Implements the concept of "mass concatenation" with a new, specialized ASTNode

[Altoids1]

One common issue with most implementations of a concatenation operator is that it's left-associative and sequential, just like any typical operator. A .. B .. C .. D allocates into memory A, then AB, then ABC, then ABCD, on and on. For very long runs of successive concatenation, this can cause very large and unnecessary reallocation events for each iteration of the process.

This patch changes the behaviour of these to forgo the typical behaviour of binary operations in favour of a "single-batch" approach, where each operand is evaluated and appended to a master string, with the goal of making the number of allocations at least semi-linear with the number of operands.

I'm not wholly satisfied with this implementation, but it's a good step towards reducing the amount of redundant malloc/free calls currently littering the João project at the moment.

This PR made this script run twice as fast compared to the main branch:

/main(args)
{
    for(Value i = 0; i < 100_000; i+=1)
    {
        String str = "This index is called '" .. i .. ", and twice its value is " .. i * 2 .. "!";
        ##print(str);
    }
}