From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26873 invoked by alias); 13 Aug 2005 00:33:57 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 26849 invoked by uid 48); 13 Aug 2005 00:33:53 -0000 Date: Sat, 13 Aug 2005 00:33:00 -0000 Message-ID: <20050813003353.26848.qmail@sourceware.org> From: "wilson at gcc dot gnu dot org" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20050802184349.23205.dpatel@apple.com> References: <20050802184349.23205.dpatel@apple.com> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug c++/23205] [4.0/4.1 Regression] [C++/unit-at-a-time] stabs debug info omitted for global const variables X-Bugzilla-Reason: CC X-SW-Source: 2005-08/txt/msg01463.txt.bz2 List-Id: ------- Additional Comments From wilson at gcc dot gnu dot org 2005-08-13 00:33 ------- The problem here is the DECL_RTL_SET_P tests in dbxout_global_decl and dbxout_symbol. Constants have a NULL DECL_RTL, and hence fail this test. This is OK, because we don't use the DECL_RTL when emitting debug info for a constant. However, everything else does need to have a valid DECL_RTL. It looks like deleting the redundant test in dbxout_global_decl and repositioning the test after the check for constants in dbxout_symbol case VAR_DECL would work. Writing and testing the patch is left as an exercise for the reader. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23205