From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 28777 invoked by alias); 2 May 2005 18:54:41 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Received: (qmail 28736 invoked from network); 2 May 2005 18:54:35 -0000 Received: from unknown (HELO mail.codesourcery.com) (65.74.133.9) by sourceware.org with SMTP; 2 May 2005 18:54:35 -0000 Received: (qmail 28026 invoked from network); 2 May 2005 18:54:34 -0000 Received: from localhost (HELO ?65.78.187.111?) (mitchell@127.0.0.1) by mail.codesourcery.com with SMTP; 2 May 2005 18:54:34 -0000 Message-ID: <42767764.8010506@codesourcery.com> Date: Mon, 02 May 2005 18:54:00 -0000 From: Mark Mitchell User-Agent: Mozilla Thunderbird 1.0.2 (Windows/20050317) MIME-Version: 1.0 To: Florian Weimer CC: Richard Henderson , Arnaud Charlet , John David Anglin , gcc-patches@gcc.gnu.org, laurent@guerby.net Subject: Re: [patch] Add Ada tasking support for hppa-unknown-linux-gnu (take 2) References: <20050429102640.GA34244@adacore.com> <200504291341.j3TDfTKr010394@hiauly1.hia.nrc.ca> <20050429144405.GA79012@adacore.com> <20050429234000.GE8858@redhat.com> <42755DB4.10402@codesourcery.com> <87psw9bjsy.fsf@deneb.enyo.de> In-Reply-To: <87psw9bjsy.fsf@deneb.enyo.de> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-SW-Source: 2005-05/txt/msg00112.txt.bz2 Florian Weimer wrote: > * Mark Mitchell: > > >>>One way to fix all of this would be to move libada out to the top >>>level, and use configure and/or generator programs that actually >>>probe the contents of /usr/include for the data. Personally I >>>don't think any other solution is in fact a solution. >> >>I agree. Ada should handle its runtime library just like the other >>languages handle theirs. > > > Would it be acceptable to add a switch to the C front end to cause it > to dump the layout of struct definitions? This would simplify things > enormously. We could implement it without this feature, especially > for well-defined structs whose fields are known (but not their order > and sizes), but it's not very efficient when cross-compiling. Yes, but even better would be to add a switch that works in a language-independent manner. -- Mark Mitchell CodeSourcery, LLC mark@codesourcery.com (916) 791-8304