Won't comment directly on the compatibility window (although I lean toward
the status quo and just requiring upgrades from the latest patch version of
a major release), but I think I've also arrived at a point where I'm not
convinced running the Python upgrade tests pre-commit is an efficient use
of resources. It's not just the hardware required, but the fact that we
have to expend mental energy to glance over the couple flaky upgrade tests
that have nothing to do with our patch, etc.

Reply via email to