From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gdb-patches-return-36873-listarch-gdb-patches=sources.redhat.com@sources.redhat.com>
Received: (qmail 12069 invoked by alias); 11 Nov 2004 21:12:26 -0000
Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:gdb-patches-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/gdb-patches/>
List-Post: <mailto:gdb-patches@sources.redhat.com>
List-Help: <mailto:gdb-patches-help@sources.redhat.com>, <http://sources.redhat.com/ml/#faqs>
Sender: gdb-patches-owner@sources.redhat.com
Received: (qmail 11966 invoked from network); 11 Nov 2004 21:12:19 -0000
Received: from unknown (HELO sj-iport-2.cisco.com) (171.71.176.71)
  by sourceware.org with SMTP; 11 Nov 2004 21:12:19 -0000
Received: from sj-core-5.cisco.com (171.71.177.238)
  by sj-iport-2.cisco.com with ESMTP; 11 Nov 2004 13:24:22 -0800
Received: from msnyder8600 (dhcp-128-107-165-92.cisco.com [128.107.165.92])
	by sj-core-5.cisco.com (8.12.10/8.12.6) with SMTP id iABLCFcp005201;
	Thu, 11 Nov 2004 13:12:16 -0800 (PST)
Message-ID: <002a01c4c833$2030e8a0$5ca56b80@msnyder8600>
From: "Michael Snyder" <michsnyd@cisco.com>
To: <gdb-patches@sources.redhat.com>
Cc: <jimb@redhat.com>, <ezannoni@redhat.com>
Subject: [RFA] stabs tweak -- forward reference types
Date: Thu, 11 Nov 2004 21:12:00 -0000
MIME-Version: 1.0
Content-Type: text/plain;
	format=flowed;
	charset="iso-8859-1";
	reply-type=original
Content-Transfer-Encoding: 7bit
X-SW-Source: 2004-11/txt/msg00247.txt.bz2

Jim, Elena, 

There was a change about two years ago, related to G++ and derived classes.
See "complain_about_struct_wipeout" for comments.

I have evidence that this change was too aggressive, and that it sometimes
interferes with the resolution of forward referenced struct pointer types.
Unfortunately, I'm not having any luck constructing a small testcase.

So I hope to convince you that the small change proposed below is harmless.
According to the comments, the earlier change was only concerned with
derived types.  If we check N_BASECLASS and see that a type has no
base classes, then it should be safe.  Yes?  

I've confirmed that the present change has positive side effects -- it 
results in forward references being resolved that were being left
unresolved before.  I don't know how to confirm that it has no negative
side effects, since I don't know how to test the earlier derived class issue.

Michael

2004-11-11  msnyder  <msnyder@redhat.com>

 * stabsread.c (read_struct_type): complain_about_struct_wipeout
 only if struct is a derived type.

Index: stabsread.c
===================================================================
RCS file: /cvs/src/src/gdb/stabsread.c,v
retrieving revision 1.77
diff -p -r1.77 stabsread.c
*** stabsread.c 30 Apr 2004 14:40:54 -0000 1.77
--- stabsread.c 11 Nov 2004 21:04:52 -0000
*************** read_struct_type (char **pp, struct type
*** 3294,3303 ****
    if (! (TYPE_CODE (type) == TYPE_CODE_UNDEF
           || TYPE_STUB (type)))
      {
!       complain_about_struct_wipeout (type);
  
!       /* It's probably best to return the type unchanged.  */
!       return type;
      }
  
    back_to = make_cleanup (null_cleanup, 0);
--- 3294,3309 ----
    if (! (TYPE_CODE (type) == TYPE_CODE_UNDEF
           || TYPE_STUB (type)))
      {
!       /* Well, the above should only apply if type *has* a baseclass.
!   Otherwise, we may interfere with the resolution of forward
!   referenced struct types.  */
!       if (TYPE_N_BASECLASSES (type) > 0)
!  {
!    complain_about_struct_wipeout (type);
  
!    /* It's probably best to return the type unchanged.  */
!    return type;
!  }
      }
  
    back_to = make_cleanup (null_cleanup, 0);