Geoffrey Hutchison wrote:
> On Feb 6, 2007, at 9:00 AM, Noel O'Boyle wrote:
>
>> OBMol.Separate was introducted in SVN r1734. Until this is fixed, you
>> may wish to use the preceding revision.
> ...
>>>> /usr/lib/python2.4/site-packages/_openbabel.so: undefined
>>>> symbol: _ZN9OpenBabel5OBMol8SeparateEi
>
>
> Aha! Yes, SWIG and Python seem to be extremely good at finding
> methods declared in headers, but which do not have implementations in
> the library.
>
> OBMol::Separate() was introduced into the SVN but has no
> implementation. Chris, this seems to be your revision. Was this a
> mistake, or did you forget the changes in mol.cpp?
A mistake, my apologies. I have removed it in SVN 1741. This killer
feature will shortly complete its extensive pre-commit testing and its
release will no doubt shake the world of chemo-informatics. Not being an
"only one sheet of paper on the desk at a time" person I struggle with
maintaining pre-release features properly.
Does SWIG have to behave this way? Is there an option to make it behave
like C++ (whose design I would trust more). Yes, yes, I can hear you
saying "a useful discipline, etc."...
Chris
|
