Re: DPScomposite behaviour, was Coordinate transformations

[Pete French]

> The DPS composite operator works on gState's, not on images

Ah, sorry, I thught we were talking about sending a compositeToPoint methodto
an NSImage (which is what I am doing so it interests me)

> Well, that is actually not true.  The mapping from User coordinates
> to device space coordinates does not preserve size.  The NSImage

True, but the way I read composite is that it merely copies device pixels, and
that if you do this via an NSImage it will only happen if the cached data
matches the destination, else it will be drawn again from scratch. I
took a look through the onoline Cocoa descriptions of the classes and they
seem to specify it a lot more clearly than the NeXT documentation does. But
not clearly enough to know exactly what happens sadly.

> of knowing which size will actually be used.  And this is right,
> the gui classes should not know nor care about device space.

Yes, that seems right - but in that case how does an NSImage know if
the bitmap representation it has cached can be composited or not ?
If I lock focus on a view, send a compositeTo method to an NSImage, then
lock focus on another view which I have scaled and do the same then the
two images should come out at different sizes in device space surely ? I
have not tried this, but this is what I would expect (the documentation does
not make this clear howevr). If though though, then the cached data from the
first operation can clearly not be used for the second ? Which imples some
knowledge by NSImage of the co-ordinate system into which it is being asked to
composite ?

hmmmm....

-bat.