`lib/core` is extended with CircularArray for an array with efficient push/pop/shift/unshift
Some number with bench_seq. I used valgrind with n=1000
push (Ir/call)
* list: 383
* unrolled list: 188
* circular array: 135
* array: 85
pop
* unroll: 180
* list: 79
* array: 66
* circ: 47
unshift
* array: 31,755 (because memmove)
* list: 385
* unroll: 322
* circ: 127
shift
* array: 36,168 (because memmove)
* unroll: 180
* list: 79
* circ: 50
insert:
* circ: 179,811 (because manual copy item by item)
* array: 27,282
* unroll: 12,235
* list: 11,500
Pull-Request: #1875
Reviewed-by: Lucas Bajolet <r4pass@hotmail.com>
Reviewed-by: Jean-Philippe Caissy <jpcaissy@piji.ca>