If a CPU's stack is 1MB and a programming language is purely pass by value, how can we pass data bigger than 1MB into functions? What happens exactly?

Question

So I hear that using pass by value, copies of the parameters are added to the call stack. Apparently, the stack size on Windows is often 1MB. Obviously though, we can easily pass around data that is far bigger than 1MB between functions/procedures (arrays of primitives/classes/hashmaps/sets or whatever)

So is my understanding of the situation wrong...or are these languages using pass by value for primitive types, but then pass by reference/pass by object model for these other data structures?

Just with a quick google, both Java & JavaScript are exclusively pass by value...so how are you able to pass around data/objects bigger than the stack size in these languages? For example, in JavaScript, it's stated: "For Array the maximum length is 4GB-1 (2^32-1)"

Answer 1

The difference between "pass-by-reference" and "pass-by-value" is that, in pass-by-reference calling convention, the function can change what the reference points to. In languages that don't support a pass-by-reference calling convention, you won't be able to do this directly; you'd have to simulate it by passing a mutable type that contains a reference to another object.

In pass-by-value, you pass references as values too. So the thing that's allocated on the stack, when you're passing an object that isn't a primitive value, is a copy of the pointer, not a deep copy of the data.

The aspect of this that gets confusing is the thing (or instance) of what you point to, if it's a mutable type, can be mutated.