> > There are some problems here:
> 1. You have limited K to int or string, but none of these types can be > subtyped. As a result K1<:K2 never applies. We can only have K1:=K2. > 2. You also claim that V is covariant. However, this holds true only for > read-only arrays. Write operations of an array use V in contra-variant > positions. > I am afraid that we cannot have sound covariant subtyping of an array, > unless we clearly separate the read interface from the write one. Hello there, Lazare, once again, thanks for the feedback: I just published a v0.2 of the RFC with changes based on your feedback if you want to take a look at it. Also, for other members of Internals, I'd be happy to get your feedback on this. I plan to start implementation as soon as we agree on an Implementation Level I should aim for. Thanks in advance. *---* *Best Regards,* *Wendell Adriel.* *Software Engineer & Architect* *https://wendelladriel.com <https://wendelladriel.com>* Em qui., 16 de jul. de 2026 às 15:18, Wendell Adriel < [email protected]> escreveu: > There are some problems here: > > >> 1. You have limited K to int or string, but none of these types can be >> subtyped. As a result K1<:K2 never applies. We can only have K1:=K2. >> 2. You also claim that V is covariant. However, this holds true only for >> read-only arrays. Write operations of an array use V in contra-variant >> positions. > > >> I am afraid that we cannot have sound covariant subtyping of an array, >> unless we clearly separate the read interface from the write one. > > > Hey Lazare, > > Thank you for the feedback. > I'm going to analyze this and update the RFC document accordingly ASAP. > > *---* > *Best Regards,* > *Wendell Adriel.* > *Software Engineer & Architect* > *https://wendelladriel.com <https://wendelladriel.com>* >
