On Sun, Sep 20, 2020 at 06:23:41PM +0200, Solene Rapenne wrote: > I don't really care about the port itself, but I think we need to > define a clear policy about python2 removal crusade.
> I get that python2 is end of life. We have lot of python2 ports that > are python2 only. Do we remove python2 in X months/year and projects > that still use python2 will have to be deleted? Do we try to delete > every port requiring python2 ? At which cost? > If we agree that python2 must disappear for 6.9 release, I would > understand the choice of deleting python2 ports. > I am always ok to delete ports that are broken or network related > and abandoned for safety reasons. But if a port works and doesn't > seem to be a security risk to run I don't see strong reasons to > remove it. Well, my thoughts are this. There are ports that are python2 that *are* very useful and needed. sthen has pointed out that Ubuntu and Red Hat have committed to providing security updates to Python 2.7 for at least a few years. So, those python2 ports we _need_, we keep. However, many python libraries are dropping python2 support. So, for me, right now, if we don't _need_ a port that is python2-only (or a python2 FLAVOR), I'd like to get rid of it, to minimize the rot in the ports tree that needing to hold back updates to support python2 causes. Yes, we can start doing special python2 versions of ports, etc. But, if something is clearly not maintained upstream, upstream recommends moving on to something else, it has no consumers, and there isn't a compelling need for it, I'd like to remove it. I'm slowly removing unneeded python2 stuff so when/if we _do_ eventually decide to remove python 2.7, it will be easier. Also so the maintenance of the rest of the python ports is easier. The Python ports tend to suffer from lack of maintenance anyway, so why not make it easier? --Kurt
