From e1f7382910e308e08ec9a72998a987f492492b8b Mon Sep 17 00:00:00 2001
From: ifv <ifv@opencascade.com>
Date: Tue, 5 Jul 2022 17:30:35 +0300
Subject: [PATCH] 0028243: Bad result of the canonical recognition

LProp/LProp_CLProps.gxx - minor modification to avoid floating overflow because of multiplication of large values
---
 src/LProp/LProp_CLProps.gxx | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/LProp/LProp_CLProps.gxx b/src/LProp/LProp_CLProps.gxx
index 8afa74d53c..499bab6cda 100644
--- a/src/LProp/LProp_CLProps.gxx
+++ b/src/LProp/LProp_CLProps.gxx
@@ -232,14 +232,15 @@ Standard_Real LProp_CLProps::Curvature ()
   {
     Standard_Real N = myDerivArr[0].CrossSquareMagnitude(myDerivArr[1]);
     // if d[0] and d[1] are colinear the curvature is null.
-    Standard_Real t = N/(DD1*DD2);
+    //Standard_Real t = N/(DD1*DD2);
+    Standard_Real t = N / DD1 / DD2;
     if (t<=Tol)
     {
       myCurvature = 0.0;
     }
     else
     {
-      myCurvature = sqrt(N) / (DD1*sqrt(DD1));
+      myCurvature = sqrt(N) / DD1 / sqrt(DD1);
     }
   }