Right now, the Clojure/Jaunt test suite does one round of clojure.test running which touches pretty much everything. It should be possible to refactor the test runner script to provide much better output, including per-test suite failures. In part this helps track the various parts of the Clojure/Jaunt codebase which literally have no tests.
This is likely to be a bit of an adventure, and may involve some classloader voodoo to create sandboxes during test loading and evaluation. I've tried to get this working a couple times before and found really interesting data dependency behavior between the test namespaces.
Right now, the Clojure/Jaunt test suite does one round of
clojure.testrunning which touches pretty much everything. It should be possible to refactor the test runner script to provide much better output, including per-test suite failures. In part this helps track the various parts of the Clojure/Jaunt codebase which literally have no tests.This is likely to be a bit of an adventure, and may involve some classloader voodoo to create sandboxes during test loading and evaluation. I've tried to get this working a couple times before and found really interesting data dependency behavior between the test namespaces.