>
> Hello ROOT developers,
>
> is it intentional that TCanvas::HandleInput doesn't have a default
> for an X/Win event? Right now if an event is different from a small
> number of events generated by mouse (press-left-button, move mouse,
> release-left-button etc) TCanvas::HandleInput just returns. For
> example it is not possible to process a keyboard hit...
The reason for that is Pick() method. Very this method defines what
kind of the object the Mouse Cursor "is pointing" (via
DistanceToPrimitive() method of TObject()). To use other events
some sort of the way to link that event with ROOT object
must be offered. For example TGLViewerImp (see below).
What is special with this example. The point is that TGLViewerImp is
a special "Canvas" object and it can contain only ONE type of the
objects, namely TPadOpenGLView.(When the normal TCanvas may contain
ANY type of the objects). So for this object it is out question which
ExecuteEvent() should be called.
//______________________________________
void TGLViewerImp::HandleInput(EEventType button, Int_t x, Int_t y) {
if (!fGLView) return;
switch (button) {
case kButton1Down:
case kButton1Up:
case kButton1Motion:
case kKeyPress :
fGLView->ExecuteEvent(button,x,y);
break;
default:
break;
}
}
Where
TPadOpenGLView *fGLView; // Pointer to Pad GL View object;
and
//____________________________________________________________________________
void TPadOpenGLView::ExecuteEvent(EEventType event, Int_t px, Int_t py)
{
switch ((EEventType)event)
{
case kButton1Down:
//*-* Save the start coordinates
fMouseInit=kTRUE;
fMouseX = px;
fMouseY = py;
break;
case kKeyPress :
{
Char_t chCharCode = (Char_t) px;
// if (chCharCode != '+' || chCharCode != '-') chCharCode |= 0x20;
MoveModelView(chCharCode, py);
break;
}
case kButton1Up:
if (!fMouseInit) break;
fMouseInit=kFALSE;
RotateView(px,py);
fMouseX = px;
fMouseY = py;
PaintScene();
break;
case kButton1Motion:
if (!fMouseInit) break;
RotateView(px,py);
fMouseX = px;
fMouseY = py;
PaintScene();
break;
default:
break;
};
}
>
> I'd expect a lot of (potential) users would like to use this
> feature.
Yes, but for that probably we need first to be able to "Select"
some object with the very mouse event (actually it should a
separate method called by the mouse events) and THEN apply another
events. (See TObjectView class descrition as example too)
Regards,
Valery
Dr. Valeri Faine (Valery Fine)
------------ ------------- Phone: +41 22 767 4921
CERN FAX : +41 22 767 7155
CH-1211 Geneva, 23 mailto:fine@mail.cern.ch
Switzerland http://nicewww.cern.ch/~fine