Gabriel Scherer
Gabriel Scherer
In the work on [Camlboot](https://github.com/Ekdohibs/camlboot), we actually benefit from code that "heeds warning 40", in the sense that it explicitly qualifies its constructors or only uses the latest-bound constructor in...
@sliquister I broadly agree that your alternative suggestions are interesting, but one aspect in which they are lacking is that there is no tooling to check that they are correctly...
Note: in recent OCaml versions binding operators support a weird form of punning, `let+ x in ...` for `let+ x = x in ...`, so one can even write ```ocaml...
I know little about jj development and/or Rust packaging policies, but I did not find the argument for vendoring the code instead of making a separate crate very convincing. (The...
I ended up here after being frustrated at how hard it is to reorder commits. I had a feature proposal (see below), and I thought to look for other issues...
I would prefer to see `caml_result` in this PR -- we decided to deprecate encoded exceptions to get rid of them, it feels weird to start introducing more of them...
I understand that it's my duty to work on adapting this PR to `caml_result`, and will try to work on it on my next hacking spree.
> we have some fast paths in our atomics when there's only a single one. So ThreadSanitizer can see the non-atomic fast paths as races with the atomics when there's...
That's an excellent, point, but do `caml_atomic_*` operations require the runtime lock? This is natural for atomics that need the write barrier, but I wouldn't have necessarily expected it from...