Python dictionary inspired hash table implementation

[Manfred_Nowak]

bearophile wrote:

> creating benchmarks is a very tricky thing

Yes. And as one may recognize: Moritz designed the benchmarks in such a 
way,
- that collisions at least aren't obvious
- that the locality of two consecutive inserts/acesses seems to be 
close to maximum

In addition: the space effciency of as low as 25% which Moritz' attempt 
allows may be inacceptable.   


Whithout some agreed scenario, i.e. specifications

- about space effciency
- about static properties of the key spaces
- about dynamic properties of the actually used set of keys and its 
insert/access sequences

every user of a dictionary will be able to claim, that his preferred 
implementation is better than any other.

-manfred