Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!sun-barr!ames!pasteur!ucbvax!decwrl!decvax!ima!think!barmar
From: barmar@think.COM (Barry Margolin)
Newsgroups: comp.lang.lisp
Subject: Re: (eval-when (load) ...
Message-ID: <40511@think.UUCP>
Date: 12 May 89 07:04:32 GMT
References: <8905120156.AA17446@anaconda.Stanford.EDU>
Sender: news@think.UUCP
Reply-To: barmar@kulla.think.com.UUCP (Barry Margolin)
Organization: Thinking Machines Corporation, Cambridge, MA
Lines: 27

In article <8905120156.AA17446@anaconda.Stanford.EDU> rit@ANACONDA.STANFORD.EDU (Jean-Francois Rit) writes:
>I think the following statement is true:
>(eval-when (load) <s-exp>) is equivalent to <s-exp> (i.e. at the top-level of
>a file.
>Am I right?

No.  It's only equivalent if the file is compiled, and only if the
<s-exp> is one which doesn't need to modify the compile-time
environment.  In order to make it equivalent whether you load the
source or the binary, you need to specify (eval-when (eval load) ...).
And if the <s-exp> is something like a DEFMACRO, and the macro is used
later in that file, the eval-when will prevent the compiler from using
it when it encounters the invocation.

Unfortunately, there isn't a portable way to write what DEFMACRO
expands into when it's processed by COMPILE-FILE.  When there's no
EVAL-WHEN around the DEFMACRO, the compiler is permitted to define it
in such a way that it is only in effect during the compilation
process.  (eval-when (compile ...) (defmacro ...)) specifies that the
macro should be fully defined in the Lisp that is running the
compiler, which is not the same thing.

Barry Margolin
Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar