Towards overcoming the transitive-closure bottleneck: Efficient parallel algorithms for planar digraphs