Optimize zero test through the sage lib

[hivert]

The default implementation of __nonzero__ in element in far from optimal. Thanks #12510 we now can detect where __nonzero__ is not properly implemented. To experiment I tried with CDF here are the timing:

With the default implementation:

p = CDF(pi)
sage: %timeit bool(p)
625 loops, best of 3: 731 ns per loop
sage: %timeit p.is_zero()
625 loops, best of 3: 965 ns per loop

with the following trivial specific implementation:

def __nonzero__(self):
    return bool(self._complex.dat[0]) or bool(self._complex.dat[1])

here is the new timing

sage: %timeit bool(p)
625 loops, best of 3: 342 ns per loop
sage: %timeit p.is_zero()
625 loops, best of 3: 869 ns per loop

This is a so common question that it would be worth removing the default implementation and implement all the specific methods.

Depends on #12510

Component: basic arithmetic

Issue created by migration from https://trac.sagemath.org/ticket/12526