On Jan 25, 2024, at 10:44 AM, Alfred M. Szmidt <ams@gnu.org> wrote:

On Jan 25, 2024, at 10:17 AM, JD Smith <jdtsmith@gmail.com> wrote:

On Jan 25, 2024, at 8:57 AM, Stefan Monnier <monnier@iro.umontreal.ca> wrote:

(cond-let
(((var value)
(dvar (derived-from var))
((has-the-right-stuff-p dvar)))
(cons 'correct dvar))

(((foo value2)
(bar (1- foo))
((< bar 0)))
(cons 'incorrect bar))

(t nil))

Personally, I was thinking of a syntax like

(my-cond
  (:let var value)
  (:let dvar (derived-from var))
  ((has-the-right-stuff-p dvar)
   (cons 'correct dvar))

  (:let foo value2)
  (:let bar (1- foo))
  ((< bar 0)
   (cons 'incorrect bar))

  (t nil))

  And it perhaps goes without saying that the two approaches could be combined :) —

  (cond-let
    (:let var-for-rest value)

Is there a reason why using a keyword here? How would multiple
bindings be handled -- or is that what the two forms:

(:let foo value2)
(:let bar (1- foo))

try to convey?

Stefan correct me if I’m missing the main point, but I think the difference motivating a special form like (:let …) comes from the desire to establish bindings for all the rest of the clauses, and to make it clear you are doing this. I don’t know the right way to handle that (or how important it is). The proposed cond* has that capability.

My cond-let idea has somewhat different goals, akin to a short-circuiting group of when + when-let’s. Its clause syntax is just a straightforward copy of when-let; hence no new syntax to learn.