James Cheney

I suggest we defer this to 0.7.3 and release 0.7.2 ASAP.

Fine with me, just means no one besides you will try using handlers, which maybe is a feature rather than bug if it's all about to change.

This is intended (for now). We have not tested with mariadb as far as I know; we can remove this once we have done so and made any changes needed....

Incidentally, our dependency doesn't rule out mariadb, it just means that links-mysql installation won't succeed unless conf-mysql is also installed. So it just has the effect of saying that while...

I don't see any particular obstacle to implementing this as described above / in ATTAPL, perhaps by defining a "restriction" on row kinds. However, we are fairly seriously considering extending...

At the risk of falling prey to Wadler's Law, this seems like it would be an incredibly confusing design, since in C/Java/JS family languages unaccompanied braces are almost always just...

I think `unsafe` should be documented (it is in the prelude) and the documentation should say don't use it. We could even make it an error to use it outside...

I think this is because after the first typechecking pass, functions are frozen (`fun_frozen` is set) and because of this, the following line https://github.com/links-lang/links/blob/0e3a7e51bc6237e34b98acc661bbfd11064200bd/core/typeSugar.ml#L4292 results in the quantifiers of the...

This change would be easy enough to make as part of #900, but we should probably discuss its ramifications first. I just experimented with this by disabling the `/* empty...

I had a very superficial try at doing this in branch `issue945`. One wrinkle I hadn't thought of is that we make `Unionfind.point` showable, for debugging. I modified this to...