Jacques Garrigue

Added a few features to the experimental implementation * automatically add pure to ground types in interfaces, and ignore it when subtyping * `[@@@pure]` and `[@@@impure]` in signature, to allow...

Yes, I've been looking for a better name for a while, but couldn't find one yet. Something like `[@@unrestricted_poly]` would be more correct, but not very palatable. I accept proposals...

Thanks for the suggestion. Another idea is `[@@logical]`, to emphasize that quantifiers follow the laws of logic. But we now also need a negation to switch off signature wide qualifiers....

## Current state (2019-11-12) As mentioned above, the current branch allows one to play with the [Furl](https://github.com/Drup/furl/pull/17) library, avoiding clumsy thunks. It extends the proposal with: * `[@@@pure]` and `[@@@impure]`...

> It also is bothers me (just a little) that the _positive_ thing is expressed by a negative statement such as _non_ expansive (or _no_ alloc, _un_ leakable, etc.) It...

Indeed, the problem with SML was that it was considering the problem from the wrong side. Being generalized is a plus, but I don't think there should be an opposite...

The original rules were unsound (they were actually unintentionally stronger than SML 90). I just fixed that, and the implementation too. The implementation now relies on a predicate `is_pure_exp` working...

Sorry for the very long delay. Too many pans to look after.

Can you indicate which version of gcc you are using?

And also which version of ocaml. Note that if your ocaml is 4.07 or newer, the latest version of LablTk is 8.06.8.