Replies: 2 comments
|
Hi @billcrook, thank you for sharing feedback! Yes, that's how it is set up today. Each project need to have dependencies added in their respective I have that as a long-term goal for Python too, but believe there's a couple of things that need some experimentation and trying out. I have chosen to implement Polylith components and bases (aka bricks) as Python namespace packages. Initially I had an idea to have them as editable dependencies (libraries), to define the dependencies for a brick in there. But I leaned towards namespace packages instead, to avoid versioning, conflicting dependencies (such as different versions of a library) between bricks. As now, the tooling support will analyze code and highlight any missing dependencies (the Another thing that I have in the back of my head long-term is to implement something in-between to what's in Clojure and the third-party dependencies in Python: a Polylith-specific toml-file (or similar) for components. Maybe that's a bad idea, not sure 😄 |
|
Great, glad you are thinking about it. For larger repos with many components and projects, I imagine the duplication of dependencies could get quite cumbersome.
My two cents, I would go as far as I could with standard dependency tooling first. IDE compatibility is important and a custom toml file would be a non-starter for some people. If you could get it to work where each component has a poetry pyproject.toml, I think IDEs like pycharm would have a much easier time with it. I have a few personal side-projects that would be better managed in this polylith style so I'm going to try it out. Thanks a bunch. |
Uh oh!
There was an error while loading. Please reload this page.
Thanks for the interesting project. I've just started to dig in but one thing that stood out is the need to redeclare transitive dependencies of components. For instance:
So, the duplication of pendulum dependency can end up being N+1 where N=number of projects that use component A.
Ideally one would declare the pendulum dependency within component A once.
Any thoughts on this?
All reactions