Behavior of nodes defined pst-eucl seems to be different from those defined in pst-plot.
In the section titled Direct node, I attempted to create a square with pq as it base inclined at a slope of line pq. Unfortunately, it produces a wrong output.
But if I buffered the rotated point with a node defined in pst-node, I got the correct output.
This inconsistent behavior might cause confusion and difficult to debug. What do you think?
\documentclass[preview,border=12pt,12pt,varwidth]{standalone}
\usepackage{pst-eucl}
\begin{document}
\section*{Direct node}
\begin{pspicture}[showgrid,saveNodeCoors,NodeCoorPrefix=N](10,10)
\pstGeonode(4,8){p}(4,6){r}
\pstRotation[RotAngle=90]{p}{r}[q]
\rput{!Nqy Nry sub Nqx Nrx sub atan}(r){\psframe(!Nqx Nqy Nrx Nry Pyth2 dup)}
\end{pspicture}
\vspace{4cm}
\section*{Indirect node}
\begin{pspicture}[showgrid,saveNodeCoors,NodeCoorPrefix=N](10,10)
\pstGeonode(4,8){p}(4,6){r}
\pstRotation[RotAngle=90]{p}{r}[temp]
\pnodes(temp){q}
\rput{!Nqy Nry sub Nqx Nrx sub atan}(r){\psframe(!Nqx Nqy Nrx Nry Pyth2 dup)}
\end{pspicture}
\end{document}

Question
Why do we need to buffer the pst-eucl nodes before passing them to \rput? Why don't make the behavior of nodes defined in pst-eucl consistent with those defined in pst-plot?