From: dooglus <dooglus@1f10aa63-cdf2-0310-b900-c93c546f37ac>
Date: Mon, 3 Mar 2008 18:11:27 +0000 (+0000)
Subject: Use a bigger epsilon when calculating bline tangents.  It gives a more accurate resul... 
X-Git-Url: https://git.pterodactylus.net/?a=commitdiff_plain;h=85dedbc544c3f802c0733b2be5a2c6cb23448e59;p=synfig.git

Use a bigger epsilon when calculating bline tangents.  It gives a more accurate result, since it reduces the effect of rounding errors.


git-svn-id: http://svn.voria.com/code@1854 1f10aa63-cdf2-0310-b900-c93c546f37ac
---

diff --git a/synfig-core/trunk/src/synfig/valuenode_blinecalctangent.cpp b/synfig-core/trunk/src/synfig/valuenode_blinecalctangent.cpp
index 51bcafc..453b781 100644
--- a/synfig-core/trunk/src/synfig/valuenode_blinecalctangent.cpp
+++ b/synfig-core/trunk/src/synfig/valuenode_blinecalctangent.cpp
@@ -125,7 +125,7 @@ ValueNode_BLineCalcTangent::operator()(Time t)const
 
 	etl::hermite<Vector> curve(blinepoint0.get_vertex(),   blinepoint1.get_vertex(),
 							   blinepoint0.get_tangent2(), blinepoint1.get_tangent1());
-	etl::derivative< etl::hermite<Vector> > deriv(curve);
+	etl::derivative< etl::hermite<Vector> > deriv(curve, 0.0001);
 
 	switch (get_type())
 	{