Make the integration test project capable of running in parallel, with all usability implications.
Should better determine a degree of parallelism through the number of cores on the machine, a factor such as 1:3 might be a useful ratio to explore, or more generally, experiment with degree of parallelism, and leave it entirely configurable. Avoid Akka though... too much ceremony and moving parts in using and debugging it.
Make the integration test project capable of running in parallel, with all usability implications.
Should better determine a degree of parallelism through the number of cores on the machine, a factor such as 1:3 might be a useful ratio to explore, or more generally, experiment with degree of parallelism, and leave it entirely configurable. Avoid Akka though... too much ceremony and moving parts in using and debugging it.
First though explore how the scala community build works, in case they have anything reusable, as the tasks are similar. On first look that's a clunk of bash but better look more carefully. They use https://github.com/typesafehub/dbuild, the exact benefit need be delineated...
Tightly related to https://github.com/CANVE/extractor/issues/4