From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 8015 invoked by alias); 24 Sep 2004 14:41:30 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 8006 invoked by uid 48); 24 Sep 2004 14:41:30 -0000 Date: Fri, 24 Sep 2004 14:41:00 -0000 Message-ID: <20040924144130.8005.qmail@sourceware.org> From: "matthew dot amos at baesystems dot com" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20040922161456.17615.matthew.amos@baesystems.com> References: <20040922161456.17615.matthew.amos@baesystems.com> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug fortran/17615] Internal compiler error using interface procedures X-Bugzilla-Reason: CC X-SW-Source: 2004-09/txt/msg02952.txt.bz2 List-Id: ------- Additional Comments From matthew dot amos at baesystems dot com 2004-09-24 14:41 ------- I think what you describe is correct: crossproduct3_R4_R8 should not be visible in the program's namespace, but the compiler should have enough information to choose cross_product3_R4_R8 from the argument types even though only cross_product is used. In my experience of other Fortran 95 compilers this is the expected behaviour, but I was not sure if this feature is present in GCC and, if it was, whether it acts like the C++ virtual keyword rather than statically linking. The code posted above is an excerpt from a highly-numerical F90 code which I'm trying to get working on an Itanium. The commercial compiler I have been given is not up to the task, so I'm exploring other options! -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17615