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
As a sidenote first you guys don't test on travis against latest numpy :)
When testing the pickle loading the default behaviour changed in numpy and breaks the test_joblib_pickle_across_python_versions test.
The most important change is in the numpy.load behaviour as inlined bellow:
[ 379s] Allow loading pickled object arrays stored in npy files. Reasons for
[ 379s] disallowing pickles include security, as loading pickled data can
[ 379s] execute arbitrary code. If pickles are disallowed, loading object
[ 379s] arrays will fail. Default: False
[ 379s]
[ 379s] .. versionchanged:: 1.16.3
[ 379s] Made default False in response to CVE-2019-6446.
The text was updated successfully, but these errors were encountered:
As a sidenote first you guys don't test on travis against latest numpy :)
When testing the pickle loading the default behaviour changed in numpy and breaks the test_joblib_pickle_across_python_versions test.
The most important change is in the numpy.load behaviour as inlined bellow:
The text was updated successfully, but these errors were encountered: