From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gdb-patches-return-47330-listarch-gdb-patches=sources.redhat.com@sourceware.org>
Received: (qmail 28729 invoked by alias); 23 Jan 2007 15:25:05 -0000
Received: (qmail 28672 invoked by uid 22791); 23 Jan 2007 15:25:02 -0000
X-Spam-Check-By: sourceware.org
Received: from mail.codesourcery.com (HELO mail.codesourcery.com) (65.74.133.4)     by sourceware.org (qpsmtpd/0.31) with ESMTP; Tue, 23 Jan 2007 15:24:51 +0000
Received: (qmail 15613 invoked from network); 23 Jan 2007 15:24:49 -0000
Received: from unknown (HELO localhost) (jimb@127.0.0.2)   by mail.codesourcery.com with ESMTPA; 23 Jan 2007 15:24:49 -0000
To: gdb-patches@sourceware.org
Subject: Re: [rfc] Plug memory leaks during gdbarch initialization
References: <20070123144523.GA6339@nevyn.them.org>
From: Jim Blandy <jimb@codesourcery.com>
Date: Tue, 23 Jan 2007 15:25:00 -0000
In-Reply-To: <20070123144523.GA6339@nevyn.them.org> (Daniel Jacobowitz's message of "Tue, 23 Jan 2007 09:45:23 -0500")
Message-ID: <m3irexvk8i.fsf@codesourcery.com>
User-Agent: Gnus/5.11 (Gnus v5.11) Emacs/22.0.50 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-IsSubscribed: yes
Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Id: <gdb-patches.sourceware.org>
List-Subscribe: <mailto:gdb-patches-subscribe@sourceware.org>
List-Archive: <http://sourceware.org/ml/gdb-patches/>
List-Post: <mailto:gdb-patches@sourceware.org>
List-Help: <mailto:gdb-patches-help@sourceware.org>, <http://sourceware.org/ml/#faqs>
Sender: gdb-patches-owner@sourceware.org
X-SW-Source: 2007-01/txt/msg00479.txt.bz2


Daniel Jacobowitz <drow@false.org> writes:
> The patch is an improvement, but doesn't fix that wart.  I don't really
> want to check it in without fixing the underlying problem but I don't
> see a good way... anyone have a suggestion?

I could imagine having each file's _initialize function start with:

void
_initialize_foo (void)
{
  static initialized;

  if (initialized)
    return;
  initialized = 1;

  ...
}

A sed script could get the job done.  Then, they could call each other
with wild abandon.  For extra points, you could do:

void
_initialize_foo (void)
{
  static int initialized;

  assert (initialized != 1);
  if (initialized)
    return;
  initialized = 1;

  ...

  initialized = 2;
}

So you'd catch circular dependencies.