Introduced an optimization where shadowed subgraphs are not queried. In this version (in contrast to the commented version), DataLeq#alwaysTrue is asynchronous, because it relies on StatixSolver#entails, which is asynchronous as well.
Before merging, there is one point of consideration. The non-internal DataLeq instances can be shared between units. That might that imply that their alwaysTrue method can be invoked from different threads at the same time (is that actually the case???). If that is the case, we might accidentally create multiple alwaysTrue instances, while we intend to create one. I don't think it gives invalid results, but it would be good to synchronize in such situations.
Introduced an optimization where shadowed subgraphs are not queried. In this version (in contrast to the commented version),
DataLeq#alwaysTrueis asynchronous, because it relies onStatixSolver#entails, which is asynchronous as well.Before merging, there is one point of consideration. The non-internal
DataLeqinstances can be shared between units. That might that imply that theiralwaysTruemethod can be invoked from different threads at the same time (is that actually the case???). If that is the case, we might accidentally create multiplealwaysTrueinstances, while we intend to create one. I don't think it gives invalid results, but it would be good to synchronize in such situations.