| Re: [paho-dev] Python - dropping support for 2.7 |
Roger Light <roger@xxxxxxxxxx> writes: > Python 2.7 went end of line in January 2020. I am proposing that the > Paho Python library finally drops support for Python 2.7 after 1st > January 2022, i.e the next non-bugfix release after this date would > not include support for 2.7. The python 2.7 EOL situation has been messy and drawn out. From my experience in pkgsrc, and also from writing a python program, the issue is that every library and program ought to have released a version that works with python 3, and thus we should be at a point where there is really no point in using 2.7. But, many places lagged in doing that. While the list is shrinking, it's still quite non-zero. In my case, I am using the NUT (network UPS tools) python module to access UPS state and report that via mqtt. NUT has a somewhat troubled maintenance status, and seems on the verge of a release after not having had one since 2017. I will try to have a look at the current repo and see if the nut module works with 3 or can be made to do so. Given that mqtt is used as glue, it would not surprise if various other random things out there have old python and people are using 2.7 to get the data into the modern world. I wonder how much pain keeping the codebase working with 2.7 is? I do think that we're getting closer to the 2.7 tipping point, where anything that needs 2.7 is just considered broken and there's no reason for anybody to accomodate it. We are way closer now than a year ago, where that didn't seem reasonable, but I don't feel like we (the wider open source community) have crossed that line. That doesn't mean it shouldn't be dropped in point cases. > Perhaps it should be called version 2.0 to celebrate the breaking > change. I am disinclined to to have a big bump for that. To me, dropping 2.7 is appropriate when either it's too painful to keep it or we believe ~nobody will notice or care.
Attachment:
signature.asc
Description: PGP signature