Re: The connection bug, was: Re: [Gzz] Summing up...

[Tuomas Lukka]

> >>>>The problem is that the connections assume that (1,1) transforms to the 
> >>>>lower right corner of the rectangle, but with (w=2, h=2)-- or 
> >>>>(w=1,h=1)--, it's still the ul corner, just one or two pixels away.
> >>>> 
> >>>>
> >>>>       
> >>>>
> >>>Wait, now I don't understand. Could you explain the problem in more 
> >>>detail?
> >>>
> >>>     
> >>>
> >>Hm. I don't know how.
> >>
> >>Ok, another example. Let's say we have a cell with the bounds (x=100, 
> >>y=100, w=50, h=20). This used to be our coordinate system, so if there 
> >>was a connection downward, it would be anchored at (x+w/2, y+h) = (125, 
> >>110). Nowadays our coordinate system is (x=100, y=100, w=2, h=2), so the 
> >>anchor point will be at (x+w/2, y+h) = (101, 102)-- for practical 
> >>purposes, the upper left corner of the cell.
> >>   
> >>
> >
> >Ahh, because of no scaling in AWT. Ok.
> >
> 
> Argh. NO. Why do you think so? NOTHING in the example above is related 
> to scaling. Setting w=2 and h=2 was done because of scaling, but the 
> effects are no different whether scaling is actually turned on or not. 
> The anchor point is (101, 102), i.e. ul corner, no matter whether 
> scaling is on or not. (Proof is that the gl connections are also between 
> the ul corners.)
> 
> Sorry for miscommunicating, but I don't know how to explain this better :-(

No, it *is* related to scaling: in the right coordinate system, (2,2) *would*
be the lower right corner. 

The point where the connection starts should be taken from the *box* coordinate
system, in which it would be right, if scaling is applied. Ok?

        Tuomas