Chutzpah uses a simple heuristic to try to match test's to line numbers.
For example, given a qunit test file it will scan the filing recording the line number of each instance of "test(" and then after tests are run match the test run to the index of each "test(" call.
This is a very fragile implementation that fails in a couple scenarios.
- If you are using a framework that adds its own verbage for tests (like Pavalov.js) this matching completely fails.
If you wrap your test in a loop, Chutzpah will fail to give correct line numbers.
I have not thought about a more robust implementation yet, thoughts are welcome.