Open hhugo opened 1 year ago
Is there anything else that needs to be done in this PR? It looks good to me.
Is there anything else that needs to be done in this PR? It looks good to me.
First, the PR should not be marked Draft if you want me to look at it.
Second, I still don't understand what we have to gain by switching to a Dune build. This is a trivial project as far as OCaml sources are concerned, but it has a C side with non-trivial configuration and dependencies. That's precisely the kind of project where Dune does not shine. Why insisting on using Dune here?
Second, I still don't understand what we have to gain by switching to a Dune build. This is a trivial project as far as OCaml sources are concerned, but it has a C side with non-trivial configuration and dependencies. That's precisely the kind of project where Dune does not shine. Why insisting on using Dune here?
Hi @xavierleroy , the main reason I wrote #73 was to be able to compose the build of Coq and ZArith.
This was essential in order to have a reasonable setup to track a zarith bug we were seeing in Coq more efficiently, but indeed, it may not be a common developer setup.
I guess the main feature ZArith users could benefit given a Dunerization would be build caching, when that is working for opam packages; some other minor features that are convenient are odoc
support, dune-release
support, hygienic builds, sound incremental builds, etc...
Continue the effort started in https://github.com/ocaml/Zarith/pull/73.
In particular, it tries to address the following comment
It's rebased on top of the recently merged new CI jobs.
List of changes: