Should we jump into Python 3.10 ?
For much I would like to start using the walrus operator, I don't think any of the new features of Python 3.8-3.10 would justify that we pushed the current boundary beyond 3.7. However, the new error information features of 3.10 seem to me like an excellent reason to upgrade. I'm especially looking forward to not seeing the "tuple is not callable" errors from typos in nipype workflows anymore, along with other similar issues giving really bad feedback to the user.
WDYT?
I would wait for 3.10.1. There are often critical bug fixes and the overall ecosystem support will be improved.
So maybe for 22.0?
Sure, definitely not thinking of putting this into the 21.0 release mix, and considering the date, there will probably not be a 21.1.
This was done in #3108.