stopped_as_optional should handle void senders

[ericniebler]

Issue by ericniebler
Friday Oct 22, 2021 at 00:55 GMT
Originally opened as NVIDIA/stdexec#222

---

... by mapping void to some unit (monostate?) type.

[lewissbaker]

@ericniebler Is the classification as P1 enhancement here based on the assumption that it's currently ill-formed to apply stopped_as_optional() to a sender that completes with set_value_t() in that we can later make this well-formed as completing with set_value_t(optional<monostate>)?

[ericniebler]

Yes