Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:2311
Return-Path: <hewei@ied.org.cn>
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Delivered-To: mailing list internals@lists.php.net
Received: (qmail 98665 invoked from network); 13 Jun 2003 10:06:32 -0000
Received: from unknown (HELO ied.ac.cn) (211.99.43.218)
  by pb1.pair.com with SMTP; 13 Jun 2003 10:06:32 -0000
Received: (qmail 15838 invoked by uid 610); 13 Jun 2003 17:22:43 +0800
Received: from localhost (sendmail-bs@127.0.0.1)
  by localhost with SMTP; 13 Jun 2003 17:22:43 +0800
Date: Fri, 13 Jun 2003 17:22:43 +0800 (CST)
X-X-Sender: hewei@web.ied.ac.cn
To: internals@lists.php.net
cc: Andrei Zmievski <andrei@php.net>
In-Reply-To: <Pine.LNX.4.44L1.0305232341150.21698-100000@web.ied.ac.cn>
Message-ID: <Pine.LNX.4.44L1.0306131656560.31342-100000@web.ied.ac.cn>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Subject: The aggregation/Zend hash bug
From: hewei@ied.org.cn (Wei He)
References: <Pine.LNX.4.44L1.0305232341150.21698-100000@web.ied.ac.cn>

Hi,

I didn't receive and emails regarding to my discovery of the bug. If it 
wasn't because the mailing list or my email server's problem, it would
turned out to be that nobody cares about it. 

Then what if, according to my own debugging, the problem is related
to zend hash?

I guess the answer to the following error is that unset($a) freed the 
object but not the all the hash tables (aggretated methods) associated 
with it.

So a new object after the unset will reuse the same memory location and
zend_hash_add will return a failure.

zend_hash.c: in function zend_hash_add_or_update

	while (p != NULL) {
		if ((p->h == h) && (p->nKeyLength == nKeyLength)) {
			if (!memcmp(p->arKey, arKey, nKeyLength)) {
				if (flag & HASH_ADD) {
					return FAILURE;
				}
				HANDLE_BLOCK_INTERRUPTIONS();

Can anyone understading zend hash functions well check why 
this would happen? 

Shouldn't unset($a) clean everything related to it?

Please! I can't live without aggregation and I can't
work any longer due to the bug.

If you really don't think it a bug. Would anyone kindly reply 
me a message so that at least I know you guys have received my 
report?

Wei He


On Fri, 23 May 2003, Wei He wrote:

> Hi,
> 
> This short script can reproduce the aggregation bug I reported above.
> 
> class bar {
>  
>    function doit()
>    {
>       print " Doing...\n";
>    }
> }
> 
> class foo {
> 
>    function foo()
>    {
>        print_r(aggregation_info($this));
>        aggregate($this, "bar");
>    }
>    
>    function spawn()
>    {
>        return new foo();
>    }
> }
> 
> $a = new foo();
> $a->doit();
> $b = $a->spawn();
> $b->doit();
> unset($a);
> $c = $b->spawn();
> $c->doit();
> 
> Besides 'unset($a)', '$a = new foo()' will also cause the same problem.
> 
> Wei He
> 
>