On 3/23/26 22:03, John Cowan wrote:
>
> On Mon, Mar 23, 2026 at 4:12 PM Daniel Ziltener <xxxxxx@lyrion.ch>
> wrote:
>
>     Okay. I adjusted the `copy` message doc accordingly, and added a
>     `deep-copy` message that is declared as optional for implementers. I
>     definitely won't be able to implement this in pure R7RS, nor in my
>     "day
>     to day" dialect of Chicken, but I think now that deep-copying is a
>     potentially valuable mechanic for some.
>
>
> I'm not sure why it can't be implemented, given a rule that you don't
> descend into arbitrary Scheme values like pairs, lists, vectors, or
> records (the last being obviously problematic)
Not descending into everything means it by definition is not a deep copy.