Comment 9 for bug 1804621

The fact is py/pack/_packPredicates.cpp defines aabb() functions for predicates that still return tuples (not lists, as I proposed). I do not think this would necessarily conflict with the use of list in aabbExtrema(), but I still can not say I know all of _packPredicates.cpp

Then, in terms of a manual search for problems in pack-related scripts, I for instance tested:

    examples/gts-horse/gts-operators.py
    examples/gts-horse/gts-random-pack-obb.py
    examples/test/pack-cloud.py
    examples/test/pack-predicates.py
    examples/packs/packs.py
    examples/gts-horse/gts-horse.py
    examples/WireMatPM/wirepackings.py

which all worked well.

Other scripts showed a possibly more problematic behavior

    examples/gts-horse/gts-random-pack.py
=> pkg/common/Facet.cpp:26 postLoad: Facet has coincident vertices 2 (0 -0 -0) and 0 (0 -0 -0)!

    examples/polyhedra/horse.py
=> RuntimeWarning: No spheres are produced by regularHexa-function

    examples/stl-gts/gts-stl.py
=> ZeroDivisionError: float division by zero in regularHexa

I actually tested these scripts with three different versions of the executable:
- Yade 1.20.0 (yade package: aabbExtrema() is a tuple)
- current trunk after #2 ie commit 7c60d78 (aabbExtrema() is a list, after the previous "suspicious" commit)
- and commit 7c60d78 + the patch from #8 (aabbExtrema() is a tuple again).

The above possibly more problematic behaviors were always present in these three cases, suggesting they do not relate at all with the present "bug"... (hence my comment in #8)

Let me know in case you have further remarks.