Next month, I’m toying with the idea of bumping the test runner to v2. Vader (our testing framework) provides g:vader_report which stores the contents of Vader’s test results buffer, and g:vader_errrors which stores the list of failed tests. So we could identify which tests failed and retrieve it from g:vader_report. To grab the test code, Vader’s parser exposes a parse function that retrieves dictionaries representing each test without running them. So between these three things, we could report out a v2 JSON without a lot of fuss.
2 Likes