Re: drawing area and motion hint mask

[Carlos Pereira <jose carlos pereira ist utl pt>]

Emmanuel Touzery wrote:
> Hello,
>
>     I'm trying to detect mouse moves in a drawing area. The default 
> behaviour without the GDK_POINTER_MOTION_HINT_MASK is as expected 
> flooding me with events which I can't handle.
Perhaps you could have something as the code below (it works for me. It 
is based in GtkGlArea / GtkGLExt examples and discussions in this list 
with Havoc Pennington years ago)

I strongly suggest you to have a look in the examples coming with GtkGlExt,

Carlos Pereira
---------------------------------------------------------
g_signal_connect (area, "motion_notify_event",
G_CALLBACK (callback_area_notify), window);

int callback_area_notify (GtkWidget *widget,
GdkEventMotion *event, void *data)
{
my_structure *structure = (my_structure *) data;
GdkModifierType state;
int width = widget->allocation.width;
int height = widget->allocation.height;
int x, y;

if (event->is_hint) gdk_window_get_pointer (event->window, &x, &y, &state);
else
 {
 x = event->x;
 y = event->y;
 state = event->state;
 }

if (state & structure->mask_button1)
 {
 }

else if (state & structure->mask_button2)
 {
 }

else if (state & structure->mask_button3)
 {
 }

/* save x,y as initial values, preparing to next scan:
difference between initial and current x, y values control speed */

structure->begin_x = x;
structure->begin_y = y;

return TRUE;
}



>     So I added that flag, but I still get way too many events.
>
>     In my handler I put this at the top:
>
>     printf ("got mouse move (%d)!\n", event->is_hint);
>
>     And when I don't set the HINT_MASK then is_hint is 0, if I set the 
> HINT_MASK then it's 1.
>     So it seems I'm setting the hint properly.
>
>     That I'm getting more events (not just one, because I never call 
> gdk_event_request_motions() or gdk_window_get_pointer() anywhere in my 
> code) it seems somehow someone else is calling 
> gdk_event_request_motions() or gdk_window_get_pointer() (the examples 
> on internet are not really clear on which one to call, or both, maybe 
> this changed in the past?).
>
>     So I put a breakpoint on both those methods in GDB. And indeed I'm 
> hitting very often this call:
>
> #0  IA__gdk_window_get_pointer (window=0x88bdde8, x=0xbff3aef8, 
> y=0xbff3aef4,
>     mask=0x0) at /build/buildd/gtk+2.0-2.16.1/gdk/gdkwindow.c:3315
> #1  0xb7c300f0 in IA__gdk_device_get_state (device=0x849c840,
>     window=0x88bdde8, axes=0x0, mask=0x0)
>     at /build/buildd/gtk+2.0-2.16.1/gdk/x11/gdkinput-x11.c:788
> #2  0xb7be5db4 in IA__gdk_event_request_motions (event=0x849c840)
>     at /build/buildd/gtk+2.0-2.16.1/gdk/gdkevents.c:889
> #3  0xb7e657d0 in _gtk_tooltip_handle_event (event=0x887b0c0)
>     at /build/buildd/gtk+2.0-2.16.1/gtk/gtktooltip.c:1287
> #4  0xb7d8720d in IA__gtk_main_do_event (event=0x887b0c0)
>     at /build/buildd/gtk+2.0-2.16.1/gtk/gtkmain.c:1643
> #5  0xb7c1434a in gdk_event_dispatch (source=0x849c358, callback=0,
>     user_data=0x0) at 
> /build/buildd/gtk+2.0-2.16.1/gdk/x11/gdkevents-x11.c:2364
> #6  0xb78a7b88 in IA__g_main_context_dispatch (context=0x849c3a0)
>     at /build/buildd/glib2.0-2.20.1/glib/gmain.c:1814
> #7  0xb78ab0eb in g_main_context_iterate (context=0x849c3a0, block=1,
>     dispatch=1, self=0x8483b80)
>     at /build/buildd/glib2.0-2.20.1/glib/gmain.c:2448
> #8  0xb78ab5ba in IA__g_main_loop_run (loop=0x8611fb0)
>     at /build/buildd/glib2.0-2.20.1/glib/gmain.c:2656
> #9  0xb7d877d9 in IA__gtk_main ()
>     at /build/buildd/gtk+2.0-2.16.1/gtk/gtkmain.c:1205
>
>     Something about tooltips. I don't have any tooltips for now in the 
> application and certainly not in this drawing area...
>     So I don't understand how come that tooltip code is getting 
> called. It prevents me from completely controlling when I'm getting 
> mouse move events. That said, I'm surely getting much less events than 
> before and maybe it's "expected". It does complicate my life though. 
> Is there a way to avoid this? What am I doing wrong?
>
>     I can't send the complete source of that program for now, it's 
> already bigger. I'm hoping that it's obvious enough that someone can 
> answer it without me having to extract the problem in a test program.
>
>     Thank you!
>
> emmanuel
> _______________________________________________
> gtk-app-devel-list mailing list
> gtk-app-devel-list gnome org
> http://mail.gnome.org/mailman/listinfo/gtk-app-devel-list