[Numpy-discussion] copy/deepcopy pull request, Travis build failure

David Warde-Farley wardefar at iro.umontreal.ca
Thu Oct 25 17:48:26 EDT 2012

I submitted a pull request and one of the Travis builds is failing:


Given my changes,


I don't see how the masked array module (responsible for the failing
test) could possibly be affected in this manner (RuntimeWarning raised
by encountering an invalid value in power()).

Moreover, I cannot reproduce it after setting NPY_SEPARATE_COMPILATION
on my own machine (whatever that does -- I'm not quite clear).

Does anyone have any insight into what's going on here? Can anyone
else reproduce it outside of Travis?


More information about the NumPy-Discussion mailing list