Tim Whiting
Partially addresses #828.
This fixes the open-resolve issue, and illustrates the problem, but likely introduces unnecessary inefficiencies. One potential solution might be to leave the effect row as is and set a flag...
Return clauses tend to pile up when using a lot of effect handlers, however most return clauses are identity functions or idempotent. We should detect and optimize these. See also...
Previously the unification didn't match the named argument eta-expansion / type matching. This could cause surprising behavior where `?@copy:(a, x:int, y:int) -> a` unified with `?@copy: (a, dx: int, dy:...
Fixes #711. The real fix would be to move all private constructor related functions out of the C header files, but currently we have them as `static inline` functions in...
Fix exit status on unix for `system` call. The returned integer is not directly interpretable as an exit status. Fixes: #797 Unfortunately, unix does this to require the user to...
If named effects do not have or use `ctl` effects, they never throw an exception if used after escaping. Personally I see this as a feature, since you can easily...
This rebases unique vectors on #716 and updates it to use the local var primitives for the functionality it implements.
It would be great to introduce multiple types in their own scope within files. Particularly to prevent errors from other parts of the file prevent the successful type check of...
#780 brought this up. The parameter/return value for `rcontext.finalize` does not make much sense in practice. `raw ctl`'s `rcontext` is intended to escape the handler's return context, and the required...