On 2025-10-20 15:52 +0000, Sergei Egorov wrote:
> - Consistency. If the constructor (list x y z) makes no promise
> about the evaluation order of `x`, `y`, and `z`, the pattern
> (~list x y z) should not impose a matching order either.

This is a good argument, but the presence of binding expressions
makes the situation more complicated.  Unspecified evaluation order
is good; unspecified scope is not.  No weakness is more deadly for
matching macros.

I second (third?) Daphne & Marc here.  If the matching order is
unspecified, then ban variable references that depend on order.
Otherwise, impose clear semantics for scope.

--
Wolfgang Corcoran-Mathe  <xxxxxx@sigwinch.xyz>