Jan Midtgaard

This just came up again in https://github.com/c-cube/qcheck/pull/364 where I ended up inserting a special case to print negative `nan` as `-nan` across all platforms, but otherwise still using `string_of_float`. I'm...

FTR, https://github.com/ocaml/ocamlfind/pull/112 fixes this. It "just" needs to be reviewed, merged, and released with the next version 🙂

Note: this issue is tracked in https://github.com/ocurrent/opam-repo-ci/issues/452. I'll keep this open for now, so that package submitters observing it have a chance of finding out.

Although not a slick `setup-ocaml` option, a good step forward would be to get a more recent `opam` into Cygwin https://cygwin.com/packages/summary/opam.html 🤓

This is excellent - thanks a ton! :partying_face: Thinking a bit a head, one could imagine having a "pure-Cygwin" CI workflow too at some future point? :thinking: At that point,...

> I thought msys2 was always using mingw but now I see an unprefixed `gcc` package along with the `mingw`-prefixed one. Do you know what is this `gcc`? Do we...

AFAICS, you've specified the wrong checksum: ``` $ wget https://github.com/Ckaf/pgn_parser/archive/v1.0.2.tar.gz [...] $ sha256sum v1.0.2.tar.gz d84c375f6acec42c8c3a64a05f6942e9c0289a8fd367a52f11099c866d9b1e22 v1.0.2.tar.gz ``` which differs from `78e455b3d4446e9cd8c191b78cb83a7edbc526894eb917ef0555346c1e5c7f86`

There's progress: Now I see a bunch of lower-bounds failures: `chess_com_api.1.0.2` and `pgn_parser.1.0.0` together fails with ``` #=== ERROR while compiling chess_com_api.1.0.2 ================================# # context 2.4.1 | linux/x86_64 | ocaml-base-compiler.4.14.2...

> I don't understand these errors about cyclic dependencies. Is it something I should worry about? Nope, no worries. There are known issues with opam-2.0: https://github.com/ocaml/opam-repository/issues/28952

OK, thanks!