I'm OK to keep the uv and non-direct dependencies experimentally for a while, but we should at least add a comment to the dependency saying why it is there, and the current issues we have with uv. ~Also if it is not much work it would be nice if it were actually used in the CI, so we can also see more in practice how much faster the CI could be with it.~
What's needed?
uv
is not really being used, so we shouldn't depend on it (developers that want to use it can install it manually on their envs).The non-direct dependencies should go, I guess they were added because of
uv
?. See:https://github.com/frequenz-floss/frequenz-dispatch-python/pull/9#discussion_r1539263157
I'm OK to keep the
uv
and non-direct dependencies experimentally for a while, but we should at least add a comment to the dependency saying why it is there, and the current issues we have withuv
. ~Also if it is not much work it would be nice if it were actually used in the CI, so we can also see more in practice how much faster the CI could be with it.~Proposed solution
No response
Use cases
No response
Alternatives and workarounds
No response
Additional context
No response