Instead of the process of reading in backend (herd/litmus/etc) results and then comparing them to each other being tangled up in the tester, I'm planning to separate them and introduce a new output format that serves as a backend-agnostic representation of backend runs. To start with, this'll just contain a straight state set and an 'is undefined?' flag, but will eventually support postconditions and whatnot.
This amounts, mostly, to exposing what is already in the Act_sim.Output.t through yojson.
Instead of the process of reading in backend (herd/litmus/etc) results and then comparing them to each other being tangled up in the tester, I'm planning to separate them and introduce a new output format that serves as a backend-agnostic representation of backend runs. To start with, this'll just contain a straight state set and an 'is undefined?' flag, but will eventually support postconditions and whatnot.
This amounts, mostly, to exposing what is already in the
Act_sim.Output.t
through yojson.