have some public attributes and optionally a set of functions that act on those attributes, that should become a class.
But what about when you have a set of related functions but no public attributes? An example would be an event manager: you might only have subscribe(), post(), and dispatch() and no public attributes; however you do have hidden attributes like a list of subscribers and an event queue that the three functions act upon. Should this be a class or a namespace?
Any time you have behavior and state it should be a class, even if the state isn't publicly accessible. One practical reason for this is it makes it easier to unit test other modules that interact with the module in question.
A class. You have a state, albeit internal, so you may want to create two event managers (two queues).
This should definitely be class, because it has internal state. What if you need more than one instance? In that case namespace can't help you, and class can.
If you want to create one of more instances of the thing that each have a lifetime, then it should be a class.
An event manager sounds like something you want to create, use and then destroy. So, it should be a class.
namespaces were only really introduced to help with very large programs. In a large program you may have multiple teams of developers each writing code. It may not be feasible for each team to make sure that nobody else uses happens to give a function or class the same name as somebody else. Or, names of things may clash with 3rd party libraries that might want to be used. Namespaces help avoid these problems. If you're not working on a huge project, you probably don't really need to bother ever using namespaces for your own code, unless you feel the need to organize your code into a few namespaces just to keep it neat. Using unnecessarily small namespaces can make code painful to work with. There's not much point in over obsessing about whether code should be in this namespace or that one, it's better to focus on making sure the code actually works.
I'm working on a lil' game engine in C++, and decided to do it all OOPily (heavy use of classes.)
It's intended to be (theoretically) cross-platform, so I have an 'Engine' class, an instance of which is created by the 'OS Module', which is WinMain for Windows (the platform I'm developing it for first.)
I have three main questions:
Is it considered poor practice to create a class that is only going to be instantiated once in the entire application? Perhaps because there is some kind of performance hit or added overhead incurred by using a class rather than a bunch of functions?
I've been planning to have WinMain create the instance of Engine as a local variable. The Engine class will be fairly large, containing classes for rendering, script parsing, file system stuff, etc. Basically, the whole game engine, apart from OS specific code, will be contained in the Engine class in some form (possibly as an instance of another class.) Is creating a local instance of my very large Engine class within the WinMain function a bad idea? Is creating a local instance a bad idea when the class will be created when the program starts, and end when the program ends? Maybe new would be better?
My plan (i/wa)s to divide the engine up into 'modules', each of which is represented by a class. The Engine class would contain an instance of almost all the other modules, like, as mentioned above, rendering, file system interaction, etc. Is using classes as containers for huge modules a bad idea from some perspective (performance, design, readability?)
Thanks for any help :)
Game engines are not prime candidates for cross-platform'ness since they usually involve efficient interaction with low-level API's (which are not cross-platofrm).
The size of a class depends on the member variables it contains, not the number of functions it implements.
Stack space is usually small (http://msdn.microsoft.com/en-us/library/ms686774%28v=vs.85%29.aspx) while the heap is theoretically as big as available RAM. So, if you have something really big, store it on the heap (with new).
The "Splash Screen": don't do all the work at the beginning of the program. Users hate it when they run the program and nothing shows on screen because your program is busy initializing something...
Lookup lazy instantiation, basically don't do things that can wait and always show something on screen.
As for specific answers:
1. No, assuming no virtual functions, there should be no performance overhead.
2. See "Splash Screen" and "limited stack space" above.
3. Modularity is generally good, just make sure each class does/represent a single "thing". But don't have so many classes you start forgetting their names and purpose :)
Is it considered poor practice to
create a class that is only going to
be instantiated once in the entire
application? Perhaps because there is
some kind of performance hit or added
overhead incurred by using a class
rather than a bunch of functions?
Not at all. This is basically because your Application class can do things like inherit and encapsulate. There's no performance hit or added overhead.
I've been planning to have WinMain
create the instance of Engine as a
local variable. The Engine class will
be fairly large, containing classes
for rendering, script parsing, file
system stuff, etc. Basically, the
whole game engine, apart from OS
specific code, will be contained in
the Engine class in some form
(possibly as an instance of another
class.) Is creating a local instance
of my very large Engine class within
the WinMain function a bad idea? Is
creating a local instance a bad idea
when the class will be created when
the program starts, and end when the
program ends? Maybe new would be
better?
Nope- this is pretty much the way it should go. Why bother heap allocating? You want automatic destruction semantics. Unless your class has a very large size, hundreds of KB or more, in which case an RAII-based heap allocation is smarter, as stack memory is quite limited. That is, a physical per-instance static size as reported by sizeof(), not including dynamic allocations.
My plan (i/wa)s to divide the engine
up into 'modules', each of which is
represented by a class. The Engine
class would contain an instance of
almost all the other modules, like, as
mentioned above, rendering, file
system interaction, etc. Is using
classes as containers for huge modules
a bad idea from some perspective
(performance, design, readability?)
This is exactly what object-orientated design is for- encapsulation, and dividing the program up into clearly defined and separated modules, then instantiating each one that you need, is exactly the idea behind object-orientated programming. The size of the concept that a class encapsulates is usually considered to be irrelevant, as long as it's one single concept. A clear modular design is great.
I would recommend using run-time inheritance for all of the platform-dependent ones, and preferably, loading them dynamically at run-time (using the OS class to perform the dynamic loading). This enforces a solid compile-time abstraction and allows for more efficient compilation.
No, doing whatever makes your design cleaner and more maintainable is recommended.
No, direct use of the freestore should be avoided whenever possible (i.e., don't use new unless you absolutely have to)
See #1
No it isn't considered bad style. In fact I know no application frameworks that go without one, and the renowned Singleton pattern is basically around the same theme (but different, mind you)
I can't imagine that your class is actually that big. If it is, make it on the heap. However, chances are that the actual contents of that class are going to be on the heap anyway (I'm sort of assuming you'll use existing container classes, that wil 99 out of 100 do dynamic allocation; this goes for STL containers as well)
What difference would it make whether you put them in the data segment, as auto's on the stack or members of a class? I think the class emphesizes the modularity and might enable you to do things more easily (like e.g. unit testing, or reusing the engine for a network-only version etc. etc.)
List down the various global states like global variable, singleton object etc.,
I read that they affect the testability of code. Can you explain alternatives to global states that help in improving the testability of the code?
Global variables and singletons are so similar from the testability viewpoint that they don’t have to be treated separately. Good OO design should be like bricks of Lego that fit together to create the resulting system. The main problem with globals is that you lose isolation. When you access your collaborators using globals or singletons, the classes are no longer separate bricks that you can easily take and reorganize, use in isolation or reuse in another project. When you take some class that depends on globals, you drag the rest of the design with it. In testing this makes problems because you have to take extra steps to cut the connections from your class to its collaborators, so that you can test it in isolation. (I’ve written more about this on my blog.)
A nice solution to the problem is supplying the class dependencies from the outside. You simply introduce some kind of mechanism (maybe a simple class) that will take all the isolated bricks and connect them into the object graph that makes up your application. This is what Inversion of Control and Dependency Injection is all about.
Start by reading singletons are pathological liars.
There are no various global states. Either something is global state or it isn't. The Singleton anti-pattern just removes various initialization headaches, but it's just the same as a global variable.
How to avoid global state? Make some local state and pass a reference or pointer to it around. In the vast, vast majority of occurrences, global state is needless. There are some instances of necessary global state, like the heap, but they're very rare and if you need to ask, then you almost certainly don't know enough to know when it's correct.