gh actions: update to opam 2.1 invocation and refresh compiler versions #135
Conversation
|
We may have to avoid depext for this only on the Windows runners. (or we can make more fundamental improvements!) |
Co-authored-by: Sora Morimoto <sora@morimoto.io>
|
Thanks for this PR. Unfortunately this changes the set of CI failures, but does not fix all. Namely, the windows CI (that uses a separate configuration file) avoided the installation of mirage-crypto-rng-async -- it looks like the new "setup-ocaml" doesn't care about that. |
|
2.1 has been temporarily disabled. v2 uses the latest opam < 2.1. Native Windows support is planned to be introduced in 2.2, so we will unlock this constraint when it is released. |
|
For issues on Windows, it may be enough to simply disable auto depext. |
|
I looked into the ocaml/setup-ocaml documentation (and examples), and tried with a subsequent commit (efb13ff) -- happy to see whether this works as expected (and hear your thoughts). indeed, the mirage-crypto-rng-async opam package is not supported on windows. |
…age, mirage-crypto-rng-async, mirage-crypto-pk and mirage-crypto-ec (0.10.4) CHANGES: - hash: ensure alignment of data, fixes arm32 test cases (mirage/mirage-crypto#143 @hannesm) - drop sexplib and ppx_sexp_conv dependencies (mirage/mirage-crypto#141 @hannesm) - install LICENSE files of different opam packages - require eqaf >= 0.8 to avoid bigarray-compat dependency (mirage/mirage-crypto#139 by @hannesm) - update to GH actions 2.1, fix windows action (mirage/mirage-crypto#135 mirage/mirage-crypto#139 @avsm @smorimoto) - drop rresult dependency (mirage/mirage-crypto#139 mirage/mirage-crypto#141 by @hannesm @reynir) - avoid deprecated Cstruct.len, use Cstruct.length instead (mirage/mirage-crypto#134 by @hannesm) - mirage-crypto-pk: adapt to mirage 4 (mirage/mirage-crypto#141 @TheLortex @samoht)
No description provided.