None compressor, dict store, shared data

[alimanfoo]

When a Zarr array uses no compression, if data are stored in a dict, currently data can end up being shared with another array, which could lead to strange and unexpected behaviour. E.g.:

In [31]: a = np.arange(20)

In [32]: z = zarr.zeros(20, chunks=10, compressor=None, dtype=a.dtype)

In [35]: z[:] = a

In [36]: z[:]
Out[36]: 
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
       17, 18, 19])

In [37]: a[:] = 0

In [38]: z[:]
Out[38]: array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])