iterutils: optimize chunked_iter

[immerrr]

Hello and thank you for such a nice library.

This PR optimizes iterutils.chunked_iter performance (2x on pypy, 5x on cpython).

Before

immerrr@mmrcomp:~/sources/boltons$ time ~/pypy-5.1.1-linux64/bin/pypy bench_chunked_iter.py

real    0m1.862s
user    0m1.824s
sys     0m0.020s
immerrr@mmrcomp:~/sources/boltons$ time /usr/bin/python2.7 bench_chunked_iter.py

real    0m12.275s
user    0m12.256s
sys     0m0.004s
immerrr@mmrcomp:~/sources/boltons$ time /usr/bin/python3.5 bench_chunked_iter.py

real    0m17.307s
user    0m17.288s
sys     0m0.008s

After

immerrr@mmrcomp:~/sources/boltons$ time ~/pypy-5.1.1-linux64/bin/pypy bench_chunked_iter.py

real    0m0.926s
user    0m0.868s
sys     0m0.044s
immerrr@mmrcomp:~/sources/boltons$ time /usr/bin/python2.7 bench_chunked_iter.py

real    0m2.361s
user    0m2.340s
sys     0m0.008s
immerrr@mmrcomp:~/sources/boltons$ time /usr/bin/python3.5 bench_chunked_iter.py

real    0m3.007s
user    0m2.988s
sys     0m0.008s

With bench_chunked_iter.py being:

from boltons.iterutils import chunked_iter
from collections import deque

try:
    xrange
except NameError:
    xrange = range


for i in xrange(100):
    deque(chunked_iter(xrange(1000000), 100), 0)

[mahmoud]

Hard to argue with numbers like that, and shorter code! Thanks!