Louis Gesbert
I don't really have an opinion on removing the magic... the cases pointed out in the above are very specific to a convoluted CI setup, and the code might still...
I think that was a planned feature in 1.0 (when it was not clear if ocamlfind would be replaced) but never got any use; it was already deprecated in 2.0...
This seems useful! The real test would be to try and actually install the generated bundles though :) One thing: the `shell/` directory is at the moment used to contain...
Thanks! > It is annoying that so many packages have "No solution". It's just because the OCaml version is fixed at bundle creation time, I think. Opam has a feature...
3 CI jobs fail: ### * The main test [Build learn-ocaml and run tests](https://github.com/ocaml-sf/learn-ocaml/actions/runs/6120233013/job/16611680057?pr=481#logs) successes on local tests then fails on the corpus ``` + docker run --entrypoint '' -v...
This sounds very reasonable — a perfect solution actually. I'd have given you a "go" already for an asak release so that this can be merged, but @erikmd promised he...
And thanks :)
> This sounds very reasonable — a perfect solution actually. I'd have given you a "go" already for an asak release so that this can be merged, but @erikmd promised...
@erikmd, this finalised version includes @nobrakal's patches so you should be able to check the partitioning feature directly with the static binary artefacts. If that works I don't think there...