Drilling down, the obvious difference is that coveralls shows python multi-line strings as misses. This seems wrong to me, since the entire multiline literal was evaluated in a statement which did execute. coverage.py does not report these as misses:
I have a python project using coverage.py, reporting to coveralls. There is a big discrepancy: coverage.py reports coverage of 96% coveralls reports coverage of 83%.
Drilling down, the obvious difference is that coveralls shows python multi-line strings as misses. This seems wrong to me, since the entire multiline literal was evaluated in a statement which did execute. coverage.py does not report these as misses:
How can I get the correct report from coveralls?