David Allsopp
David Allsopp
Yes, we could do (as an aside, we already have `caml_thread_t`, so for macOS it could just include a pointer to the string for the initial name and `caml_thread_start` would...
It wasn’t so much the naming of an arbitrary thread, but I could certainly imagine it being the thread which _creates_ another thread being the point where the name is...
Looking at the referenced commit in Tsdl, I think it's worth conceptually separating out libraries which use hand-rolled stubs (like graphics above, and sqlite3) and ctypes-based ones, like Tsdl. For...
I'm not sure (without digging into it further than I have time to at the moment) whether the fix would be in ctypes or in the libraries themselves - but...
Obviously OCaml 5.3.0 is not yet even branched, but the winpthreads package is already needed for `ocaml-variants.5.3.0+trunk`, and I have a proposed back-port of the MSVC port to 5.2.0 I...
> The packages are not constrained to Windows as the _opam packages_ aren't particularly Windows-specific (and, indeed, at some point cross-compilers will certain want the FlexDLL sources). How embarrassing, I...
It is using a mingw build of OpenSSL (from [`mingw64-x86_64-openssl`](https://cygwin.com/packages/summary/mingw64-x86_64-openssl.html) but the “stable” version is very out of date. setup-ocaml could mitigate this if either a mechanism to specify additional...
Yes - by installing both `mingw64-x86_64-openssl=1.1.1w-0.1` and `mingw64-i686-openssl=1.1.1w-0.1` in windows.ts (as above)
> PS: nope sadly `opam exec -- dune runtest` is not sufficient to get (cygwin) `sh` in your path... If/When cygwin executable are added to the path by opam is...
@Octachron had privately me asked about the s/shall/will - I think the original "shall" is valid (in fact, pedantically, I think it's more correct than "will" in the contexts in...