From owner-sc22wg5@open-std.org Mon Dec 8 19:40:41 2008 Return-Path: X-Original-To: sc22wg5-dom7 Delivered-To: sc22wg5-dom7@www2.open-std.org Received: by www2.open-std.org (Postfix, from userid 521) id 69A4DC178E5; Mon, 8 Dec 2008 19:40:41 +0100 (CET) X-Original-To: sc22wg5@open-std.org Delivered-To: sc22wg5@open-std.org Received: from mailrelay1.lrz-muenchen.de (mailrelay1.lrz-muenchen.de [129.187.254.106]) by www2.open-std.org (Postfix) with ESMTP id BCD54C178E4 for ; Mon, 8 Dec 2008 19:40:40 +0100 (CET) Received: from [129.187.15.179] ([129.187.15.179] [129.187.15.179]) by mailout.lrz-muenchen.de with ESMTP; Mon, 8 Dec 2008 19:40:13 +0100 Message-Id: <493D6A0D.2090705@lrz.de> Date: Mon, 08 Dec 2008 19:40:13 +0100 From: Reinhold Bader User-Agent: Thunderbird 2.0.0.18 (X11/20081112) MIME-Version: 1.0 To: longb@cray.com, sc22wg5@open-std.org Subject: Re: (j3.2006) (SC22WG5.3759) Response on the TR29113 draft N1761 References: <20081207203535.92039C178D6@www2.open-std.org> <20081208171349.2CAD1C178E4@www2.open-std.org> In-Reply-To: <20081208171349.2CAD1C178E4@www2.open-std.org> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: owner-sc22wg5@open-std.org Precedence: bulk Bill Long schrieb: > > Reinhold Bader wrote: > > Issue 6 - referencing or defining assumed rank entities: > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > We need some rules to deal with this. > > > There are rules currently - basically that the only thing allowed on the > Fortran side is to pass an assumed-rank dummy as an actual argument to > another procedure. Are you saying that the scope of the TR needs to be > expanded to specify ways for things like this to be used in Fortran? Hmm - I guess my feeling simply is that if we do it, we should do it right. It works a bit like polymorphism, only with respect to rank instead of type. Besides, if we really really want the functionality for C interop *only*, why did we not choose the semantics described by me for DIMENSION(**) in the first place? The formal problem with this of course is that it can be argued that it does not fall under the scope of the TR. > Beyond the rank remapping already allowed for pointers? Which presently only treats the reverse case, namely the target being 1D (I believe). Note that the > main purpose of adding this feature was for MPI-like interfaces, where > the only objective was to pass the object to a C function and give the > function tools to correctly interpret the object. Yes. Note that I'm not yet sure the MPI Forum is going to swallow this since the MPI vendors apparently don't want to rewrite their whole infrastructure to use descriptors ... even though it provides the technically better solution. > > Cheers, > Bill Regards Reinhold