Edge tolerance become so large after stitch (sew)

Hi All,

Attachment is a .brep file, and it contains only two faces. Try to stitch them using sew API (tol=0.01). After stitching, one edge's tolerance changed from 0.5 to 116. I have reproduced this problem in 6.3, 6.5.3 and 6.6.0. The raw two faces have no topological or geometrical error. I think I have reported the similar bugs two times ( including OCC MantisBT system), but on feedback at all. In fact, this .brep file is extracted from a large iges file, the stitching result is not good in raw file. Many topological or geometrical error happens after stitching.
Thanks.

Ding

Forum supervisor's picture

Dear Cauchy Ding,

the reported case is recorded as an additional example in issue #24174

Regards...

Cauchy Ding's picture

OCC Supervisor,

Thanks.

Ding

Cauchy Ding's picture

Hi All,

I think it's caused by the 3d curve. The 3d curves of two edges have lots of control points. In fact, the pcurve of the edge on the face is degree=1 and contains only 2 control points, but after buildcurve3d, the 3d curve contains more than 100 control points and degree=14, it's so bad.
In fact, I have try to rebuild curve3d by reducing GeomAbs_C1 to GeomAbc_C0, change tol from 1.0e-5 to 1.0e-3, and changing maxDegree from 14 to 3. But the 3d curve still contains lots of control points.
I have attached the raw iges file.
Thanks.

Ding