Am Mo., 7. Nov. 2022 um 15:45 Uhr schrieb John Cowan <xxxxxx@ccil.org>:
>
>
>
> On Mon, Nov 7, 2022 at 9:20 AM Marc Nieper-Wißkirchen <xxxxxx@gmail.com> wrote:
>
>>
>> > If an unspecified return could be "zero or more values", then the same
>> > expression could return a different number of values each time. A test
>> > can't reliably catch that.
>
>
> It certainly can, using (receive vals expr vals), which binds to a list of the values returned by expr and then returns them.  There are a variety of equivalent expressions using let-values and call-with-values.
>
> That said, it is IMO better to stick with the existing R7RS-small convention and then leave it to be considered by Committee C (which I believe already has an issue for this) at a later stage.  Having an inconsistent convention benefits nobody.

The sample implementation currently returns zero values, which should
catch errors early when people misunderstand the semantics of
"independently" with the semantics of "begin".