Currently, JTR only terminates browsers when all tests are done in all requested browsers.
It's not a problem for our CI as we only run JTR on a single browser at a time, but it makes local test runs much more annoying.
For example, running the following in jQuery:
npm run test:unit -- -v -c jtr-isolate.yml --browserstack 'IE_11' --browserstack 'Chrome_latest' --browserstack 'Firefox_latest' --browserstack 'Safari_latest'
usually leads to modern browsers finishing quickly and waiting while IE is running tests. Since they wait for a long time, the browsers are idle for too long and... JTR (or something under the hood) thinks it's because the browser got stuck, restarts the browser and re-runs the last module tests on it (here: tween).
One possible solution to this would be to fix this detection, but needlessly keeping the browser running in idle mode is also not great; just terminating the browser when it's done would solve both issues at the same time.
Currently, JTR only terminates browsers when all tests are done in all requested browsers.
It's not a problem for our CI as we only run JTR on a single browser at a time, but it makes local test runs much more annoying.
For example, running the following in jQuery:
usually leads to modern browsers finishing quickly and waiting while IE is running tests. Since they wait for a long time, the browsers are idle for too long and... JTR (or something under the hood) thinks it's because the browser got stuck, restarts the browser and re-runs the last module tests on it (here: tween).
One possible solution to this would be to fix this detection, but needlessly keeping the browser running in idle mode is also not great; just terminating the browser when it's done would solve both issues at the same time.