DiscussionsIssue archiveOCCT:Modeling Algorithms

Archived issue #0022969

Wrong confusion in BRepLib_MakeEdge

CommunityOCCT:Modeling Algorithmsclosed4 public notes

Search issues

Description

I am afraid I have found a bug in
BRepLib_MakeEdge::Init(const Handle(Geom_Curve)& CC,
           const TopoDS_Vertex& VV1,
           const TopoDS_Vertex& VV2,
           const Standard_Real pp1,
           const Standard_Real pp2)
in file src\BRepLib\Attachment 1 (CXX) at line 766:
Standard_Real epsilon = Precision::Confusion();

This "epsilon" is used to check if two parameters on a curve are coincident.
But Precision::Confusion() is a 3d tolerance, not 1d.
It seems to me Precision::PConfusion() should be used, which is made for parameters.

Additional information

fixed in attachment

Public activity

4 archived notes

Participants are labeled by their role within this record.

01Commenter 1
Added additional file with description - Attachment 2 (DOCX).
02Commenter 2
Fix has been applied and reviewed without remarks.

Branch http://svn/svn/occt/branches/OCC22969 is ready for testing.

Please test.
03Commenter 2
Dear Commenter 1,
   Workbench KAS:dev:mkv-22969-occt was created from SVN branch http://svn/svn/occt/branches/OCC22969
   (and mkv-22969-products from trunk) and compiled on Linux and Windows platforms.
   
   There are not regressions in mkv-22969-products regarding to KAS:dev:products-20120203-opt
   
   See results in /QADisk/occttests/results/KAS/dev/ mkv-22969-products_09022012/lin
   See reference results in /QADisk/occttests/results/KAS/dev/products-20120203-opt_03022012/lin
   See test cases in /QADisk/occttests/tests/ED
04Commenter 1
Integrated into trunk of occt repository

Date: 2012-02-10 16:57:17 +0400 (Fri, 10 Feb 2012)
New Revision: 10421

Modified:
   trunk/src/BRepLib/Attachment 1 (CXX)