When I run coverage run setup.py test and then coverage report, the output properly reports the coverage at 88% and shows the missing lines, but it exits with a status of 0, where it should fail. If I add --fail-under=100` flag to the command line, it properly fails.
Here is my .coveragerc file:
branch = True
source = querybuilder
# Have to re-enable the standard pragma
pragma: no cover
# Don't complain if tests don't hit defensive assertion code:
show_missing = 1
@doismellburning I'm sorry, but this change, as small as it seems, depended on some deeper changes that are not in 3.7.1. Also, it's difficult for me to keep two development streams going on a side project.