From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 24858 invoked by alias); 8 Oct 2007 07:13:51 -0000 Received: (qmail 24843 invoked by uid 22791); 8 Oct 2007 07:13:50 -0000 X-Spam-Check-By: sourceware.org Received: from nikam-dmz.ms.mff.cuni.cz (HELO nikam.ms.mff.cuni.cz) (195.113.20.16) by sourceware.org (qpsmtpd/0.31) with ESMTP; Mon, 08 Oct 2007 07:13:45 +0000 Received: from localhost (occam.ms.mff.cuni.cz [195.113.18.121]) by nikam.ms.mff.cuni.cz (Postfix) with ESMTP id 026305B8B9; Mon, 8 Oct 2007 09:13:41 +0200 (CEST) Received: by localhost (Postfix, from userid 16202) id 013F0939EA; Mon, 8 Oct 2007 09:13:40 +0200 (CEST) Date: Mon, 08 Oct 2007 07:13:00 -0000 From: Jan Hubicka To: Michael Matz Cc: Jan Hubicka , Richard Guenther , Alexandre Oliva , gcc-patches@gcc.gnu.org Subject: Re: Put scope blocks on a diet Message-ID: <20071008071340.GH2375@kam.mff.cuni.cz> References: <20070724180235.GM24519@kam.mff.cuni.cz> <20070806201102.GF4460@kam.mff.cuni.cz> <84fc9c000710031119q3a1672cep5c98a642aa1100cb@mail.gmail.com> <20071007223722.GH2688@kam.mff.cuni.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.9i 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: 2007-10/txt/msg00404.txt.bz2 > Hi, > > On Mon, 8 Oct 2007, Jan Hubicka wrote: > > > > > > > In the long run, we should generate (and output) debug information > > > for types and declarations early and simply remember the DIE when > > > refering to it instead of keeping our hands on the types and decls > > > itself. > > > > This seems to be bit dificult to implement given that inlining and other > > optimizations modify/build new types.... > > I certainly hope that middle-end transformation don't build many new > types. Actually I sure hope they build no new types at all, but I'll > better not say none, you never know :) What I had in mind is that inlining is building copies of types in new copied scopes, that also needs DIEs and since we are slowly getting into busyness of transforming datastructures (like in struct reorg), we should be able to compensate this in debug info instead of emitting debug info for original layout and writting data in modified. Honza > > > Ciao, > Michael.