Re: [gtkmm] Probably a stupid Gtk+ question

[Nathan Hurst <njh hawthorn csse monash edu au>]

Paul Davis wrote:

> > > a *lot* to do with it. binding scheme/perl/ruby/php/python to C++ is
> > > to all effects and purposes so hard as to be impractical. when you add
> > > in the instability of the C++ ABI, even if you loved C++, its a pretty
> > > clear choice.
> > >
> > >      
> > This is not particularly true.  I've written python interfaces to pure 
> > C++ libraries without trauma.  I don't think even FuzzyBaseClass isn't 
> > the problem that people claim it is.
> >    
>
> you either to have export "extern C" static functions, or you have to
> know how names are mangled. you can't call a C++ function unless
> one of those two things has been done. either one of them is pretty
> problematic. 
Sure, but that's no worse than having to call functions things like 
gtk_window_new()  Your main implementation can still be done in a clear 
and simple manner.  The interface could even be automatically generated.

njh