From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-172563-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 17992 invoked by alias); 10 May 2006 09:06:21 -0000
Received: (qmail 17977 invoked by uid 22791); 10 May 2006 09:06:19 -0000
X-Spam-Check-By: sourceware.org
Received: from py-out-1112.google.com (HELO py-out-1112.google.com) (64.233.166.177)     by sourceware.org (qpsmtpd/0.31) with ESMTP; Wed, 10 May 2006 09:06:17 +0000
Received: by py-out-1112.google.com with SMTP id c30so2089844pyc         for <gcc-patches@gcc.gnu.org>; Wed, 10 May 2006 02:06:15 -0700 (PDT)
Received: by 10.35.91.15 with SMTP id t15mr964773pyl;         Wed, 10 May 2006 02:06:14 -0700 (PDT)
Received: by 10.35.88.2 with HTTP; Wed, 10 May 2006 02:06:14 -0700 (PDT)
Message-ID: <19c433eb0605100206y4bbbcbbbo535c31582ee5fcc@mail.gmail.com>
Date: Wed, 10 May 2006 10:44:00 -0000
From: "=?ISO-8859-1?Q?Fran=E7ois-Xavier_Coudert?=" <fxcoudert@gmail.com>
To: GFortran <fortran@gcc.gnu.org>, gcc-patches <gcc-patches@gcc.gnu.org>
Subject: [patch]
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
X-SW-Source: 2006-05/txt/msg00383.txt.bz2

:ADDPATCH fortran:

I tried to understand PR 24549
(http://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D24549), an ICE after
error recovery. We see a statement we don't understand, we skip to the
next and there is the ICE, because we use a value for gfc_new_block
that has been freed (it was allocated for the last statement). I don't
see any reason why, in reject_statement(), we don't set gfc_new_block
=3D NULL, leading to the following patch:

Index: parse.c
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- parse.c     (revision 113603)
+++ parse.c     (working copy)
@@ -1292,7 +1292,7 @@
 static void
 reject_statement (void)
 {
-
+  gfc_new_block =3D NULL;
   gfc_undo_symbols ();
   gfc_clear_warning ();
   undo_new_statement ();


So, I'd like to know, from someone who's accustomed to gfortran error
recovery mechanism, if there is a reason for not setting clearing
gfc_new_block. And, if not, is the above patch approved (I'll add a
testcase and the obvious ChangeLog entries, of course).

Thanks,
FX

PS: I know this is patch submission is not in very good order, but I'm
at work and it's such a simple patch, I couldn't keep it for later.