Closed hrshdhgd closed 1 year ago
That’s not how it is supposed to work…
The constraints.txt
should never be updated manually; it’s strictly the output of PIP’s dependency resolution algorithm.
If there’s a Python package for which you would like to enforce some specific constraints, that should be done in the requirements.txt.full
file, e.g.:
semsql >= 0.3
the constraint will then be picked up by the update-constraint
action next time it is run (it can be run manually if there’s an urgent need to to create an ODK image with the new semql version), which will then update the constraints.txt
file, taking into account the specific constraint on the package.
(@matentzn You broke the build the last time you tried to update constraints.txt
, don’t you remember? :P )
It was not manually updated, Harshad ran the script; don't worry. I also immediately ran the workflow again afterwards!
My bad.
But in that case it’s strange that the previous execution of the action earlier today didn’t update semsql, but the manual invocation a few hours later did…
Indeed. I think what happened is that an OAK release today dealt with a bad dependency; not sure though
fixes #831