David Allsopp

Oh, by the by - if it turns out that the option can be better optimised, my suggestion instead would have been to retrieve the magic value from C at...

Yes! And indeed in https://github.com/dra27/ocaml/pull/240/commits/562a49751e423ec4bfc12eaca8b4b6dd14d181ed there's a a logging function for the OCaml side. Key to all of it - and a flaw in OCaml's Unix API at the moment...

Incidentally, I haven't double-checked the details on clambda, but the fact that both flambda 1 and 2 optimise optional arguments properly means that the approach in this specific PR should...

Can I cross-reference a quite long exchange on the original referencing issue about this (https://github.com/dbuenzli/topkg/issues/11) - the short summary is that I don't think anything should be changed (except fixing...

I can see a case (though I still find it unnecessary) as a halfway house for the symlink (or copy) of the build artefact having `.exe` added - but not...

Just FTR, the regression is in fact #12811 (in 5.2.0), and the use of `thread_local` for `caml_state` in C++ mode - see https://github.com/ocaml/ocaml/pull/14220#issuecomment-3254629771

Full review to follow, but I disagree about splitting the PR - the errorcheck part is in a separate commit, so not much easier to review in another PR and...

> Is it a big issue if the Windows port does not implement ERRORCHECK and deadlocks instead? My impression was that ERRORCHECK is there to catch programming errors (in particular,...

Something else is going on here - it shouldn’t be necessary to do this in a package build - I’ll endeavour to have a look tomorrow/next week

Given (facetiously), that `" "` has a much more bloated representation than `' '`, I'm not sure (non-facetiously) why having a function to cut on a character separator and (perhaps...