From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-46810-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 17929 invoked by alias); 22 Feb 2002 03:05:43 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 17748 invoked from network); 22 Feb 2002 03:05:34 -0000
Received: from unknown (HELO mail.cs.tu-berlin.de) (130.149.17.13)
  by sources.redhat.com with SMTP; 22 Feb 2002 03:05:34 -0000
Received: from platon.cs.tu-berlin.de (matzmich@platon.cs.tu-berlin.de [130.149.25.107])
	by mail.cs.tu-berlin.de (8.9.3/8.9.3) with ESMTP id EAA12296;
	Fri, 22 Feb 2002 04:04:26 +0100 (MET)
Received: (from matzmich@localhost)
	by platon.cs.tu-berlin.de (8.9.3/8.9.3) id EAA17069;
	Fri, 22 Feb 2002 04:04:26 +0100 (MET)
Date: Thu, 21 Feb 2002 19:19:00 -0000
From: Michael Matz <matzmich@cs.tu-berlin.de>
X-X-Sender:  <matzmich@platon>
To: Andreas Jaeger <aj@suse.de>, Jakub Jelinek <jakub@redhat.com>
cc: <gcc@gcc.gnu.org>
Subject: Re: 252.eon / current GCC
In-Reply-To: <ho7kp72fmc.fsf@gee.suse.de>
Message-ID: <Pine.GSO.4.33.0202220357140.14671-100000@platon>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-SW-Source: 2002-02/txt/msg01250.txt.bz2

Hi,

On Thu, 21 Feb 2002, Andreas Jaeger wrote:

> [... a compile error with 252.eon from SPEC2000 ...]

A reduced testcase for this problem is:

-----------------
template <class T>
class Wrapper {
  public:
    Wrapper (T& a);
    Wrapper (const Wrapper<char>& ref);
};

template <class T>
class Element {
public:
  T * operator[](int x);
};

void test()
{
  char bla = 42;
  Element< Wrapper <unsigned char> > elem;
  elem[1][1] = Wrapper<char> (bla);
}
------------------

The error is cause by this patch:

2002-02-20  Jakub Jelinek  <jakub@redhat.com>
        * typeck.c (cp_pointer_int_sum): Renamed from
        pointer_int_sum, call pointer_int_sum.

In that process a certain thing got lost, namely the completion of the
inner TREE_TYPE on which size_in_bytes() is applied.  The below patch
fixes it for me.  I did not test the patch further.  Neither bootstrapped
nor regtested.  No time because of moving, sorry.


Ciao,
Michael.
-- 
2002-02-21  Michael Matz  <matz@suse.de>

	* typeck.c (cp_pointer_int_sum): Complete inner type which
	is used later by size_in_bytes().

Index: typeck.c
===================================================================
RCS file: /cvsroot/gcc/gcc/gcc/cp/typeck.c,v
retrieving revision 1.388
diff -u -p -r1.388 typeck.c
--- typeck.c	20 Feb 2002 23:05:42 -0000	1.388
+++ typeck.c	22 Feb 2002 02:54:39 -0000
@@ -4076,8 +4076,17 @@ cp_pointer_int_sum (resultcode, ptrop, i
      enum tree_code resultcode;
      register tree ptrop, intop;
 {
-  if (!complete_type_or_else (TREE_TYPE (ptrop), ptrop))
+  tree res_type = TREE_TYPE (ptrop);
+
+  if (!complete_type_or_else (res_type, ptrop))
     return error_mark_node;
+
+  /* pointer_int_sum() uses size_in_bytes() on the TREE_TYPE(res_type)
+     in certain circumstance (when it's valid to do so).  So we need
+     to make sure it's complete.  We don't need to check here, if we
+     can actually complete it at all, as those checks will be done in
+     pointer_int_sum() anyway.  */
+  (void) complete_type (TREE_TYPE (res_type));

   return pointer_int_sum (resultcode, ptrop, fold (intop));
 }