Revision: 73137
http://sourceforge.net/p/brlcad/code/73137
Author: starseeker
Date: 2019-05-22 21:30:21 +0000 (Wed, 22 May 2019)
Log Message:
-----------
Need more adjustment to the vert normal calculation logic
Modified Paths:
--------------
brlcad/trunk/src/libbrep/cdt.cpp
Modified: brlcad/trunk/src/libbrep/cdt.cpp
===================================================================
--- brlcad/trunk/src/libbrep/cdt.cpp 2019-05-22 19:42:10 UTC (rev 73136)
+++ brlcad/trunk/src/libbrep/cdt.cpp 2019-05-22 21:30:21 UTC (rev 73137)
@@ -410,7 +410,17 @@
* during surface point builds, we use the normals as a curvature guide.
* When we hit singular trims, we use the normal calculated below to ensure
* the normal is always the same at the singularity regardless of the
- * particular UV point on the singular trim we are evaluating. */
+ * particular UV point on the singular trim we are evaluating.
+ *
+ * TODO - rather than averaging, we probably want to get a per-face
+ * normal for each vert. When we hit singular trims we can do a lookup
+ * by vert, but we still want a normal specific to the face since
+ * an average will tend to visually corrupt sharp edges. Let's do this -
+ * iterate over all trims, build the set of singular trims, for those trims
+ * get the associated verts and for each vert and each face associated with
+ * that vert that has a singular trim, calculate a normal based on the
+ * surface evaluation at that point from non-singular trims in the face
+ * that share the same vertex point.*/
if (!s_cdt->w3dnorms->size()) {
for (int index = 0; index < brep->m_V.Count(); index++) {
ON_BrepVertex& v = brep->m_V[index];
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
_______________________________________________
BRL-CAD Source Commits mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-commits
