Just wondered if this (oldie) was still on a list / someone's radar ?
Select edges around a cyl end face and use above op., - observe weirdness.
pp
Select edges around a cyl end face and use above op., - observe weirdness.
pp
11-22-2012, 02:12 AM
I can confirm this, if it even needed confirming.
11-25-2012, 01:27 AM
It is difficult to rationalise this behaviour.
It seems intuitive that all the edges should remain even when they are all selected, but should the edge length of a single edge change if it is the only one selected? In the case of a single edge, it is probably expected that the edge simply move along the expected normal (the average of the face normals on either side of the edge), and not change length. If the edge were to change length this would more likely be a vertex normal move. In the case of two adjacent edges, their shared vertex is pulled by one edge not both. The position of the vertex is not averaged. Actually, the vertex is positioned for each edge successively, and comes to rest relative to the last edge moved. A solution might be to have an accumulator sort all the vertices which are shared by more that one edge and then average their position for the final output, thus taking into account all the edge normals equally. I guess that wasn't so difficult
11-25-2012, 06:15 AM
I've added a fix to my github.
branch: rj/extrude_edge_normal
11-25-2012, 06:34 AM
Woo! Great job sir.
|
|