[SciPy-dev] Is it ok to depend on ctypes for scipy code ?

[Robert Kern]

David Cournapeau wrote:
> Robert Kern wrote:
>> The linking process on Windows requires an extension module to actually have the
>> initfoo function.
> Don't tell me this is again windows doing things totally differently 
> than everybody else for the sake of being incompatible :)

<shrug> Dynamic linking is inherently platform specific. Everyone does it
differently. It just happens that Windows' choices gets in the way in this
particular instance.

> On windows, 
> the major problem I have with ctypes for pyaudiolab is detecting the 
> library, because numpy.distutils does not find the dll if the .lib is 
> not there (the library I depend on, sndfile, is compiled by mingw by the 
> main author).

Of course, that's not a problem with ctypes, nor even numpy.distutils. It's just
an unsupported use of numpy.distutils.

> Of course, I do not pretend there is no problem, I just don't understand it.

I think you may be reading my statement out of context. Andrew asked about
building shared libraries for ctypes' use by abusing distutils to build a fake
extension module from it. My response was providing the reason why that doesn't
work in general.

-- 
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