From time to time we are noticing that tests should be extended to cover "real" framework usage (e.g. as in #784).
An approach to pro-actively discover such cases could be to collect code coverage metrics on a framework when executing the benchmark requests, so we get a graphical view within the code what has been touched and what might be missing.
From time to time we are noticing that tests should be extended to cover "real" framework usage (e.g. as in #784).
An approach to pro-actively discover such cases could be to collect code coverage metrics on a framework when executing the benchmark requests, so we get a graphical view within the code what has been touched and what might be missing.