conda-forge core meeting 2023-12-13
Add new agenda items under the Your __new__() agenda items
heading
Attendeesβ
Name | Initials | GitHub ID | Affiliation |
---|---|---|---|
Dave Clements | DPC | tnabtaf | Anaconda |
Daniel Ching | DJC | carterbox | Argonne |
Marcel Bargull | MB | mbargull | Bioconda/cf |
Filipe Fernandes | FF | ocefpaf | conda-forge |
John Kirkham | JK | jakirkham | NVIDIA/cf |
Matthew Becker | MRB | beckermr | cf |
10 people total
Standing itemsβ
- [ ]
From previous meeting(s)β
- (JK) Miniforge 23.10
- Are we ready to make this release?
- We think all outstanding issues have been resolved
- Yes, we are ready to make the release.
- JK will follow up with Mark to get this moving.
- (JK) CUDA 11.8
- 75% of feedstocks are migrated
- 20% in PR
- 5% are waiting, backlogged, or erroring
- Think we are ready to close out this migration.
- 95% is a good number.
- We will close out this migration.
- 11.2 is in early stages of going away.
- CUDA 11.2 Docker images planned for removal May 2024
- 11.8 can use the same driver version as 11.2 so migrating to 11.8 should be pretty easy for package users.
- https://docs.nvidia.com/cuda/archive/11.8.0/cuda-toolkit-release-notes/index.html#cuda-major-component-versions (please see "Version Compatibility" table)
- More details in the compatibility guide: https://docs.nvidia.com/deploy/cuda-compatibility/index.html
- Cloud providers all support 11.8
- Kaggle is a good lagging indicator (other CSPs move a bit sooner)
- Upgraded to CUDA 11.8 in May 2023
- https://github.com/Kaggle/docker-python/pull/1254/commits/9e00acf188b781000ec2d7763b09d9b03883cfe7
- Kaggle is a good lagging indicator (other CSPs move a bit sooner)
- Thoughts on when we should drop 11.2?
- No strong objections
- Will propose an issue on 11.2 to discuss when we are ready to drop it.
- (JK) CUDA 12.x
- CUDA 12.0 Migration halfway through most packages, another third in PR.
- Planning to start adding CUDA 12.x (x > 0) packages
- Not planning migrations for CUDA 12.x at this time
- Discussion about whether packages constraints are sufficient
- (JK) Conda + libmamba
- This is a month old.
- These issues have been resolved.
- (MB) Confirmed resolved.
- (JK) Public visibility of Alma images on Quay
- Can't see them on the Quay website. You can pull them.
- Is there a setting we need to set.
- This has been fixed by IF
- (HV) Archive k* ecosystem (see last comment here, has five +1's from core)
- dead as a doornail, constant headache for migrations
- archiving is reversible, so let's finally bite that bullet?
- Can leave instructions in feedstock README (or a pinned issue) if someone comes along who wants to revive; however unlikely that is...
- they have on the order of 1000 downloads. They are real dead, but they constantly show up and block migrations.
- HV: We should bite the bullet because no one is going to arrive and update these packages.
- Archiving can be undone easily.
- FF: we could implement "Packages that need help!"
- These were archived, not because they are bad, but because they just weren't being maintained.
- HV: Yes. We can automate this.
- HV will try, but will ask for support as needed.
- (HV) Migration for
error_overlinking: true
?- already being set for new feedstocks in staged-recipes, should roll out to existing ones too (eventually).
- would be a good opportunity to do
{{ stdlib }}
-related changes (e.g. remove implicit run-export to C/C++ stdlib --> must be specified in recipe,error_overlinking
will find missing instances; if not necessary, package dependencies get slimmed by migration π₯³) - JK: Documentation in conda-build is a bit spotty
- MB:
- Apply to existing feedstocks.
- Maybe do this incrementally (opt-in like arch-migrator)?
- HV: Yes. Can roll this out slowly.
- HV will start an issue.
Active votesβ
- [ ]
Your __new__()
agenda itemsβ
- (MRB) libxcrypt
- (HV) Close boost 1.82 migration and archive old boost-cpp feedstock
- YES, DO THIS
- (DPC) Our next call is Dec 27.
- Should we cancel?
- THERE WILL NOT BE A MEETING ON THE 27th.
- (WV) Do we have any technical debt to take care of? (MRB: LLOL)
- There is another company that helps with technical debt on open source projects.
- Is there anything that can be fixed in a reasonalbe amount of time
- Reasonable: maybe 2-4 weeks.
- Part of Sovereign Tech Fund. Company in Berlin, with about 10 people.
- Reach out to Wolf if you have ideas.
Pushed to next meetingβ
- (DPC) Google Summer of Code 2024
- Advice from NumFOCUS for 2024:
A key lesson learnt from last year is that Googleβs approach to selecting projects is not well suited to umbrella organizations like NumFOCUS. This year, we heavily recommend projects which can apply on their own to do so. We encourage you to explore and propose projects in the AI/ML space. We also expect more potential GSoC contributors to be looking for AI/ML related ideas.
- From Google
We are looking forward to having more AI/ML open source organizations involved in GSoC 2024
- Is there interest in a conda-forge for GSOC 2024?
- Is there interest from conda-forge in a conda-wide application?
- Or would that be an umbrella organization? π
- Advice from NumFOCUS for 2024:
- (DPC) Conda-Forge off of twitter now, damn it.
- (JK) Almalinux 8
CFEPsβ
- [ ]