Transparent pixels and report position

I want to clarify something that is giving me some trouble.

For reporting position VASSAL users the center pixel in the image, ignoring any Non-Rectangular trait, right?

For stacking purposes VASSAL uses the bottom left corner of the border, now considering the Non-Rectangular trait, right?

I had so much trouble making my images of the same size and in the proper position so that they would stack nicely independent of their transparent areas (if you overlap all of them in the image editor it would flow nicely), that now some pieces report the wrong location because the visible part isn’t centered on the image (eg, the piece is just on the top or bottom of the image).

Basically I was expecting the report position to follow the Non-Rectangular trait center, and stacking to use the real image size when positioning each image in the “right” place.