Incorrect bitwidth when scope 0 int

[nmacedo]

Currently, when the scope for ints is set to 0, integer atoms are not created in the universe but Kodkod is still passed a bitwidth that allows arithmetic operations as long as they are not cast back into tuple set. This bitwidth is the one needed to calculate the cardinality of the universe.

Although I have my doubts about this behaviour in general, at the moment this bitwidth is not being correctly calculated (it's using log base e rather than log base 2).

org.alloytools.alloy/org.alloytools.alloy.core/src/main/java/edu/mit/csail/sdg/translator/A4Solution.java

Line 428 in d1b71cf

solver_opts.setBitwidth(bitwidth > 0 ? bitwidth : (int) Math.ceil(Math.log(atoms.size())) + 1);