I want to exclude a residue from XYZ refinement, but I noticed that my atom selection "not (chain A and resid 100)" is ok for reciprocal-space, but not real-space?

Someone encounter a similar problem 10 years ago: http://www.phenix-online.org/pipermail/phenixbb/2012-December/019297.html

Is it a known bug? Any clues?

Thanx in advance