On 2019/09/24 17:42, Kurt Mosiejczuk wrote: > On Tue, Sep 24, 2019 at 04:43:32PM -0400, Daniel Dickman wrote: > > If we keep python2 for the next release it means we will have those > > ports out there in 2020 and many projects have announced that Jan 1 > > 2020 is going to be the end for python2 support. > > The main *Python* has announced that deadline. Plenty of projects > in the ecosystem have *already* dropped support for python 2.7. > pytest 5.x+ is Python 3 only. We've got a number of modules in > the tree that aren't the latest version because 2.x support was > dropped.
imo, if there's a good reason to keep the py2 version (used by something else in the ports tree, or possibly if it includes a useful compiled iextension) then split the port. but if the py2 version isn't really useful and is holding back an update, drop the py2 version. (when upstream stops supporting py2.x, I bet there will still be fixes done by various OS packagers, or maybe we could switch it to tauthon). > > That means if vulnerabilities are found next year, we will have to > > come up with fixes while upstream will end up diverging quite a bit > > from the last legacy python2 releases they will have made. > > I think that ship has already sailed. 6.6-beta was tagged weeks ago and the > release is not far away. Really, I think we should have a deprecation > notice in the 6.6 upgrade instructions about python 2 if we are thinking > of going that way. > > All that ignores that there are a *lot* of projects that haven't finished > their python 3.x support yet. OfflineIMAP's python 3 support is labelled > as "STALLED". Removing python2 at this point is basically setting fire > to large portions of the ports tree. absolutely. removing py2 from ports is really not an option at this time, there are way too many useful things that still need it.
