Hi Drew, I think it is a good style to be really picky with unexpected warnings; even deprecation warnings. However, some dependencies are rather slow in updating their code to remove the deprecation warnings, and so some workaround is necessary. Probably @pytest.mark.filterwarnings would be better to handle this, but this is something for upstream. I wanted a minimally invasive, dedicated patch, due to the freeze.
Cheers Ole On 15.03.19 06:40, Drew Parsons wrote: > Source: skimage > Followup-For: Bug #924606 > > The skimage tests are doing something passing strange, if they're > giving an error when a Warning is no longer emitted. > > Anyway, thanks for uploading the patch that sorts it out, Ole. > > Drew
