Closed dokempf closed 2 years ago
I would not consider this to be an accidental side effect, the directory structure seems to be more mature now. The executable is now also properly installed and reachable in $PATH
, so if you only need to run this on the command line you can just do this now:
$ pip install lcov_cobertura
$ lcov_cobertura ....
Thanks, that helps of course. Whether intentional or not, I would still consider it a regression. After all, both of these methods can be implemented to achieve backwards compatibility (e.g. by providing __main__.py
).
Yes, that's true. I first noticed the change because of breaking CI pipelines ;)
Yes, you cannot import a package name, and in this case you get exactly the same behavior if you try importing the module name:
python -m lcov_cobertura.lcov_cobertura
Converts LCOV coverage data to Cobertura-compatible XML for reporting.
Usage:
lcov_cobertura.py lcov-file.dat
lcov_cobertura.py lcov-file.dat -b src/dir -e test.lib -o path/out.xml
By default, XML output will be written to ./coverage.xml
Fixed by a07765117c23073e36cfd2e88e3793c6ae57045d and released in version 2.0.2 https://pypi.org/project/lcov-cobertura/2.0.2/
Between
1.6
and2.0
there has been a regression w.r.t. the command line entry point. Doingpython -m lcov_cobertura
would work in 1.6, but produce the following error in 2.0:Most probably an accidental side effect of the
setup.py
->setup.cfg
transition.