You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Jan 16, 2024. It is now read-only.
The setuptools, wheel, and pip packages are in the requirements.txt now. That's not ideal. As a first thought, what I think should happen is the update_requirements make target should do two pip freeze operations, one with and one without --all specified. Then do a diff (perhaps using comm -3 requirements.txt setup_requirements.txt or perhaps just a grep) to pull those three out, and save the three versions in "setup_requirements.txt". Either way, the dependabot action will need updated to check that new file as well as requirements.txt, and the first pip install can just do pip install -U -r setup_requirements.txt.
Possibly with the pulp-requirements.txt itself being mangled with a sed in the container instead of having the Makefile do a sed -i. This starts to get a tad complicated:
comm -3 \
<(/opt/pulp/bin/pip freeze -r <( sed '/following requirements/,$d' /opt/pulp/pulp-requirements.txt ) --local | sort) \
<(/opt/pulp/bin/pip freeze -r <( sed '/following requirements/,$d' /opt/pulp/pulp-requirements.txt) --local --all | sort)
The setuptools, wheel, and pip packages are in the requirements.txt now. That's not ideal. As a first thought, what I think should happen is the update_requirements make target should do two pip freeze operations, one with and one without
--allspecified. Then do a diff (perhaps usingcomm -3 requirements.txt setup_requirements.txtor perhaps just a grep) to pull those three out, and save the three versions in "setup_requirements.txt". Either way, the dependabot action will need updated to check that new file as well as requirements.txt, and the firstpip installcan just dopip install -U -r setup_requirements.txt.More explicitly...
Possibly with the pulp-requirements.txt itself being mangled with a sed in the container instead of having the Makefile do a
sed -i. This starts to get a tad complicated: