Set default python version 3.9 and fix strenum problem#135
Set default python version 3.9 and fix strenum problem#135
Conversation
jacomago
requested review from
anderslindho,
ralphlange,
shroffk,
simon-ess and
tynanford
|
I tried to get this check in the pre-commit, but adding all ruff lints wasn't enough, I had to add type checking... Which made the PR huge. So I'll stick with just this change for now. |
jacomago
force-pushed
the
set-default-python-version-lower
branch
from
7dd8ea3 to
e3862b6
Compare
| version="1.5" | ||
| readme = "README.md" | ||
| requires-python = ">=3.6" | ||
| classifiers = [ |
There was a problem hiding this comment.
Do you not need 3.6 anymore? I'm happy to remove.
There was a problem hiding this comment.
Ah, I read "default" in the PR description as planning to get rid of 3.6-3.8
It's up to the community what to do. I think there's value in making life easiest on the people using/deploying software in these open source projects. So it is good to keep support for older versions when it's not too much hassle and I would say keep 3.6. But I am fine moving to 3.9 too (RHEL 9 default) if no one else wants 3.6 support.
There was a problem hiding this comment.
I'm wary of going beyond the python supported versions https://devguide.python.org/versions/. 3.9 is already end of life. Wouldn't be surprised if that starts breaking the ci soon.
There was a problem hiding this comment.
Eh, I think the hassle is worth it to help the lab setups where there might just be one person trying to maintain all this EPICS software. But just my opinion
There was a problem hiding this comment.
+1
At ITER - at 1 million records with less than 10 of ~160 plant systems delivered - we are struggling to migrate everything to RHEL 8 (that uses Python 3.6 as standard).
There was a problem hiding this comment.
Side-point, but you might want to have a look at conda(/mamba, etc.) @ralphlange
Re py3.6, I am a bit torn: there may be sites that are limited to it, but EOL'd more than 4 years ago, and is becoming increasingly costly to maintain support for; that is at least what we have seen for the projects where we have tried to do so.
There was a problem hiding this comment.
I know there are better solutions - it's simply not my decision to make.
Almost all ITER plant systems are in-kind contributions, and many take 5-10 years between prototyping and delivery. Which means we're getting deliveries based on versions/technologies of 5-10 years ago.
If supporting old versions is getting costly, drop it. We'll find ways.
Please don't drop it just because it's old.
There was a problem hiding this comment.
We discussed it today and said that we either will build (but not test) from 3.6 or else maybe set up a dedicated (limited support) release branch for 3.6.
jacomago
force-pushed
the
set-default-python-version-lower
branch
from
e3862b6 to
2f39dc5
Compare
jacomago
force-pushed
the
set-default-python-version-lower
branch
from
dfbfa02 to
3f12593
Compare
Co-authored-by: anderslindho <44849690+anderslindho@users.noreply.github.com>
jacomago
force-pushed
the
set-default-python-version-lower
branch
from
3f12593 to
7368eb7
Compare
|
5 participants
No description provided.