Re: bug (?) and question

[Owen Taylor <owt1 cornell edu>]

Robert Roebling <roebling@ruf.uni-freiburg.de> writes:

>   Hi,
> 
> I'm using 0.99.8 and I wonder if it was
> intentional to have gtk_entry_button_press()
> always return FALSE. It happens to break
> my library (wxGTK).

Currently all event handling routines in GTK return FALSE - presumably
so that the events are available to the parent windows. It's a bit
dubious behavior, but since wrapper libraries need to conform to GTK,
not vice-versa, could you explain:

 What you are trying to achieve,
 How you are trying to do it,

and

 What happens?

Perhaps there is a simple workaround. If not, it would nice to have
details of the justification before making a change.

[ Perhaps that sounds a bit to harsh. I don't mean that having things
like wxGTK work isn't important. I just mean that 95% of all GTK
programs now use the raw library, or libraries that translate the GTK
API directly into other languages - so I'm hoping that the any changes
needed will have a justification that can be translated into things
people will want to do in programs using the native GTK API ]
 
> I'd like to connect the the motion event.
> I have set the motion mask when creating
> the window but I only get exactly one
> motion event where I expected a constant
> flow of events as long as the mouse is
> moving over my window. What am I doing
> wrong. Maybe it has something to do with
> the drag stuff that I guess also depends
> on motion events.

You are probably setting GTK_POINTER_MOTION_HINT mask, which says (to
X), send me an event only when a button press occurs (or some other
circumstances), or after I call
gdk_window_get_pointer()/gtk_widget_get_pointer. (The both call
XQueryPointer).

This is nice for restricting the flow of events so that X doesn't
get ahead of your program.

Regards,
                                        Owen