While discussing with Mick in an external discussion, he raised the fact that the project approach has always been to deprecate features/support in one major version and remove them in the following one. Due to the fact that 4.0 took so long, we did not respect that approach for the Windows scripts and it might make sense to do the same for python 2 as it is EOL but we should also consider that it might not be the best way of doing things.
A big part of the problem here is caused by the fact that we do not have any idea of when the next major release will be. It might be good to discuss to have more visibility on this point before deciding about our approach for python 2. I will fire a thread to discuss the plan for the future major release. On Tue, Jan 26, 2021 at 11:02 AM Benjamin Lerer <benjamin.le...@datastax.com> wrote: > According to some study I found here > <https://www.activestate.com/resources/white-papers/python-2-eol-survey-results/> > it seems that not all the organizations are ready regarding python 3. I am > not a python expert. Does anybody know what are the > practicalities/hurdles that users can face when upgrading and what is the > expected cost of keeping support for 2.7 until the next major? > > > > On Tue, Jan 26, 2021 at 10:39 AM Benjamin Lerer < > benjamin.le...@datastax.com> wrote: > >> We have a check in CQLSH that prevents it to start if the python version >> is lower than python 2.7. I am fine dropping support for python 2 as long >> as we make it explicits in the NEWS.txt and that we add a warning in CQLSH >> to notify python 2 users that it is not supported anymore and might not >> work properly. >> >> +1 >> >> >> >> >> On Tue, Jan 26, 2021 at 8:46 AM Sumanth Pasupuleti < >> sumanth.pasupuleti...@gmail.com> wrote: >> >>> +1 (nb) for dropping support for python2; I agree 4.0 major release is a >>> good time to do this, given python2 is already EOL. >>> >>> On Mon, Jan 25, 2021 at 2:00 PM Yifan Cai <yc25c...@gmail.com> wrote: >>> >>> > +1 nb. >>> > We probably also want to set a milestone to get rid of the python2 >>> > compatible code completely, if we are going in the direction that drops >>> > python2 support in 4.0 and retains the python2 compatible code. In 4.x >>> or >>> > 5.0? >>> > >>> > On Mon, Jan 25, 2021 at 9:24 AM Ekaterina Dimitrova < >>> e.dimitr...@gmail.com >>> > > >>> > wrote: >>> > >>> > > I support the idea, we are not removing python2-compatible code >>> > > +1 >>> > > >>> > > On Fri, 22 Jan 2021 at 15:14, Adam Holmberg < >>> adam.holmb...@datastax.com> >>> > > wrote: >>> > > >>> > > > As you may recall, CASSANDRA-10190 [1] introduced Python 3 support >>> for >>> > > > cqlsh. This change will be landing in 4.0. In the course of >>> development >>> > > and >>> > > > discussion spanning years, it was decided to retain support for >>> Python >>> > 2. >>> > > > In the meantime, Python 2 sunsetted (a year ago [2]). I hadn't >>> seen a >>> > > > discussion about whether we intend to carry on support for Python >>> 2, so >>> > > I'm >>> > > > raising one here. >>> > > > >>> > > > 4.0 is a major release and we have an opportunity to drop support >>> at >>> > this >>> > > > milestone. It has been mentioned that it will not be acceptable to >>> do >>> > in >>> > > a >>> > > > minor or patch release, so if it's not done for 4.0, we will need >>> to >>> > wait >>> > > > for the next major. I do understand that many in the project would >>> like >>> > > > majors on a more frequent interval post-4.0, but at this time we >>> don't >>> > > know >>> > > > when that will be. >>> > > > >>> > > > I advocate for dropping support ASAP. I expect that users should >>> not be >>> > > > inconvenienced by this -- I am not aware of a major distro that >>> has not >>> > > had >>> > > > python3 for years. Dropping python2 support does not mean that we >>> would >>> > > do >>> > > > work to rip out python2-compatible code, just that we wouldn't >>> > advertise >>> > > > support and any package requirements would be adjusted. We benefit >>> by >>> > > > removing the need to test multiple runtimes, and we wouldn't be >>> > concerned >>> > > > with fixing python2-specific issues that may arise on the EOL >>> runtime >>> > > [3]. >>> > > > >>> > > > I look forward to the discussion. >>> > > > >>> > > > -- >>> > > > Adam Holmberg >>> > > > e. adam.holmb...@datastax.com >>> > > > w. www.datastax.com >>> > > > >>> > > > [1] https://issues.apache.org/jira/browse/CASSANDRA-10190 >>> > > > [2] >>> https://urldefense.proofpoint.com/v2/url?u=https-3A__www.python.org_doc_sunset-2Dpython-2D2_&d=DwIBaQ&c=adz96Xi0w1RHqtPMowiL2g&r=Jad7nE1Oab1mebx31r7AOfSsa0by8th6tCxpykmmOBA&m=crUhrXldycHqWKNgLahLKHtKL6pfe5rjVY5WXeujMCc&s=kU1jF8FeM_Jy4RoyW51BF9BMZV8gnSfQIG_GAup5PiQ&e= >>> > > > [3] https://issues.apache.org/jira/browse/CASSANDRA-16400 >>> > > > >>> > > >>> > >>> >>
