[C++-sig] Re: Pyste-generated extentsion Aborts Python

Jim Wilson wilson at afn.org
Fri Jul 4 03:25:27 CEST 2003


Nicodemus,

You wrote (regarding my abbreviated class Fl):

> This does not seems like a good idea... the policies are *really* 
> needed. If you want to expose only a few functions, then you can 
> always exclude() the ones that you don't want.

I revised my fltk.h:

#include </usr/local/include/FL/Fl.H>		// Where my FLTK
#include </usr/local/include/FL/Fl_Window.H>	//  headers reside

and fltk.pyste (to exclude the troublesom functions):

f = Class("Fl", "fltk.h")
exclude(f.readqueue)
exclude(f.first_window)
exclude(f.next_window)
exclude(f.modal)
exclude(f.grab)
exclude(f.belowmouse)
exclude(f.pushed)
exclude(f.focus)
exclude(f.selection_owner)
exclude(f.get_boxtype)
w = Class("Fl_Window", "fltk.h")
exclude(w.icon)
exclude(w.child)
exclude(w.current)
exclude(w.array)
exclude(w.resizable)
exclude(w._ddfdesign_kludge)

And *now* I remember why I did it the way I did.  Pyste silently
generates (a somewhat longer) fltk.cpp (available renamed) at:

	http://www.afn.org/~wilson/fltk.cxx

But when compiling it, gcc goes berserk and generates all sorts of
cryptic template diagnostics.  Since I don't understand templates,
I'm hopelessly lost!  Hence my pragmatism ;).

If it'll do any good, I'm happy to tgz or zip the FLTK headers for
your entertainment.

Jim












More information about the Cplusplus-sig mailing list