Floating-Point Precision Issues =============================== Upstream says they probably will not support any architectures besides amd64: https://github.com/flexible-collision-library/fcl/issues/474 Other architectures seem to suffer from various amounts of floating point precision issues. For now, we keep building libfcl on all Debian architectures, but be advised that in certain edge cases, the library may return wrong results -- Timo Röhling Wed, 18 Nov 2020 23:01:40 +0100