conda-forge core meeting 2023-10-04¶
Add new agenda items under the Your __new__() agenda items heading
Attendees¶
Name |
Initials |
GitHub ID |
Affiliation |
|---|---|---|---|
Dave Clements |
DPC |
tnabtaf |
Anaconda |
Daniel Ching |
DJC |
carterbox |
Argonne National Laboratory |
Marcel Bargull |
MB |
mbargull |
Bioconda/cf |
John Kirkham |
JK |
jakirkham |
NVIDIA/cf |
Eric Dill |
ED |
ericdill |
Anaconda/cf |
Filipe Fernandes |
FF |
ocefpaf |
conda-forge |
Carl Anderson |
CPA |
barabo |
Anaconda |
Matt Becker |
MRB |
beckermr |
cf |
15 people total
Standing items¶
[ ]
From previous meeting(s)¶
[x] (JK) NumPy 2.0 planning
https://github.com/conda-forge/conda-forge.github.io/issues/1997
https://github.com/conda-forge/conda-forge-repodata-patches-feedstock/issues/516
HV: Should be possible to only build against 2.x, result will be ABI-compatible with 1.2x
IF: It will not be ABI compatible if the package author changes
NPY_TARGET_VERSIONmacro. Need ways to ensure that it does not happen.
NumPy 2 coming out early next year.
What’s our timeline?
JK: Ask the NumPy devs to embed the NPY_TARGET_VERSION in the resulting binaries as a string?
Will continue further discussion.
[x] (JL) FYI the creation of a conda “build tools” team under conda governancy policy (still federated until team figures out team charter) for conda-build and hopefully other build tools, welcome to join:
[ ] https://github.com/conda/conda-build/issues/4698
Active votes¶
[ ]
Your __new__() agenda items¶
[x] (HV) Drop PyPy on aarch/ppc?
PyPy support is already stretched very thin, to the point that we considered dropping it (c.f. here), haven’t done a pypy3.10 migration yet, various high impact bugs remain, etc.
We have no CI on the feedstock for aarch/ppc (cannot cross-compile, travis times out), and manual builds put extra strain on maintenance, and sometimes just don’t happen (e.g. missing 7.3.12 & 7.3.13, the last two versions), meaning missed bugfixes etc.
Rather than risk PyPy support overall by overwhelming the PyPy devs (or demanding that someone from core do manual builds for every PR), I suggest to drop PyPy on aarch/ppc; usage numbers for those builds are on the order of 1 in 10’000 users. This is not great, but IMO more sensible vis-à-vis the very constrained PyPy-resources.
Isuru does these builds. This is not causing larger problems in CF, therefore, it’s up to maintainers (Isuru) to make this decision?
Axel: But it has been a problem for some downstream packages.
Axel: will report problems as they arise.
AI: If we want to drop PyPy ARM/PPC, raise an issue on the webpage repo or pinning repo to discuss
[x] (HV) Compiler bump: GCC 13 & Clang 16
PR is ready, relevant info in the OP.
Isuru: Clang 16 now, GCC 13 in a bit.
[x] (JRG) Intention to submit a CZI EOSS6 letter of intent with Wolf.
Rough draft: https://hackmd.io/m_dJizu9TWGRI8RprLO-Yg
Theme: the future of conda-build, transitioning to new tooling (
boa,rattler-build) and improved build reproducibility in conda-forge.Interested in knowing if this is a terrible idea.
Deadline: Oct 16th.
[x] (JRG) GPU CI status:
Passing CI for x64
cupyat https://github.com/conda-forge/cf-autotick-bot-test-package-feedstock/pull/466 with global cirun config (simpler conda-smithy changes)admin-requestsPR ready. Jaime to review / debug in the coming days.
[x] (IF) m2-recipes https://github.com/conda-forge/staged-recipes/pull/24112
[x] (WV) Sovereign Tech Fund contract
[ ] libsolv - Rust port
[ ] rattler-build conda-forge integration
[ ] conda-forge reproducibility evaluation
[x] (WV) Please don’t forget to vote for conda steering council :)
Pushed to next meeting¶
[ ]
CFEPs¶
[ ]