Repository navigation
1.18 Release Planning #19764
Description
I think it is time to start planning for 1.18. It would be great to release rather sooner than later since we had a bunch of non-trivial changes recently.
cc @JukkaL
| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
I think it is time to start planning for 1.18. It would be great to release rather sooner than later since we had a bunch of non-trivial changes recently.
cc @JukkaL
| Back | FazBrowse Home | New Git URL |
Activity
ilevkivskyi commented on Aug 29, 2025
Btw one thing worth discussing is what should we do about fixed format cache in this release:
I think the second option may be best, it would be good to know if anyone relies on JSON format in any way (especially taking into that hopefully fixed format cache will soon be default).
emmatyping commented on Aug 30, 2025
I would definitely announce it, ideally with a suggestion of how people can benchmark it/try it out. Might be useful to get more data on performance and it's always good to catch bugs early.
CoolCat467 commented on Aug 30, 2025
From 1.17.0 release planning, #18615 is still yet to be merged
sterliakov commented on Aug 30, 2025
I agree that announcing new cache mode as experimental gives the most benefit: we get more early testers while warning everyone that they use it at their own risk.
I would really appreciate it if #19347 could make its way into this release - it's a trivial 3-line bugfix diff for a problem with no viable workaround.
ilevkivskyi commented on Sep 2, 2025
Btw I just noticed we still have argparse.SUPPRESS for --allow-redefinition-new. I think we should announce it (as experimental) in 1.18, and deprecate --allow-redefinition in 1.19 (assuming all goes well). Although implementation of the former may be less polished, the logic looks much more natural/intuitive to me.
KevinRK29 commented on Sep 3, 2025
Hey, I'll be the release manager for this release. I'll be cutting the release branch by today and plan on releasing by mid next week!
KevinRK29 commented on Sep 4, 2025
I cut the release branch 1.18 from this commit acacc9eba333701efbc1beedfe87245d4743bf94. I can cherry pick any fixes as well
brianschubert commented on Sep 4, 2025
(that's acacc9e from Sept 1 for the lazy :-)
brianschubert commented on Sep 4, 2025
#19792 (comment) uncovered some silently failing tests from a new feature added in #19099. Ideally those tests / the feature should be fixed, but if that proves troublesome or would hold up the release, another option would be to revert #19099 on the release branch
KevinRK29 commented on Sep 4, 2025
Version bumped on master to 1.19.0+dev in 4007c7d
KevinRK29 commented on Sep 4, 2025
Hey @BobTheBuidler, would you be able to fix the issue in 06f97b2 that @brianschubert has mentioned in time for the planned release next week?
BobTheBuidler commented on Sep 4, 2025
I think merging #19217 will fix that, I'm working on wrapping it up as we speak
20 remaining items
ilevkivskyi commented on Sep 13, 2025
#19845 should be included in a point release (and we definitely need one).
BobTheBuidler commented on Sep 14, 2025
Can we make sure this one gets into 1.18.2?
mypyc/mypyc#1140
#19846
DanielNoord commented on Sep 15, 2025
@JukkaL Could you decide on whether python/typeshed#14701 is significant enough that it warrants a typeshed sync? It is problematic enough for that we can't adopt the new version so we would appreciate it 😄
cdce8p commented on Sep 15, 2025
Typeshed was actually synced earlier today in #19848. The PR is small enough that I think it could be cherry-picked as is.
JukkaL commented on Sep 15, 2025
We can also just cherry-pick python/typeshed#14708 to a point release using misc/sync-typeshed.py.
JelleZijlstra commented on Sep 15, 2025
mypyc/mypyc#1139 might need attention.
hauntsaninja commented on Sep 15, 2025
Summary post for 1.18.2
Things to cherry pick:
Unresolved issues:
BobTheBuidler commented on Sep 16, 2025
I added the run test @JukkaL requested on #19846 , should be good now
BobTheBuidler commented on Sep 16, 2025
Since we have a viable workaround shown on mypyc/mypyc#1139, and no diagnosis or fix yet, does it make sense to release a 1.18.2 to resolve the 4 things we've already addressed and just push a 1.18.3 once we have a real fix implemented? The 4 things we did fix all seem like pretty big deals.
Chainfire commented on Sep 16, 2025
Isolated #19376 as a cause of TypeErrors. Basic async inheritence broken. Somewhat surprised no test case caught this (every single one of our mypyc-using projects breaks on this)
Update changelog for 1.18.2 (#19873)
Update changelog for 1.18.2 (#19873)
KevinRK29 commented on Sep 19, 2025
1.18.2 has been released with the following commits cherry picked from here #19764 (comment)
ilevkivskyi commented on Oct 1, 2025
I think this one can be closed now, it is time to start thinking about 1.19 :-)