DiscussionsIssue archiveOCCT:Modeling Algorithms

Archived issue #0024874

sprops algorithm computes a different values for shapes, which are got on Windows and Linux platforms

Open CASCADEOCCT:Modeling Algorithmsclosed3 public notes

Search issues

Description

This bug appeared in "boolean bopfuse_complex J6" test case after applying fix from CR24613_1 branch.

There are two shapes in attached files "Attachment 3 (BREP)" and "Attachment 2 (BREP)", which are result of "bopfuse" operation in the test case "boolean bopfuse_complex J6".

Steps to reproduce

On the one hand, try the following script:
Draw[]> restore Attachment 3 (BREP) sh1
Draw[]> restore Attachment 2 (BREP) sh2

Draw[]> sprops sh1 1.0e-12
#Mass : 325.882

Draw[]> sprops sh2 1.0e-12
Mass : 175.227




Values are different. But on the other hand:
Draw[]> restore Attachment 3 (BREP) sh1
Draw[]> restore Attachment 2 (BREP) sh2

Draw[]> explode sh1 f
sh1_1 ... sh1_22

Draw[6]> explode sh2 f
sh2_1 ... sh2_22

#I.e. both sh1 and sh2 contains 22 faces

#Try the script from attached file
Draw[]> source script
Expected square of sh1 = 164.14876500000003
Expected square of sh2 = 164.14876500000003
Delta = 0.0

I.e. both shapes have equal squares (because Delta = 0), which are 164.14876500000003.

Public activity

3 archived notes

Participants are labeled by their role within this record.

01Commenter 1
Dear Igor!
Please review branch CR24874.



Dear testers!
Ask to pay attention to that very many regressions are expected, because square will compute in a new fashion. Differences in squares should not be considered as regressions and corresponding testes must be corrected according to their new behavior.

02Commenter 2
The reason of sprops difference is different shapes on Wind and Lin, but not algo of sprops.
It is necessary to find out why bopfuse produces different result.
03Commenter 3
Well!

TopExp_Explorer in area computing algorithm applied correctly and shared faces must be computed repeatedly.

It is not a bug. Issue can be closed without fix.