Thank you. But still I am not sure that the space of infinitely differentiable functions with compact support in X is a core for the Dirichlet Laplacian (If we suppose that the domain X is regular). If yes, this space should be dense in the domain of the laplacian with respect to the graph norm.