@kha I added `qsort` for sorting environment extension entries, but I am
wondering if we could use it as a benchmark in our paper. It is a pure implementation; it is
fast; it implements the real quick sort algorithm with in-place updates if the
input array is not shared. If the array is shared it performs
a single copy and then switches to in-place updates.
The array dimension is now stored inside the array.
The main motivation is that it reflects the actual runtime implementation.
We need to store the array size to be able to GC it.
So, it feels silly to have the array size stored in each array object,
but we cannot use this information.
For example, in the `hashmap` we implemented the bucket array using
`array`, and we store the `size` of the array.
Same for the Lean3 `buffer` object.
The `buffer` object doesn't even need to exist.
The actual `array` implementation is the `buffer`
