Another case of float imprecision leading to endless loop. INcreasing a bit 'noise threashold' seems to work OK. Not a regression, but might be nice to have in 2.78a.
14 KiB
14 KiB