#1121-updateBound() call for PQTorus geometry updates

[louhy]

Proposed fix for issue #1121. Since it's visual not sure if there's any good options for a test, but we'll have at least one which uses this shape when I update the pending physics GImpact shape test.

[stephengold]

For testing, how about using it in a Geometry, reading the bounding volume, and comparing against the expected bounds?

[louhy]

Went a bit overkill on this one... world bounds tests added for pretty much all shapes (seemed weird to just do it for PQTorus). Test fails without fix, succeeds with fix.

[stephengold]

Very nice. Thank you.

[stephengold]

How about stricter tests of the bounding volumes? Either verify that all the vertex positions are within the bounding volume, or check the bounds against hand-calculated minima, or both.

[louhy]

See how that one looks, I did make a few assumptions but I believe they're safe. Should be a more useful test like this I think. Put it together pretty quick, so double check that it makes sense. Reversing the vertex component order to 2,1,0 made it fail as I expected but I didn't check much else.

[stephengold]

Looks good to me. Unless there's some objection in the next day or so, I'll integrate this PR.