From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23401 invoked by alias); 27 Mar 2012 19:13:21 -0000 Received: (qmail 23338 invoked by uid 22791); 27 Mar 2012 19:13:19 -0000 X-SWARE-Spam-Status: No, hits=-2.3 required=5.0 tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW X-Spam-Check-By: sourceware.org Received: from mail-gx0-f175.google.com (HELO mail-gx0-f175.google.com) (209.85.161.175) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 27 Mar 2012 19:13:05 +0000 Received: by ggcy3 with SMTP id y3so262602ggc.20 for ; Tue, 27 Mar 2012 12:13:04 -0700 (PDT) MIME-Version: 1.0 Received: by 10.236.200.197 with SMTP id z45mr27116348yhn.99.1332875584093; Tue, 27 Mar 2012 12:13:04 -0700 (PDT) Received: by 10.100.146.5 with HTTP; Tue, 27 Mar 2012 12:13:04 -0700 (PDT) In-Reply-To: <4D9545D6-4857-4BCE-8BF9-C824304F4B8C@comcast.net> References: <4D9545D6-4857-4BCE-8BF9-C824304F4B8C@comcast.net> Date: Tue, 27 Mar 2012 19:13:00 -0000 Message-ID: Subject: Re: [patch][RFC] bail out after front-end errors From: Steven Bosscher To: Mike Stump Cc: GCC Mailing List , GCC Patches Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable X-IsSubscribed: yes Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2012-03/txt/msg01747.txt.bz2 On Tue, Mar 27, 2012 at 8:59 PM, Mike Stump wrote: >> and errors from the middle end are mostly for exotic >> code (involving asm()s and the like). Bailing out after parse errors >> is therefore IMHO the right thing to do for the common case. > > Some are for very exotic things, yes, but not all of them. =A0Actually, I= started reviewing them, none of them seem that exotic to me. =A0Take for e= xample: > > int f() { return this->i; } =A0 // { dg-error "" "no member named i" } > > Really, exotic? I expect it is not difficult to retain almost all of these C++ errors by lowering (i.e. gimplifying) in the front end, or by having some intermediate stage. It certainly wouldn't be acceptable to loose all those diagnostics, just like that. I'm trying to figure out how to avoid that! But it seemed prudent to ask what other people think of this kind of change before spending loads of time on it :-) Thanks for your comments! Ciao! Steven