So, if you let Pants generate the setup.py for you (as described here) then Pants knows to rebuild A and B if C changes (instead of satisfying A and B from cache, say)

But the interesting question is how to know that you have to release them

So that depends on how you want it to work

For example, would you want a pants command that lists the distributions that depend on C?

I guess my question is what you want the workflow to look like.

Pants does have all the dependency data necessary to know that A and B depend on C, but how to use that depends on how you want the user workflow to be.

For example, how do you generate version numbers for the distributions?

I should also mention that using a tool like Pants means you often don't have to publish distributions at all - you can consume dependencies locally, in the repo. You should only need to publish if other repos need to consume these dependencies.

Happy to chat further, maybe on #general? The more we learn about your use-case the better the solution we can find.