Anil Madhavapeddy
Anil Madhavapeddy
The latest patchset on this branch improves by: - added github actions, to test on macos and windows. Windows now builds on mingw64 - packaging fixes for ctypes _vs_ ctypes-foreign,...
We could simply drop the unthreaded variant. We could do this in stages: merge the existing port, and subsequently remove it (and the use of variants). We could also propose...
Merged against #651 and activated ocaml-ci on my fork. I'm seeing some failures on arm32 I need to chase down: https://ci.ocamllabs.io/github/avsm/ocaml-ctypes/commit/96700a9886dfea4551192b2cc75de9d1e845b859/variant/debian-10-ocaml-4.10_arm32 ``` #18 7.805 test_returning_errno alias tests/test-returning-errno/runtest (exit 1) #18...
To broaden that, all the Lwt-based tests seem to fail in the PR at the moment, but only on Linux but not macOS.
Thanks for looking @bobot! The main worrying thing about this PR is a test case failure (on the existing branch) on Linux for the Lwt tests, but only when running...
It would be great to root-cause the underlying reason why behaviour changes. The ctypes Makefiles are rather precise at present, and I'm nervous about proposing a large-scale change to the...
Fantastic!! Great detective work @emillon.
I'm looking into this. It smells a bit like a compiler bug in 4.07 to do with packing, since it's very hard to reproduce.
did you have cstruct.4.0.0 pinned while building this @XVilka ?
Awesome! I'll take a look at the Mirage end later this afternoon, and review the dune changes as well.