These tests are known to only be executed partially or not at all. So
we always get incomplete, missing, and sometimes flaky, coverage in
the test functions that are expected to fail.
This change updates the ``coverage.py`` config to prevent said tests
from influencing the coverage level measurement.
This allows for e.g. Jedi to infer types (it checks the name).
It was only used to support Python 3.5.0/3.5.1, where this is is not
available in the `typing` module.
Ref: https://github.com/davidhalter/jedi/issues/1472
Uses `TYPE_CHECKING = False` in `_pytest.outcomes` to avoid having to
work around circular import.
This should increase coverage for subprocesses, where previously
`source` paths were used only from the config file, but not the initial
`--source` argument.
- Skips pypy for coverage, reports only py37 to coveralls
- tox: allow for TOXENV=py37-coverage
- tracks coverage in subprocesses, using coverage-enable-subprocess, and
parallel=1
- removes usedevelop with doctesting to match `--source` being used with
coverage
- keep coveralls for now, used with AppVeyor
Sviatoslav Sydorenko
Ran Benita
Daniel Hahler
hugovk
Tyler Goodlet
Bruno Oliveira