Strong reasons for the distribution were already discussed above, including:
* we don't have the ability/mapower to remove everything at once
Won't it take more total effort to remove things piecemeal rather than all in one go, though?
* by removing stuff gradually, we have longer time period to fix any
issues that it brings
My understanding was that it's now too late to remove things from Fedora 28, and we want Python 2 gone by the time Fedora 30 is released. That only gives us Fedora 29 as wiggle room, which doesn't seem all that gradual to me.
In any case, once we start removing Python 2 components, it seems to me that the message to users is, "Python 2 can't be relied upon in this release". That being the case, if we did go ahead with this staged removal, would it be helpful to think of this change the other way around? Rather than removing Python 2 from Fedora 30 but starting that work during Fedora 29, we're removing it from Fedora 29 but not completing all of that work until Fedora 30?
--
Peter Oliver