In 2.7.1, 2.7.2, and 2.7.3, the following model:
var -100..100: a;
var -10..10: c = 2 * a;
flattens to the following flatzinc:
var -100..100: a:: output_var;
solve satisfy;
which is incorrect. The flattened model has a = -100 as a solution, but this would have c = -200 which violates the bounds on c in the original model.
In 2.7.1, 2.7.2, and 2.7.3, the following model:
flattens to the following flatzinc:
which is incorrect. The flattened model has
a = -100as a solution, but this would havec = -200which violates the bounds oncin the original model.