From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 9720 invoked by alias); 14 Oct 2004 18:40:34 -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 9706 invoked from network); 14 Oct 2004 18:40:33 -0000 Received: from unknown (HELO mail.codesourcery.com) (65.74.133.9) by sourceware.org with SMTP; 14 Oct 2004 18:40:33 -0000 Received: (qmail 24387 invoked from network); 14 Oct 2004 18:40:32 -0000 Received: from taltos.codesourcery.com (zack@66.92.218.83) by mail.codesourcery.com with DES-CBC3-SHA encrypted SMTP; 14 Oct 2004 18:40:32 -0000 Received: by taltos.codesourcery.com (sSMTP sendmail emulation); Thu, 14 Oct 2004 11:40:32 -0700 To: gcc-patches@gcc.gnu.org Subject: Re: RFC Speed up dbxout.c a bit References: <877jptn0qm.fsf@codesourcery.com> <20041014132402.GA8948@nevyn.them.org> <87u0sxkzf6.fsf@codesourcery.com> <20041014172512.GA15773@nevyn.them.org> From: Zack Weinberg Date: Thu, 14 Oct 2004 18:48:00 -0000 In-Reply-To: <20041014172512.GA15773@nevyn.them.org> (Daniel Jacobowitz's message of "Thu, 14 Oct 2004 13:25:12 -0400") Message-ID: <87is9dkusf.fsf@codesourcery.com> User-Agent: Gnus/5.110003 (No Gnus v0.3) Emacs/21.3 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-SW-Source: 2004-10/txt/msg01206.txt.bz2 Daniel Jacobowitz writes: > On Thu, Oct 14, 2004 at 10:00:29AM -0700, Zack Weinberg wrote: >> Daniel Jacobowitz writes: >> >> > Unfortunately this won't work. Did you test on a target which breaks >> > up stab strings? GDB (and probably other consumers) do a similar >> > process of only putting the string back together in certain places. >> > The majority of the code will fall down if you insert backslashes >> > arbitrarily. >> >> Ugh. I guess I didn't test on such a target. Do you know what the >> constraints are? If it's a simple rule like 'only at semicolons' it >> won't be that hard to do after the fact. > > They're quite arbitrary. Take a look at STABS_CONTINUE (and one call > in dbxread.c; you'll find it if you search for "cretinous"). Looking at that, it seems wisest to put back the old semantics of continuing only at explicitly marked points. I think I have figured out how to do that in the new framework. zw