You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In v3.7, if you constructed a coverage object like this: coverage(config_file=".coveragerc"), it would be OK with the file not existing.
In 4.0a4, if you construct a coverage object that way, it will raise an exception: CoverageException: Couldn't read '.coveragerc' as a config file. Constructing it with coverage(config_file=True) will work the same in both versions: it will be fine with the file not existing.
The nose-cov plugin uses cov_core, which always supplies a config_file option.
Should it always be OK for the specified config file to be missing? If the user uses coverage run --rcfile=xyzzy.ini, shouldn't it complain if that file doesn't exist?
Suppose you have no .coveragerc file.
In v3.7, if you constructed a coverage object like this:
coverage(config_file=".coveragerc")
, it would be OK with the file not existing.In 4.0a4, if you construct a coverage object that way, it will raise an exception:
CoverageException: Couldn't read '.coveragerc' as a config file
. Constructing it withcoverage(config_file=True)
will work the same in both versions: it will be fine with the file not existing.The nose-cov plugin uses cov_core, which always supplies a config_file option.
Should it always be OK for the specified config file to be missing? If the user uses
coverage run --rcfile=xyzzy.ini
, shouldn't it complain if that file doesn't exist?The text was updated successfully, but these errors were encountered: