like the assumption with v1 pytest is that you can always

import pdb; pdb.set_trace()

because the pytest process will always have the console -- v2 gives us (maybe) parallelism, but modelling the real constraints probably requires ensuring each pytest subprocess has exclusive access to the console, which means they wouldn't be able to run in parallel. and it probably wouldn't even apply for remote execution? not sure quite how to configure all of that but seems fun and cool

replied on the other thread