Skip to content

Conversation

@kit-ty-kate
Copy link
Member

This is a starting point for a broader discussion about the goal of each opam libraries and in particular opam-core.

In my opinion OpamSWHID isn't really useful outside of the context of opam-format and moving it there allows us to avoid an unnecessary dependency towards swhid_core in `opam-core.

However we can also question ourselves on whether opam-core itself is a useful library to have on its own. As far as i know, no-one uses it without also using opam-format so maybe merging the two could be envisioned.

Another question could also be "what is our threshold for generic things to put into opam-core vs. the other library?". For example #5892 leaves OpamSystem.patch where it was but now a new dependency is required, but this function is only used in opam-repository so it could be better to just move it to that library instead.

So far after thinking about it a little i'm thinking that we could merge opam-core into opam-format then that leaves:

  • opam-client for UI and commands stuff
  • opam-state for the low-level internals
  • opam-repository for things related to fetching and repository composition
  • opam-solver for solver stuff
  • opam-format as a base (format and all the generic modules)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

PR: WIP Not for merge at this stage

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant