Create Pipfile.locks with the Python version that is used in the respective image

[jiridanek]

Aha, understood. When you updated Pipfile.lock before, you did not run whatever command you used on Python version 3.8, but you used a newer Python?

Originally posted by @jiridanek in #572 (comment)

[jiridanek]

From the discussions below, pip has --python option that can be specified to run resolution as if on a different Python version.

• pip download: python_version in requirements file and --python-version option does not work pypa/pip#11198 (comment)
• Allow changing the entire dependency resolution “context” (Markers, dependency resolution and cross compiling) pypa/pip#10050
• Various package-index filtering flags do not affect the environment markers pypa/pip#11664

It does not seem possible to put something into Pipfile itself to pin Python version.

[guimou]

It does not seem possible to put something into Pipfile itself to pin Python version.

Of course you can. Example: https://github.com/opendatahub-io-contrib/workbench-images/blob/8c1c0cf735c6754fdaa332f83056dc386549ff75/base/c9s/3.11/Pipfile#L14

It's in the requirements file that you cannot set it. But that should not be a problem, because you use the requirements file in an already installed Python environment, so you have already chosen your version.

[jiridanek]

@paulovmr mentioned that architecture could be culprit, if locking runs on arm64/aarch64 or amd64/x86_64