[Bf-blender-cvs] [63e0d3a] blender-v2.77-release: Fix T47705: Freestyle line glitch.

[Tamito Kajiyama]

Commit: 63e0d3a47fdae8a42cd584f13d0530ea57ca0e6f
Author: Tamito Kajiyama
Date:   Tue Apr 5 16:08:45 2016 +0900
Branches: blender-v2.77-release
https://developer.blender.org/rB63e0d3a47fdae8a42cd584f13d0530ea57ca0e6f

Fix T47705: Freestyle line glitch.

The addressed issue is a regression from Blender 2.75, after the internal
switch from double to single precision floating-point numbers in the
Freestyle code base.  Face normal calculations require the higher
precision during the computations, even though the results can be stored
as single precision numbers.

===================================================================

M	source/blender/freestyle/intern/winged_edge/WEdge.cpp

===================================================================

diff --git a/source/blender/freestyle/intern/winged_edge/WEdge.cpp b/source/blender/freestyle/intern/winged_edge/WEdge.cpp
index 87ca3a4..99aa2d2 100644
--- a/source/blender/freestyle/intern/winged_edge/WEdge.cpp
+++ b/source/blender/freestyle/intern/winged_edge/WEdge.cpp
@@ -640,18 +640,19 @@ WFace *WShape::MakeFace(vector<WVertex *>& iVertexList, vector<bool>& iFaceEdgeM
 	vector<WVertex *>::iterator it;
 
 	// compute the face normal (v1v2 ^ v1v3)
-	WVertex *v1, *v2, *v3;
+	// Double precision numbers are used here to avoid truncation errors [T47705]
+	Vec3r v1, v2, v3;
 	it = iVertexList.begin();
-	v1 = *it;
+	v1 = (*it)->GetVertex();
 	it++;
-	v2 = *it;
+	v2 = (*it)->GetVertex();
 	it++;
-	v3 = *it;
+	v3 = (*it)->GetVertex();
 
-	Vec3f vector1(v2->GetVertex() - v1->GetVertex());
-	Vec3f vector2(v3->GetVertex() - v1->GetVertex());
+	Vec3r vector1(v2 - v1);
+	Vec3r vector2(v3 - v1);
 
-	Vec3f normal(vector1 ^ vector2);
+	Vec3r normal(vector1 ^ vector2);
 	normal.normalize();
 	face->setNormal(normal);