parsing: strange behaviour, bug?

Hi all, ; data is a file of about 30 Mo ; data is an ascii format like ABCDEF AVGJHJ ... POIVGA data: read %data ; this should give me all tokens but the only thing done is something like a memory leak parse data none ; making some rules allow me to parse data ; the memory use double at the beginning (60 Mo and then end at 30 Mo) parse data rules what can happen exactly? is it a bug? These tests are done both on NT, XP and Solaris OS with same results, and hardware is at least 512 Mo RAM. Thank you for any feedback, GreG.