The garbage collector does scan the stack -- to see what things in the heap are currently being used (pointed to) by things on the stack.
It makes no sense for the garbage collector to consider collecting stack memory because the stack is not managed that way: Everything on the stack is considered to be "in use." And memory used by the stack is automatically reclaimed when you return from method calls. Memory management of stack space is so simple, cheap and easy that you wouldn't want garbage collection to be involved.
(There are systems, such as smalltalk, where stack frames are first-class objects stored in the heap and garbage collected like all other objects. But that's not the popular approach these days. Java's JVM and Microsoft's CLR use the hardware stack and contiguous memory.)
I am a little confused on both pointers and reference.
From my understanding pointers are addresses in memory.
In general, they are both addresses in memory. Usually one speaks of pointers if arithmetic can be done on them and of references if not, but there are some language-specific variations.
If I pass a variable using pointers and reference to a function, any
manipulations of the variable in the function will change the original
variable.
Yes.
But there are three reasons for passing something by reference:
- To avoid copying large amount of data to the stack.
- To modify the original variable.
- To support polymorphism, as the generated code does not know in advance how large the object will be and how to properly copy it.
Code using the second option is always more difficult to read than code using return values and if it is not used consistently, it is a sure way to create real mess. To distinguish this case from the other two C++ has the const modifier. Unfortunately Java and C# don't (and their authors don't seem to understand why it is useful; might have something to do with the fact that many people sell const in C/C++ as optimization tool which it certainly isn't).
I read online that passing by reference is not encouraged because it can get messy code easily.
High level languages hold everything by reference, because of the reason 3 above. They may pass the reference by value (just like passing pointer in C; the function can't change where the pointer points, but it can change content of the pointed object) or by reference (that is pointer-to-pointer; the function can make the variable point to different object), but it is always a reference to the actual object.
Java does that except for primitive numeric types and C# does it for everything that is class.
But passing by reference for the reason 2 above is indeed discouraged. One simply does not normally expect function to modify it's arguments, so better not do it.
I also read here (https://stackoverflow.com/questions/7058339/c-when-to-use-references-vs-pointers), that we should avoid pointers if we can.
That is C++-specific. References can't be set to nullptr (0 in older C++) so you don't have to check and references can't be deleted, so you don't have to think whether you should.
My question is when should pointers and reference be used?
Wherever they have to be.
I know pointers are used in Xcode for strong references and reference
counting. Since, higher level languages like c# have garbage
collector, does that mean we shouldn't ever use pointers and reference
in them?
Managed languages (C#, Java, ...) don't have pointers at all. Pointers allow access to raw memory and that would interfere with the managed runtime. So they don't have them (C# does, in the special, "unsave", extension, to allow interfacing with plain old C libraries).
But references can't be not used in them. Everything is a reference. Because allocated memory can only be handled by pointer/reference. And because of polymorphism. So all variables of object types are references in Java and C# and all variables are references in anything dynamically typed like smalltalk, perl, python, ruby, javascript etc.
Best Answer
I'm assuming you mean reference counted smart pointers and I'll note that they are a (rudimentary) form of garbage collection so I'll answer the question "what are the advantages of other forms of garbage collection over reference counted smart pointers" instead.
Accuracy. Reference counting alone leaks cycles so reference counted smart pointers will leak memory in general unless other techniques are added to catch cycles. Once those techniques are added, reference counting's benefit of simplicity has vanished. Also, note that scope-based reference counting and tracing GCs collect values at different times, sometimes reference counting collects earlier and sometimes tracing GCs collect earlier.
Throughput. Smart pointers are one of the least efficient forms of garbage collection, particularly in the context of multi-threaded applications when reference counts are bumped atomically. There are advanced reference counting techniques designed to alleviate this but tracing GCs are still the algorithm of choice in production environments.
Latency. Typical smart pointer implementations allow destructors to avalanche, resulting in unbounded pause times. Other forms of garbage collection are much more incremental and can even be real time, e.g. Baker's treadmill.