conda-forge core meeting 2023-08-23

Add new agenda items under the Your __new__() agenda items heading

• Zoom link
• What time is the meeting in my time zone
• Last week's meeting

Attendees

Name	Initials	GitHub ID	Affiliation
Jaime Rodríguez-Guerra	JRG	jaimergp	Quansight
Dave Clements	DPC	tnabtaf	Anaconda
Filpe Fernandes	FF	ocefpaf	conda-forge
Cheng H. Lee	CHL	chenghlee	Anaconda/cf
Matthew R Becker	MRB	beckermr	cf
Marcel Bargull	MB	mbargull	Bioconda/cf
Katherine Kinnaman	KK	kathatherine	Anaconda
John Kirkham	JK	jakirkham	NVIDIA/cf

17 people total

Standing items

• [ ]

From previous meeting(s)

• [ ]

Active votes

• [ ]

Introductions

• Ryan Keith (AnacondaInc)
• James Loy (AnacondaInc)

Your __new__() agenda items

• (HV) Drop MacOS <10.13
  • Announcement PR; TODO list, context & discussion. Ecosystem is forcing us to move on (libcxx-trick doesn't help anymore); ≪1% users affected
  • Blocking abseil & grpc migration already, soon also libcxx 17
  • Do we want to formulate a more specific support policy? (10.13 is EOL for 3 years already; CPython might jump to 10.15/11.0)
    • (CHL) Could consider following Apple's lifecycle, but be aware that may force us to drop osx-64 before we're ready to. (Expect ~3 yrs left on Apple's support for such hardware)
    • (JK) If we do a NEP-29 style annoucement, we should align with other projects to make sure conda/conda-forge is not the odd ecosystem out.
  • MB: macOS 10.14 (Mojave) is the last version to support 32-bit (https://support.apple.com/en-us/HT208436)
  • MB: We could point people with old Macs to https://github.com/dortania/OpenCore-Legacy-Patcher or the like
  • Action items:
    • Someone check the announcement PR to make sure we have all the technical details correct
    • Add __osx>=10.13 as run_exports to the appropriate compilers packages <- actually a clobber file
      • Docs: https://docs.conda.io/projects/conda-build/en/stable/resources/variants.html#partially-clobbering-recipes
      • I think actually we want recipe_append.yaml (https://docs.conda.io/projects/conda-build/en/stable/resources/variants.html#appending-to-recipes)
      • This will add to run, not clobber it.
      • issue here: https://github.com/conda-forge/conda-forge-ci-setup-feedstock/issues/260
• (HV) Scipy for py312 on win will need a fortran compiler (that's not gfortran)
  • Switch to meson requires uniformity of compilers that invalidates current MSVC + gfortran approach
  • Either flang, lfortran (not yet ready), ifx (waiting for Intel follow-up)
  • Been working on flang for half a year, could use some help
  • This will very likely block 3.12 migration (on windows at least)
• (JRG) Miniforge and Mambaforge are now identical (both ship conda-libmamba-solver and mamba).
  • Note: conda will switch to conda-libmamba-solver as default in the next release (23.9).
• (JK) CUDA 12 arch migration
  • https://github.com/conda-forge/conda-forge-pinning-feedstock/pull/4799
  • Merge! Merge! Merge!
• (JK) CUDA 11.8
  • https://github.com/conda-forge/conda-forge.github.io/issues/1981
  • John will open a migrator PR
• (JK) GLIBC 2.28
  • Is alma linux a GLIBC option?
  • MRB: WE are in a holding pattern until dust settles.
  • CDTs: we will try and drop as many as possible for cos8
  • issue: https://github.com/conda-forge/cdt-builds/issues/66
• (JK) Windows ARM
  • Getting a machine is challenge.
  • Physical machine is an option. IF we do that it goes to Isuru. This is for debugging, not CI
  • OK! Going to Isuru.
• (MRB) repodata patches now use yaml files - yay yaml! - i'm moving things over bit by bit, fixing bugs in the process
• (JRG) Repology.org integration
  • https://github.com/repology/repology-updater/issues/518
  • https://github.com/conda-forge/conda-forge.github.io/issues/593

Pushed to next meeting

• (HV) [bikeshedding at the end] -dev vs. -devel
  • came up in boost unification, current PR uses the latter based on Isuru's rationale
  • matches Anaconda naming & CDTs, does not match recent CUDA feedstocks, tangentially related to distro discussion (RHEL vs. Debian). We should try to choose one.
  • JRG: Our own data:
    • -dev suffix: 33 packages
    • -devel suffix: 10 packages
• (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
• (JK) Start using conda's libmamba solver
  • https://github.com/conda-forge/conda-forge-ci-setup-feedstock/pull/259
  • https://github.com/conda-forge/docker-images/pull/240

CFEPs

• [ ]