Adding a "fixmap" datastore: Fixed-size mmap() files

[Rick van Rein]

Hi Matthias,

Thanks for responding.

> the tests are written for a lossless database, and that is not going to
> change.

Fair enough.  I suppose my best bet then is to either add other exact
output (output will be consistent, but different from lossless stores)
or skip tests that make non sense for this backend.

>> For another version, I am brooding on a generalisation that spreads data
>> more evenly around the database entries, a bit like in a hologram.
> 
> Sorry to say that this is too little information for me to grok your
> idea.  Can you elaborate this a bit?

Maybe I shouldn't have mentioned it yet, the "brooding" is still going
on.  I need to think a bit more first.

> What "word clashes" do you think of and what problems do you see?

I allocate a fixed number of slots and use a hash to decide where each
word goes.  This leads to slots being occupied by multiple words which
can disturb the perfect outcome.

-Rick