Re: unable to have correct signal prototype in documentation.

[Owen Taylor <otaylor redhat com>]

"Matthias Clasen" <Matthias Clasen poet de> writes:

> > and gtk-doc always give me, no matter what the prototype is : 
> > -------------------------
> > "menu-activate"
> >             void        user_function      (LimdiDoc *limdidoc,
> >                                             gpointer user_data);
> > -------------------------
> > 
> > 
> > Marshaller for this signal is limdi_marshal_VOID__POINTER_POINTER so I
> > had, at least, expected :
> > ------------------------- 
> > "menu-activate"
> >             void        user_function      (LimdiDoc *limdidoc,
> >      gpointer arg1,
> >      gpointer arg2,
> >                                            gpointer user_data);
> > -------------------------
> >
> > For information :
> > 
> > i'm using 
> > gtk-doc 0.9
> > and gtk+-2.0.3
> > 
> > Any idea ?
> 
> gtk-doc is just a hack :-( The signal prototypes for know gtk signals are
> hardwired in gtkdoc-scangobj somewhere. We probably should replace that
> by some magic to figure out the prototype based on the marshaller, as you 
> say. 

gtk-doc queries the signal system for the signal signatures.
(Though I think you can probably override them in 
foo-overrides.txt if you are making extensive use of G_TYPE_POINTER.)

The problem here is probably a mis-registered signal.  (Are you
passing in the right parameter type arguments, along with the
marshaler?)

Regads,
                                        Owen