了解作用域数组的存储方式

Question

I'm confused how arrays are stored in the executable when they're within functions and the like.

With the code below I believe space for the value of the three ints to be stored is made in the executable.

#include <stdlib.h>

int main() {
    int arr[] = {rand(), rand(), rand()};
}

I was thinking for an array in a function, each call would use the same space as the other calls to store their array. But then I thought recursive calls would overwrite each others arrays. I don't get how space for them is left aside, especially with recursion and an unknown number of calls are made. I tried to make an example:

#include <stdio.h>
#include <stdlib.h>
#include <stdbool.h>

bool myfunc(unsigned val) {
    int arr[] = {rand() % 10, rand() % 10};

    if (val == 0)
        return true;

    myfunc(val - 1);

    printf("%d %d\n", arr[0], arr[1]);
    return false;
}

int main() {
    myfunc(rand() % 50);
}

I get arrays elements are stored one after the other. But how is there space for them (in the stack?) when it's unkown how many there will be?

Answer 1

Objects defined with local scope in a function body with automatic storage (without a static keyword) are stored in temporary storage and reclaimed automatically upon exiting the corresponding scope.

Typical modern architectures use the runtime stack for this: the stack pointer register is decremented upon entering the scope by a fixed amount if the size is known at compile time, as is the case for your examples: int arr[] = {rand() % 10, rand() % 10}; defines an array of 2 int , usually 8 bytes.

If the size is only known at runtime, the stack pointer is decremented by a value computed at runtime, possibly causing undefined behavior if this size is too large. The stack pointer is restored to its original value upon leaving the function.