Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:12686
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Message-ID: <5.1.0.14.2.20040909111354.040b6570@127.0.0.1>
Date: Thu, 09 Sep 2004 11:17:18 -0700
To: john@coggeshall.org,internals@lists.php.net
In-Reply-To: <1094753081.26977.22.camel@unix-101-34.hq.communityconnect.
 com>
References: <5.1.0.14.2.20040909101402.02f6a650@127.0.0.1>
 <5.1.0.14.2.20040909101402.02f6a650@127.0.0.1>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Subject: Re: [PHP-DEV] Fwd: [ZEND-ENGINE-CVS] cvs: ZendEngine2 /
  README.ZEND_VM
From: andi@zend.com (Andi Gutmans)

At 02:04 PM 9/9/2004 -0400, John Coggeshall wrote:
>Personally I would like to see the hooks for op-codes stay in place, I
>think they offer a lot of possibilities for extensions.  Andi, what do
>you mean "create the handler execution architecture"? I'm a little
>confused to what you are referring to.

I mean that what we have to day is the function handler dispatch 
architecture (i.e. what it is in PHP 5.0.x). In PHP 4, we had switch().
Again, I think the opcode hook could be solved for people who want to do so 
by having such a version floating around in the address space for 
extensions that want to use it. It's a bit more memory but it's memory that 
should be shared.

>PS -- I hated the switch() method ;)

It was definitely uglier and that's why I made the change but in many cases 
the "goto and/or switch" dispatch methods are faster.

Andi