From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <cgen-return-1932-listarch-cgen=sources.redhat.com@sourceware.org>
Received: (qmail 10682 invoked by alias); 11 May 2006 11:52:18 -0000
Received: (qmail 10669 invoked by uid 22791); 11 May 2006 11:52:18 -0000
X-Spam-Check-By: sourceware.org
Received: from mail.agora-net.com (HELO mail.agora-net.com) (67.59.132.6)     by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 11 May 2006 11:51:46 +0000
Received: from ttn by mail.agora-net.com with local (Exim 4.50) 	id 1Fe9fG-0007H5-3z 	for cgen@sources.redhat.com; Thu, 11 May 2006 07:49:06 -0400
From: Thien-Thi Nguyen <ttn@glug.org>
To: cgen@sources.redhat.com
In-reply-to: <17506.11439.818364.259602@casey.transmeta.com> (message from 	Doug Evans on Wed, 10 May 2006 11:10:55 -0700 (PDT))
Subject: Re: [patch][commit] New (if (...) (...) (...)) Test Allowed at Top Level  of the Input
Reply-to: ttn@glug.org
References: <44622986.3090309@redhat.com> <17506.11439.818364.259602@casey.transmeta.com>
Message-Id: <E1Fe9fG-0007H5-3z@mail.agora-net.com>
Date: Thu, 11 May 2006 11:52:00 -0000
X-IsSubscribed: yes
Mailing-List: contact cgen-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:cgen-subscribe@sourceware.org>
List-Archive: <http://sourceware.org/ml/cgen/>
List-Post: <mailto:cgen@sourceware.org>
List-Help: <mailto:cgen-help@sourceware.org>, <http://sourceware.org/lists.html#faqs>
Sender: cgen-owner@sourceware.org
X-SW-Source: 2006-q2/txt/msg00027.txt.bz2

   From: Doug Evans <dje@transmeta.com>
   Date: Wed, 10 May 2006 11:10:55 -0700 (PDT)

   Blech.

   Dave Brolley writes:

    > I've committed the attach patch. I allows the psecification of
    > 
    > (if (application-is? <appname>)
    >     (true expr)
    >     (false expr))
    > 
    > at the top level of the CGEN input. This test corresponds to (eq? 
    > application <appname>) which is found in many places in the CGEN source. 
    > For example
    > 
    > (if  (application-is? SID-SIMULATOR)
    >      (include "sid-macros.cpu")
    >      (include "sim-macros.cpu"))

why not allow computed `include' instead?  (perhaps that is already
supported?)  something like:

(include EXPR)

then EXPR can be all manner of `if', `cond' or whatever.  e.g.:

(include (if (application-is? SID-SIMULATOR)
             "sid-macros.cpu"
             "sim-macros.cpu"))

(include (cond ((application-is? SID-SIMULATOR)
                "sid-macros.cpu")
               (else "sim-macros.cpu")))

(include (case (application->symbol APPLICATION-OBJECT)
           ((SIMULATOR) "sid-macros.cpu")
           (else "sid-macros.cpu")))

going from `if' to `cond' gets you extensibility.  going from `cond' to
`case' gets you regularity and some namespace separation, providing you
are in the mood for a certain style of symbolic manipulation...

thi