Update v2.8.1 (Pyarrow,Flask,Boto3)

[rxm7706]

Flask <4 https://github.com/mlflow/mlflow/blob/v2.8.1/requirements/core-requirements.txt#L7 pyarrow <15 https://github.com/mlflow/mlflow/blob/v2.8.1/requirements/core-requirements.txt#L16 Pydantic <3 https://github.com/mlflow/mlflow/blob/v2.8.1/requirements/gateway-requirements.txt#L5 Boto3 <2 https://github.com/mlflow/mlflow/blob/v2.8.1/requirements/gateway-requirements.txt#L10

Checklist

• [x] Used a personal fork of the feedstock to propose changes
• [x] Bumped the build number (if the version is unchanged)
• [x] Reset the build number to 0 (if the version changed)
• [x] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• [x] Ensured the license file is being packaged.

AttributeError: 'MinimalDistribution' object has no attribute '_referenced_files' Appears to be related to setuptools <60 cap.

[conda-forge-webservices[bot]]

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.

[rxm7706]

@conda-forge-admin, please rerender

[rxm7706]

AttributeError: 'MinimalDistribution' object has no attribute '_referenced_files'

Failures with linux_aarch64_mlflow_variantdefaultpython3 linux_aarch64_mlflow_variantskinnypython3

linux_ppc64le_mlflow_variantdefaultpython3 linux_ppc64le_mlflow_variantskinnypython3

osx_arm64_mlflow_variantdefaultpython3 osx_arm64_mlflow_variantskinnypython3

All variants were working fine in 2.8.0 https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=813657&view=results

[rxm7706]

@conda-forge-admin, please rerender

[github-actions[bot]]

Hi! This is the friendly automated conda-forge-webservice.

I tried to rerender for you, but it looks like there was nothing to do.

This message was generated by GitHub actions workflow run https://github.com/conda-forge/mlflow-feedstock/actions/runs/6890158067.

[BenWilson2]

@rxm7706 It looks like the setuptools pin is overridden in package install? https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=824883&view=logs&j=ab84e6e6-af01-55f5-f9b2-33f6b533d940&t=d401fd58-9bfe-556c-7c0f-d8d15fa93c97&l=126

According to this thread: https://github.com/pypa/setuptools/issues/3743 versions >=65.6.0 have separate stubs that have some unpredictable patching behavior with distutils.

Why does the 2nd package plan generation ignore the setuptools pin that you introduced?

[rxm7706]

Why does the 2nd package plan generation ignore the setuptools pin that you introduced?

Hi @BenWilson2 - I am not sure what's going on here, Might need @conda-forge-admin help here.

I was actually watching MLFLOW - because of the Pyarrow CVE - because we ran into https://nvd.nist.gov/vuln/detail/CVE-2023-47248 - that flagged MLFLOW.

I'll take a look & try a few more things - but I don't believe its related to the recipe package version upgrades.. Everything worked just fine with your last merge a few weeks ago - and the failures are all within arm64, ppc64le & aarch64 - linux_64, win_64 and osx_64 work without an issues -- which makes it even harder to debug locally..

Thank You.

Last successful build on all archs.. https://github.com/conda-forge/mlflow-feedstock/runs/18164745795

[rxm7706]

@conda-forge-admin, please rerender

[github-actions[bot]]

Hi! This is the friendly automated conda-forge-webservice.

I tried to rerender for you, but it looks like there was nothing to do.

This message was generated by GitHub actions workflow run https://github.com/conda-forge/mlflow-feedstock/actions/runs/6903520000.

[BenWilson2]

@rxm7706 looking good :) The only other thing we can do is remove psutil as a requirement where it appears. It's now an optional import for MLflow (we changed that in the 2.8.1 release). Aside from that minor change, I think we can use this branch for the 2.8.1 release!

[BenWilson2]

@xhochy Could you take a look at the changes here and let us know if there are any potential issues with making these modifications?

[xhochy]

The openssl pin in host sounds really awkward. That shouldn't be there.

[xhochy]

@conda-forge-admin please rerender

[rxm7706]

The openssl pin in host sounds really awkward. That shouldn't be there.

Thanks @xhochy ! @BenWilson2 - I think this is GTG - with all the fixes and corrections from @xhochy Regards.