The additional semantic is that resources always need to be reset if they have been used: it isn't something the running test has a say in, as is the case for the dirty flag.
With that as a base, the old "clean on dirty" semantics can easily be built on top.
The additional semantic is that resources always need to be reset if they have been used: it isn't something the running test has a say in, as is the case for the dirty flag.
With that as a base, the old "clean on dirty" semantics can easily be built on top.