0024772: Intersection of cylinder and cone produces incorrect results

Notes
(0069011) apv (tester) 2017-08-01 10:49	Problem described in issue is reproduced on current state of OCCT.

(0076833) nbv (developer) 2018-06-19 13:09	This problem is limitation for the current version of the intersection algorithm. The reason is following: 1. Cone-Cylinder intersection algorithm produces ALine. Aline is internal (for OCCT) 3D-intersection curve. 2. This ALine is projected to the every intersected surfaces (in order to obtain correspond p-curve). For that, ALine is discretized on several points, every point is projected to the surfaces and after that obtained set of 2D-points is approximated to the p-curve. In case of near-plane cones projection of 3D-point to the cone cannot be computed correctly because projector cannot decide from which side of the apex the projection point should be chosen (for valid cones, the nearest point is chosen but for near-plane cones correspond distances differ insignificantly). 3. Moreover, projection is used while computation of value of the ALine. So, methods IntAna_Curve::Value(...) and IntPatch_ALine::Value(...) work wrong with near-plane cones. Possible ways to fix: 1. Use bi-parametric intersection algorithm (IntPatch_PrmPrmIntersection). This algorithm initially works in 2D-space. So it does not made any projections. Consequently, it will produce valid result. 2. Replace existing intersection algorithm with new one like already used while Cylinder-Cylinder intersecting. New algorithm will work in 2D-space too and is oriented to special types of intersected surfaces (analytical cases: Cylinder-Cylinder, Cylinder-Cone, Cone-Sphere etc.) The bright advantage of 2nd step in compare with 1st is the fix 0023178 and 0026894 where bi-parametric intersector worked too long but could not find any result. New algorithm currently works perfectly. I am going to implement the way 2 while fixing the issue 0024418.

(0080373) nbv (developer) 2018-10-26 13:20	Dear Mikhail, We are looking forward your decision. Is this issue possible to be fixed? P.S. Please look at the messages 0024772:0076833 and 0027705:0080371.

(0080517) msv (developer) 2018-10-30 12:39	I propose to go by 2nd way.

(0080519) msv (developer) 2018-10-30 12:43	When have time.

Issue History
Date Modified	Username	Field	Change
2014-03-28 01:02	Roman Lygin	New Issue
2014-03-28 01:02	Roman Lygin	Assigned To	=> ifv
2014-03-28 01:02	Roman Lygin	File Added: archive.zip
2014-03-28 01:03	Roman Lygin	File Added: r_1.png
2014-09-12 14:36	abv	Target Version	=> 7.1.0
2016-10-28 11:54	msv	Target Version	7.1.0 => 7.2.0
2017-07-21 11:34	msv	Target Version	7.2.0 => 7.3.0
2017-08-01 10:49	apv	Test case number	=> bugs modalg_7 bug24772
2017-08-01 10:49	apv	Note Added: 0069011
2017-12-05 17:01	msv	Target Version	7.3.0 => 7.4.0*
2018-05-24 14:17	nbv	Relationship added	related to 0027705
2018-06-19 10:38	nbv	Relationship added	related to 0029807
2018-06-19 13:09	nbv	Note Added: 0076833
2018-06-19 13:10	nbv	Relationship deleted	related to 0029807
2018-06-19 13:10	nbv	Relationship added	related to 0024418
2018-10-26 13:20	nbv	Note Added: 0080373
2018-10-26 13:21	nbv	Assigned To	ifv => msv
2018-10-26 13:21	nbv	Status	new => feedback
2018-10-30 12:39	msv	Note Added: 0080517
2018-10-30 12:43	msv	Note Added: 0080519
2018-10-30 12:43	msv	Assigned To	msv => nbv
2018-10-30 12:43	msv	Status	feedback => assigned