issue#262 error in bounding box computation

[mpoudot]

…edge or surface

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 48.87%. Comparing base (8210518) to head (b7f4b79).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #239      +/-   ##
==========================================
+ Coverage   48.86%   48.87%   +0.01%     
==========================================
  Files         466      466              
  Lines       50515    50523       +8     
==========================================
+ Hits        24682    24693      +11     
+ Misses      25833    25830       -3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nicolaslg]

BEWARE : copy-pasting the sphere creation command does not work properly.

The BRepBndLib::AddClose can not be applied in our case, as it expects the shape to be polygonal with planar faces (see https://dev.opencascade.org/doc/refman/html/class_b_rep_bnd_lib.html#a1c5a09bca57f8145e163c59693a45bbf).
The analyzer.IsValid() does not check that property (see https://dev.opencascade.org/doc/refman/html/class_b_rep_check___analyzer.html#a99892412af7e25fb4d5c19c7d54a41fb)

I think checking whether the returned bounding box is flat is not sufficient. Considering a volume and the created block

ctx.getGeomManager().newSphere (Mgx3D.Point(0, 0, 0), 1, 125)
ctx.getTopoManager().newFreeTopoOnGeometry ("Vol0000")

Currently in the branch is returned the following block :

When always calling BRepBndLib::Add(shape, box) instead we have this behavior, which is what we expect :

But when the sphere is scaled, the call to BRepBndLib::Add(shape, box) will return a bounding box that is not "correct"

ctx.getGeomManager().scaleAll(1.000000e+01, 2.000000e+01, 1.000000e+01)

Whereas calling BRepBndLib::AddOptimal(shape, box) gives

BRepBndLib::AddOptimal(shape, box), or rather BRepBndLib::AddOptimal(shape, box, false) might be the better choice; I have not managed to exhibit differences when the third parameter useTriangulation==false. I guess it is better still to put it at false considering we want to behave the same way whether a triangulation is attached to the shapes or not.