Last week Jan integrated the new atomic step with the enclosure interpreter and started working on the top-level strategy. The idea has been to use the same constraint propagation technique employed to intersect enclosures with mode invariants and event guards during event handling. By composing event guards with the enclosures computed for the free variables in the guard and contracting the domain of the resulting predicate it should be possible to contract time segments around sub-segements where an event may occur. The advantage with this approach is that much of the necessary infrastructure is already in place, but a disadvantage is that it generally cannot show that an event has happened with certainty.
