[Numpy-discussion] setmember1d memory leak?

[Robert Kern]

Charles R Harris wrote:

> This would be easy to add. I could put in an option side='mask' that
> would return 1 if the object is found, 0 otherwise. The type would be
> integer rather than boolean but I don't see that as a big problem. If I
> add that, I would like to change the keyword to mode instead of side and
> that brings up the question of how to change the interface. It is easy
> to use the same meaning for both mode and side for a while, but it would
> be nice to issue a deprecation warning for the latter and then remove it
> after some fixed period of time. This is a policy question and I think
> the numpy team needs a policy for such things.

If the semantics of the returned value is going to change so greatly, I would
prefer that they be separate functions. I think integer 1s and 0s instead of
booleans *is* a large problem. Such an array can't be used correctly as a mask
in fancy indexing. Inevitably someone will try and wonder why they get an array
with only the first and second elements of the source array.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco