// X is a FLOAT that starts at the whole number 1. At the top of the loop the // checker makes X an unknown of the sort its value had (a whole number), and // X is 0.5 on the second pass. // MUST-NOT-PROVE: assert@8 DECLARE X : FLOAT X <- 1 FOR I <- 0 TO 2 ASSERT X <= 0.0 OR X >= 1.0 X <- X / 2 NEXT I