Re: Wrapping gtkimageview

[Andreas Volz <lists brachttal net>]

Am Tue, 04 Mar 2008 21:15:55 +0100 schrieb Murray Cumming:

> On Tue, 2008-03-04 at 20:10 +0100, Andreas Volz wrote:
> > Thanks. This was the reason. The wrapping is not yet finished, but
> > it's
> > going on. What is the current policy. Do you prefer the mm wrapper
> > library to stay on the authors website or do you centralize it in
> > the gtkmm SVN?
> 
> Generally
> a) If you already have a GNOME svn account then feel free to put it in
> gnomemm.

No I've not yet a SVN access to GNOME.

> b) If it's a very commonly used library, I'd like it to be in GNOME's
> svn, in gnomemm, and I would try to get an account for you. But I
> don't know how useful/used GtkImageView is.

I'm in contact with the developer of GtkImageView. I could ask for a
list where this widget is used. Maybe we decide this later.

> > Would you think someone from this list would review the wrapper if
> > it's
> > running to be sure no wrapper newbie mistakes are done?
> 
> Sure. Just point us at cvs/svn/whatever when you are ready.

Ok.

I wrote the first tests and the wrapper is working so far. Not really
completed, but enough to add it to a gtkmm container.

But I've some problems to connect the signals. See my code from the .hg
file:

_WRAP_SIGNAL(void on_mouse_wheel_scroll(), "mouse-wheel-scroll")

If I compile this I get:

imageview.cc: In member function 
'virtual void Gtk::ImageView::on_on_mouse_wheel_scroll()':
imageview.cc:339: error: too few arguments to function

See the prototype of the C signal:

void user_function (GtkImageView *view, GdkScrollDirection direction,
                    gpointer user_data)

And the generated def file for the signals:

(define-signal mouse-wheel-scroll
  (of-object "GtkImageView")
  (return-type "void")
  (when "last")
  (parameters
    '("GdkScrollDirection" "p0")
  )
)

The question is why the gpointer parameter isn't generated automatic.
Should I simply add it to the def file? Or is there a way to generate
it?

regards
Andreas