[parser] MemoryCopy() calls: Prevent buffer overflow by replacing har…

[avx0]

…d-coded arguments

In future, if a dev edits the second arg and miscalulates the corresponding 3rd arg, there will be a buffer overflow or the string (2nd arg) will be cut short. This commit prevents that.

[avx0]

MemoryCopy() does not do input validation on its args.

1. Is count == size(srcPtr)?
   I wrote some code to set count equal to size(srcPtr), ignoring the user's input of count (this prevents out-of-bounds write / buffer overflow if user inputs count > size(srcPtr). I don't know the reliability of my size() function, please test it and suggest modifications.

// my edit -- begin
int size(char *a)
{
     int sz = 0;
     if (sizeof(a) != 8) return sizeof(a);
     while (a[sz] != '\0') sz++;
     return sz;
}
// my edit -- end

static void MemoryCopy(void *dest, const void *src, unsigned int count)
{
     char *srcPtr = (char *)src;
     char *destPtr = (char *)dest;

// my edit -- begin
     count = size(srcPtr);
// my edit -- end

// out-of-bounds write
     for (unsigned int i = 0; i < count; i++) destPtr[i] = srcPtr[i];
}

2. size(destPtr) >? size(srcPtr)
   This issue still remains. There is no reliable way of checking it.

[raysan5]

@avx0 I'm merging this change but note that I usually expect the user to provide reliable data, I can't consider all the breaking situations resulting from a missuse of the library, it only adds code complexity and maintenance burden, and only to cover a very minimum number of use-case situations that hardly ever will happen.