conda-forge / pypy-meta-feedstock

A conda-smithy repository for pypy-meta.
BSD 3-Clause "New" or "Revised" License
1 stars 8 forks source link

Update to 7.3.11; remove duplicated CI jobs #24

Closed h-vetinari closed 1 year ago

h-vetinari commented 2 years ago

I saw in #23 that each pypy version gets built 3 times (once per cpython version). I don't quite see where the rerender picks up a supposedly unpinned - python in meta.yaml, but the following cbc-hack would do the trick. Not saying it's pretty, just a proposal.

Edit: now that https://github.com/conda-forge/pypy3.6-feedstock/pull/99 has landed, update the version here.

Closes #26

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

h-vetinari commented 2 years ago

Ping @conda-forge/pypy-meta :)

conda-forge-webservices[bot] commented 1 year ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

h-vetinari commented 1 year ago

Updated to take into account #26

mattip commented 1 year ago

Close/reopen to rerun the travis builds

h-vetinari commented 1 year ago

@conda-forge-admin, please restart ci

h-vetinari commented 1 year ago

Travis is not picking up the restarts. As a feedstock maintainer, you should be able to trigger a rerun in the GH UI under the "Checks" tab.

h-vetinari commented 1 year ago

Ah, wait, just saw the following banner on travis:

Builds have been temporarily disabled for public repositories due to a negative credit balance. Please go to the Plan page to replenish your credit balance or alter your Consume paid credits for OSS setting.

h-vetinari commented 1 year ago

So travis is out of credits, but things seem to work fine in emulation. PTAL @conda-forge/pypy-meta, would be great to have the new pypy releases for work in scipy/grpc/arrow etc... :)

mattip commented 1 year ago

Thanks @h-vetinari

h-vetinari commented 1 year ago

OK, this is now causing some unexpected side-effects. Due to pypy having a newer python "version" (referring to the stdlib portion here) than we have in CPython, this now will update random 3.9 environments from CPython to PyPy:

The following packages will be UPDATED:

  python                          3.9.15-h4de0772_0_cpython --> 3.9.16-0_73_pypy

IMO, using PyPy should remain opt-in, and we shouldn't switch the python implementation under the feet of users. So perhaps we might need to mark those builds as broken for now?

CC @conda-forge/core

mattip commented 1 year ago

oy. Is there a way to prevent this?

mattip commented 1 year ago

CI passes after rerendering the 3.8.16 PR

mattip commented 1 year ago

It seems there will be cpython 3.9.16 very soon, and cpython 3.8.16 was recently made available. Going forward, can we add a test to the build here that will prevent this? Something like conda search and make sure the proper cpython build exists?

jakirkham commented 1 year ago

Yeah we should do something better here. Testing or avoiding this issue altogether. Raised issue ( https://github.com/conda-forge/pypy-meta-feedstock/issues/27 ). Please feel free to add more there