dynamic extent closure

[ueda@shpcsl.sharp.co.jp (UEDA masaya)]

Dear MCLers:

I found some examples declaring dynamic extent closure in MCL final
reference manual and source code files.

I looked for old info-mcl mail that refers to dynamic-extent. I
could found some profitable comments, but I can not exactly understand
the benefit of the declaration yet.

What is the merit of dynamic extent closure? I guess no garbage
collection needs for the closure as they are allocated on stack. Is
it right?

Any suggestions would be appreciated.

   From: Bill St. Clair <bill@cambridge.apple.com>
   Date: Wed, 21 Aug 91 16:50:17 -0400

      Date: Wed, 21 Aug 91 13:59:33 CDT
      From: engber@aristotle.ils.nwu.edu (Mike Engber)
      Subject: what does (declare (downward-function)) do?

      The subject line says it all. I see no mention in Steele.
      I assume it allows for some optimization.
      -ME

   DOWNWARD-FUNCTION will no longer exist in MCL 2.0 final.

   It is a Lisp Machine concept providing an alternative way to declare
   that a closure has DYNAMIC-EXTENT.  We decided to get rid of it
   because the extent of the closure is confusing.

--- Thanks ---
Masaya UEDA
Information System R&D Center, SHARP Co.
2613-1 Ichinomoto, Tenri, Nara 632, JAPAN