We’ve discussed this before, but I think we always overcomplicated it re: dependabot.
I think a more generalized solution is that, in any CI run, if one of our python requirements files has been touched, we should re-generate the lockfiles and either a) commit, or b) fail out.
For dependabot, we should commit for sure - for non-dependabot, I’m a little less certain.
I can add this functionality as I start unwrapping the generate-workflows into separate functions.
We’ve discussed this before, but I think we always overcomplicated it re: dependabot.
I think a more generalized solution is that, in any CI run, if one of our python requirements files has been touched, we should re-generate the lockfiles and either a) commit, or b) fail out.
For dependabot, we should commit for sure - for non-dependabot, I’m a little less certain.
I can add this functionality as I start unwrapping the generate-workflows into separate functions.