[Numpy-discussion] SWIG, typemaps, 2D argout arrays

[Sebastian Haase]

On Tue, Oct 14, 2008 at 7:51 PM, T J <tjhnson at gmail.com> wrote:
> On Tue, Oct 14, 2008 at 1:02 AM, Sebastian Haase <haase at msg.ucsf.edu> wrote:
>> b) I don't want to use Python / numpy API code in the C functions I'm
>> wrapping - so I limit myself to "input" arrays! Since array memory
>> does not distinguish between input or output (assuming there is no
>> copying needed because of alignment or contiguity issues) the only
>> implication of this is that I have to allocate the array outside the C
>> functions.
>
> To clarify, are you saying that you don't make any use of numpy.i?

I have my own (older) version -- I think Bill Spotz has improved
numpy.i by now to do what mine does (supposedly using cleaner code).

> The functions I'm writing are not making any explicit use of the
> Python/numpy API, but I'm sure it shows up in the *_wrap.c file that
> SWIG creates.

If you are using INPLACE arrays (this is I think the word numpy.i
uses) than there is no additional mem allocation. This is/was the most
important feature of all this numpy/SWIG idea.

>
>> input<->conversion transparently for my --- do you need to handle
>> non-contiguous arrays ?
>
> Probably not.
>
So you are fine;
do you know the output array shape beforehand ?

- Sebastian