On Windows, the python file have to run with environment variable PYTHONIOENCODING=utf-8. We could not run the test without that environment variable on master.
The way to force it using utf-8 without that environment variable is to add # -*- coding: utf-8 -*- at the second line of python files.
-BEGIN VERIFY SCRIPT-
sed -i 's/#!\/usr\/bin\/env python3/#!\/usr\/bin\/env python3\n# -*- coding: utf-8 -*-/g' $(git ls-files -- '*.py')
-END VERIFY SCRIPT-
Tend to NACK. It seems easier to export PYTHONIOENCODING=utf-8
It seems easier to export PYTHONIOENCODING=utf-8
For me, the python file should be locale independent. There is no reason that we should run python script with some specific environment variable. Also, you can't guarantee that everyone run the python script from command line like someone can run it with double-clicking the python file. Furthermore, this would prevent encoding issue forever by using utf-8 explicitly. Shouldn't explicit better than implicit?
We require python3, which follows PEP 3120. Explicitly setting the encoding should only be needed when it is not utf-8.
Oh, it appears to be a mistake running the script with python2. I remember I met this issue before when I was testing #14007, but it seems that the problem is the ouput encoding, I'll try to solve it in another way. Closing this, sorry for annoying.