srfi-145 assume is allowed to do nothing.  So (maybe-ref maybe) can just be evaluated #f if maybe is Nothing?

By the way, can't (assume (just? obj) message) be enough?
 

On Tue, Jun 16, 2020 at 9:27 PM Marc Nieper-Wißkirchen <xxxxxx@nieper-wisskirchen.de> wrote:
Am Mi., 17. Juni 2020 um 04:39 Uhr schrieb John Cowan <xxxxxx@ccil.org>:
>
> While this is unobjectionable in itself, it's hard to know where to stop.  One of the SRFI 145 examples is (assume (exact-integer? x) ...).  This could be changed to (assume-exact-integer x ...), but I don't think it would be very useful to add this macro to a SRFI.

That's a good point.

So the only reasonable way to have an implementation of the semantics
of assume-just in a SRFI would be to make the parameter FAILURE of
maybe-ref optional again so that the default is (lambda () (assume #f
"just expected")).

What do you think?