I have confirmed that there seems to be no relationship between the angle in styleObj and the map rotation in map->gt.rotation_angle. On the other hand, the map rotation is applied to label angles before rendering (seen via use of rotation_angle in mapdraw.c with label.angle).
I am not entirely clear how symbol rotation is achieved. I see there is an msRotateSymbol()
function called from mapgd.c extensively if style.angle is non-zero. I'm not sure how it is
done in other contexts.
I see the original map rotation work was done before we developed the RFC process, otherwise I would suggest updating and revoting the map rotation RFC.
I'm afraid fixing symbol rotation is not something I have time to pursue on my own currently, but I'm willing to provide supporting information on how map rotation works and is applied to labels to anyone willing to work on symbol support. I do think we need to contemplate whether we ought to be applying map rotation automatically to all symbols or not. In the case of labels we only applied the map rotation if the labels had a non-zero angle.
For now I have removed the 5.4 milestone as I don't feel that is realistic. This might either belong in 6.0 or possibly a 5.4 point release.
