Support Qureg.__int__ for multiple bits (little-endian)

[Strilanc]

I think this is the natural thing to do when asking for the int in a quregister. (Existing tests weren't changed because they happened to not check for this.)

A reasonable follow-up change would be to remove the len==1 constraint from __bool__ and __non_zero__, and have them check if __int__ is non-zero instead. I can do that in this PR if you want.

[thomashaener]

This is actually what our first implementation did, but we decided to drop this feature.

Rather than adding this functionality to Qureg, I would suggest to add a new quantum type Quint (and eventually also Qufixed, Qufloat, ...). This has the advantage that such a conversion must be performed explicitly (it's a different interpretation of a Quregafter all): a = Quint(qureg1) and it would also behave as expected in terms of operations:

...
a = Quint(qureg1)
b = Quint(qureg2)
a += b # as opposed to concatenation: qureg1+qureg2
Measure | a
m_a = int(a) # is actually a+b

I think this explicit conversion helps avoid bugs such as if bool(qureg): ..., where users might have wanted all(bool(q) for q in qureg).