Remove forced config file default of .coveragerc #508
Conversation
|
@ionelmc @graingert Hello! When you get a moment can you please take a look at this? I can't figure out why the failures are occurring. |
|
Ooof sorry for delay, CI approved now. At first glance it seem fine. Please also update AUTHORS/CHANGELOG.rst |
|
Done! I don't know how to fix the tests though. |
|
Can I please get help with this? |
|
Hello, sorry for tardy response - can you rebase on the master? The CI was cleaned up. |
|
Done! |
|
I'm not sure how to fix |
|
I still can't figure it out |
| @@ -98,7 +102,7 @@ def set_env(self): | |||
| if os.path.exists(config_file): | |||
| os.environ['COV_CORE_CONFIG'] = config_file | |||
| else: | |||
| os.environ['COV_CORE_CONFIG'] = os.pathsep | |||
| os.environ['COV_CORE_CONFIG'] = '' | |||
There was a problem hiding this comment.
I believe this change is the cause of all the regressions.
Note that we have this in embed.py and you need to keep it working:
if cov_config == os.pathsep:
cov_config = True
src/pytest_cov/engine.py
Outdated
| @@ -70,6 +70,10 @@ def __init__(self, cov_source, cov_report, cov_config, cov_append, cov_branch, c | |||
| self.topdir = os.getcwd() | |||
| self.is_collocated = None | |||
|
|
|||
| # If unset (the default), indicate fallback behavior for other files like pyproject.toml. | |||
| # See https://github.com/nedbat/coveragepy/blob/6.2/coverage/control.py#L144-L146 | |||
| self.config_file_choice = self.cov_config or True | |||
There was a problem hiding this comment.
Also I'd simply rename this to config_file. Maybe it's just me but a dropdrown widget pops up in my head when I see "choice" - too much Django I guess xD
|
Can you please help? |
|
@ofek this looks old and needs conflict resolution. Try rebasing once more? |
|
CI was failing and I don't think the maintainer has time for this unfortunately so there's no point. If you really want me to I will but I feel like it's a waste of time. |
|
Sorry, I know this feels bad, having to rebase over and over. But I did point in code commends at the potential cause of the regressions. In hindsight I should have asked you to describe more clearly what this PR is supposed to do and help figure out a good solution instead of reviewing a solution that I am not sure what is supposed to fix. Maybe that would had gotten things moving faster. Can we restart this? Is there a problem with pytest-cov loading the coverage configuration from a pyproject.toml? |
There has been support for
pyproject.tomlever since 5.0 nedbat/coveragepy#664 (comment) and based on https://coverage.readthedocs.io/en/6.2/config.html it will automatically find the right file (logic here: https://github.com/nedbat/coveragepy/blob/6.2/coverage/config.py#L493-L519)I tested this locally and now config can properly be read from the other fallback files