Difference in running coverage vs coverage as a plugin
I have recently implemented
coverage to provide your awesome tool to people who use
The issue I am experiencing is that coverage running as a plugin is not reporting the same numbers compared to running coverage itself.
What is odd is that coverage as a konira add-on, reports coverage for lines in a method but not the method line itself (?).
The coverage implementation for
konira  borrows most of the code from the pytest-coverage plugin (as I knew this was a working approach).
I am following your documentation to enable and run coverage along with
konira, calling coverage.start() before the tests are run and then create the report.
To replicate this problem with konira's own tests:
konira -d cover
which konira -d && coverage report -i --include='*'
If you clone the konira repository  and install it in a virtualenv and run the above commands you should be able to see the difference.
I am not expecting you to debug my code :) But I think I am doing things correctly and I might be hitting a bug here. Any help is greatly appreciated.