From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <pthreads-win32-return-600-listarch-pthreads-win32=sources.redhat.com@sources.redhat.com>
Received: (qmail 3123 invoked by alias); 3 Jan 2003 15:30:49 -0000
Mailing-List: contact pthreads-win32-help@sources.redhat.com; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:pthreads-win32-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/pthreads-win32/>
List-Post: <mailto:pthreads-win32@sources.redhat.com>
List-Help: <mailto:pthreads-win32-help@sources.redhat.com>, <http://sources.redhat.com/ml/#faqs>
Sender: pthreads-win32-owner@sources.redhat.com
Received: (qmail 3114 invoked from network); 3 Jan 2003 15:30:44 -0000
Received: from unknown (HELO mailout09.sul.t-online.com) (194.25.134.84)
  by 209.249.29.67 with SMTP; 3 Jan 2003 15:30:44 -0000
Received: from fwd07.sul.t-online.de 
	by mailout09.sul.t-online.com with smtp 
	id 18UTmI-0000MF-0D; Fri, 03 Jan 2003 16:30:30 +0100
Received: from pseudo.t-online.de (520069358420-0001@[217.225.139.223]) by fwd07.sul.t-online.com
	with esmtp id 18UTm1-0Lr70yC; Fri, 3 Jan 2003 16:30:13 +0100
Message-Id: <5.1.0.14.0.20030103155101.00c88690@pop.t-online.de>
Date: Fri, 03 Jan 2003 15:30:00 -0000
To: pthreads-win32@sources.redhat.com
From: beckmann.bremen@t-online.de (Thomas Beckmann)
Subject: pthread_key_create destructor
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-Sender: 520069358420-0001@t-dialin.net
X-SW-Source: 2003/txt/msg00000.txt.bz2

Hi,

I have a problem with pthread_key_create ().
I'm using the latest(2002-11-4) pre-build pthreadVC.dll on Windows XP with 
MSVC++ 6.0.
I wrote the following example to explain the problem:

#include <pthread.h>
#include <windows.h>
#include <stdio.h>

pthread_key_t key;

void destructor (void *ptr)
{
	printf ((char*)ptr);
	free (ptr);
}

int main (int argc, char *argv[])
{
	char *ptr;

	pthread_key_create (&key, &destructor);
	if ((ptr = pthread_getspecific(key)) == NULL) {
		ptr = strdup ("test\n");
		pthread_setspecific (key, ptr);
	}
	return 0;
}

I think, this is a valid example, and it works fine as a release version 
but the debug version crashes inside the destructor() function .
The address of ptr inside destructor() is the same as inside main() but 
when having a look at the memory pointed to by ptr, it is initialized to 
zero at the end of main but unspecified inside of destructor().
It seems, that the address of ptr is already outside the applications 
address space or somehow else invalid, so the destructor is called too late.


Thanks for any help,

Thomas Beckmann