I had a tremendously embarrassing failure in a sheet metal part designed in Solidworks due to a bug in the way bend allowance is calculated. The long story short is that the flat pattern developed by SW was about 10% dimensionally inaccurate which resulted in a rotating part binding up in the field and going blammo. This was years and years ago as a fresh young designer, so I reported the issue to my reseller and the response was essentially "yeah, you're gonna have to take that into account." I've since learned that this sort of idiosyncratic difference in calculating certain values, particularly things like bend radii and bend allowance, things like beam extensions on structural frames, etc, is par for the course for CAD packages. Pretty much every engineer I know has learnt this lesson the hard way.
I guess it's not really a bug per se, just hidden behaviour that you have to learn about and take into consideration.
I guess it's not really a bug per se, just hidden behaviour that you have to learn about and take into consideration.