From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 9574 invoked by alias); 13 Jun 2008 15:35:30 -0000 Received: (qmail 9565 invoked by uid 22791); 13 Jun 2008 15:35:29 -0000 X-Spam-Check-By: sourceware.org Received: from NaN.false.org (HELO nan.false.org) (208.75.86.248) by sourceware.org (qpsmtpd/0.31) with ESMTP; Fri, 13 Jun 2008 15:35:07 +0000 Received: from nan.false.org (localhost [127.0.0.1]) by nan.false.org (Postfix) with ESMTP id 14E5998402 for ; Fri, 13 Jun 2008 15:35:06 +0000 (GMT) Received: from caradoc.them.org (22.svnf5.xdsl.nauticom.net [209.195.183.55]) by nan.false.org (Postfix) with ESMTP id EBAE19835A for ; Fri, 13 Jun 2008 15:35:05 +0000 (GMT) Received: from drow by caradoc.them.org with local (Exim 4.69) (envelope-from ) id 1K7BIv-0001LX-9H for gdb-patches@sourceware.org; Fri, 13 Jun 2008 11:35:05 -0400 Date: Fri, 13 Jun 2008 19:43:00 -0000 From: Daniel Jacobowitz To: gdb-patches@sourceware.org Subject: Re: [FYI] Inlining support, rough patch Message-ID: <20080613153505.GA4971@caradoc.them.org> Mail-Followup-To: gdb-patches@sourceware.org References: <20080613152754.GA4220@caradoc.them.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20080613152754.GA4220@caradoc.them.org> User-Agent: Mutt/1.5.17 (2008-05-11) X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2008-06/txt/msg00242.txt.bz2 On Fri, Jun 13, 2008 at 11:27:54AM -0400, Daniel Jacobowitz wrote: > Enjoy. Forgot one thing I meant to add. There are several test cases in the patch. These have been very useful while developing it; I've run them on five architectures and a variety of compiler versions. I didn't fix all the failures on some of those versions. I'm not sure what to do with them. They're really hard to maintain at no failures; every time you change one of the tests to fix a failure on one platform, you run a higher than usual risk of breaking it on another platform. Tested for optimized code behavior are hard :-( Also, I did test it on non-toy programs, obviously. It does a passibly good job stepping through the dwarf2 reader. But the behavior is a little quirky, especially due to scheduling. And GDB is very sensitive to compiler debug info bugs. And GCC has a lot of them. Overall, with this patch merged I think the next places to go for optimized code debugging are the two GCC branches currently testbeds for improved optimized debug info generation. After that, it's back to the age-old debugging scheduled code problem. -- Daniel Jacobowitz CodeSourcery