Revision: 73848
          http://sourceforge.net/p/brlcad/code/73848
Author:   starseeker
Date:     2019-09-06 19:31:03 +0000 (Fri, 06 Sep 2019)
Log Message:
-----------
The refinement of edges has to be paired with surface sampling, or we quickly 
get into unworkable territory.  NIST7 Solid1.s won't quite work even in this 
basic mode because of a corner meeting that produces an inconvenient triangle.  
That problem gets heightned drastically when we sample more deeply.  Need to 
look carefully at surface sampling to make sure density near edges is adequate.

Modified Paths:
--------------
    brlcad/trunk/src/libbrep/cdt.cpp

Modified: brlcad/trunk/src/libbrep/cdt.cpp
===================================================================
--- brlcad/trunk/src/libbrep/cdt.cpp    2019-09-06 18:17:10 UTC (rev 73847)
+++ brlcad/trunk/src/libbrep/cdt.cpp    2019-09-06 19:31:03 UTC (rev 73848)
@@ -175,6 +175,7 @@
     s_cdt->edge_segs_3d[trim.Face()->m_face_index].Insert(p1, p2, (void *)pe);
 }
 
+#if 0
 struct rtree_loop_leaf {
     struct ON_Brep_CDT_State *s_cdt;
     int loop_index;
@@ -229,6 +230,7 @@
     // Keep checking for other loops - we want the smallest target length
     return true;
 }
+#endif
 
 double
 median_seg_len(std::vector<double> &lsegs)
@@ -1020,7 +1022,7 @@
 
     // Sample the surface, independent of the trimming curves, to get points 
that
     // will tie the mesh to the interior surface.
-    GetInteriorPoints(s_cdt, face.m_face_index);
+    //GetInteriorPoints(s_cdt, face.m_face_index);
 
     cdt_mesh::cdt_mesh_t *fmesh = &s_cdt->fmeshes[face.m_face_index];
     fmesh->f_id = face.m_face_index;
@@ -1629,6 +1631,7 @@
 
        }
 
+#if 0
        // On to tolerance based splitting.  Process the non-linear edges first 
-
        // we will need information from them to handle the linear edges
        //
@@ -1963,6 +1966,8 @@
            s_cdt->e2polysegs[edge.m_edge_index].insert(new_segs.begin(), 
new_segs.end());
        }
 
+#endif
+
        // Split singularity trims in 2D to provide an easier input to the 2D 
CDT logic.  NOTE: these
        // splits will produce degenerate (zero area, two identical vertex) 
triangles in 3D that have
        // to be cleaned up.

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

Reply via email to