![](https://secure.gravatar.com/avatar/1fee087d7a1ca17c8ad348271819a8d5.jpg?s=120&d=mm&r=g)
Le 21/04/2021 à 13:34, Victor Stinner a écrit :
On Wed, Apr 21, 2021 at 1:24 PM M.-A. Lemburg <mal@egenix.com> wrote:
Isn't that an educational problem ? Adjusting reporting of warnings isn't all that hard:
A common practical problem is a project CI which pulls the most recent verisons of 3rd party dependencies and suddenly break if a new deprecation warning is raised by such project.
If a warning breaks CI, does it mean that the CI set non-default warnings flags (i.e. turn them into errors)? But we're talking about the default flags.
As a concrete data point, some packages in the scientific ecosystem have decided to use FutureWarning instead of DeprecationWarning, precisely because DeprecationWarnings are silent by default and users never see them. This is an ecosystem where users are often non-professional developers, and the fact that DeprecationWarning is hidden by default hits them hard, because they are not aware that they need to change something.
Regards
Antoine.