Bug in Compound location handling

[MatthiasJ1]

a = GridLocations(60, 30, 2, 2) * extrude(Rectangle(50, 20), -5)
b = pack([Compound(children=[a[0]]), *a[1:]], 1)

If you fuse b as is, it will be fixed afterwards which shows this is not a bug in pack.

[gumyr]

Note for self - here is a hypothesis on what might be happening:

There were changes in the code in November that eliminated "needless" Compound wrappers around objects - a common occurrence with OCCT operations - where "needless" is defined as a Compound that has only one member. In this case the singleton Compound has a non zero Location so stripping that off changes where the enclosed shape is ultimately located.

If this is the problem then this "unwrapping" procedure needs to re-locate the enclosed shape.

[gumyr]

Maybe not:

box = Compound(children=[Solid.make_box(1, 1, 1)]).locate(Pos((2, 0, 0)))
assembly = Compound(children=[box]).moved(Pos((0, 2)))
print(assembly.show_topology("Solid"))

unwrapped = unwrap_topods_compound(assembly.wrapped)
print(unwrapped, Location(unwrapped.Location()))

             Compound at 0x7b341b3a2a40, Location(p=(0.00, 2.00, 0.00), o=(-0.00, 0.00, -0.00))
└──          Compound at 0x7b341a5abdc0, Location(p=(2.00, 0.00, 0.00), o=(-0.00, 0.00, -0.00))
    └──      Solid    at 0x7b341a5abe20, Location(p=(0.00, 0.00, 0.00), o=(-0.00, 0.00, -0.00))

<OCP.TopoDS.TopoDS_Solid object at 0x7b341a5ded70> Location: (position=(2.00, 2.00, 0.00), orientation=(-0.00, 0.00, -0.00))

The resulting Solid is in the correct Location.