Sam Lindley
Sam Lindley
I discussed this with @SquidDev and I explicitly told him that we should have a flag for this, and I stand by this. It is an experimental feature that it...
Yep. I haven't looked at the code at all recently, but we definitely rely on the order in which types are unified. It affects the types that are inferred, and...
In theory, that shouldn't affect correctness, but there may be other subtleties here.
Yes, the effects should also be flexible. I suggest that we don't use the term "relaxed" value restriction for our approach, as that term is used by Jacques Garrigue to...
Yeah. That's the purity restriction. We could easily implement it in Links - @yallop and I discussed doing something similar over a decade ago, but somehow never got round to...
The fix was only ever implemented for non-recursive functions. If we just change the original example slightly to make `g` recursive: var f = id(id); sig g : forall a....
Isn't the issue that we should have some tests that exercise the "CP syntax" at all? Or is there really some specific non-trivial interaction with local functions that you have...
The first phase of this was done and merged in #859 , but I've left the issue open because there's more to do. What remains includes tidying up (e.g. by...
I haven't looked at this closely, but I did notice immediately that the original code shouldn't work because you would need to freeze any variable of type `List(a)` in order...
Related to #1016.