Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:71369
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.217.170 as permitted sender)
MIME-Version: 1.0
Sender: yohgaki@gmail.com
In-Reply-To: <1390227486.3439.134.camel@guybrush>
References: <CAGa2bXZEXx7H+atr3LK8OLLQGSd9a-6+Gey+RhPwP0d7EvTBmw@mail.gmail.com>
 <52DCDDDB.70000@tobin.nl> <CAGa2bXZ8=XYE3kK8iBuwirb-u+OUTDve8USDzb8DaNSgMNvJcg@mail.gmail.com>
 <1390227486.3439.134.camel@guybrush>
Date: Tue, 21 Jan 2014 08:06:25 +0900
Message-ID: <CAGa2bXaPV2uftOPJ3i++qYu2azyV-UTqLtzKfgAfCHgCa-J6AQ@mail.gmail.com>
To: =?UTF-8?Q?Johannes_Schl=C3=BCter?= <johannes@schlueters.de>
Cc: Bas van Beek <bas@tobin.nl>, "internals@lists.php.net" <internals@lists.php.net>
Content-Type: multipart/alternative; boundary=089e0160b79a48d5cf04f06ef325
Subject: Re: [PHP-DEV] [VOTE] Introduce session.lock, session.lazy_write and session.lazy_destory
From: yohgaki@ohgaki.net (Yasuo Ohgaki)

--089e0160b79a48d5cf04f06ef325
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Hi Johanns and Stas,

  - session_close() - close session, *do* write, unlock file

We have session_write_close() (session_commit() is alias), so
it may be better to remain as it is.

On Mon, Jan 20, 2014 at 11:18 PM, Johannes Schl=C3=BCter
<johannes@schlueters.de>wrote:

> I do think this also should be application defined, not configuration
> defined. So using functions for that is better than ini options. Only
> the individual script nows what it needs, global configuration for
> tuning one application will break another on the same system and makes
> it harder to write portable appliations.
>

Stas and your discussion sounds reasonable to me, too. I don't mind
files_ext implementation to be a PECL module as a reference implementation.

It would be a module that does not provide any function, but only a
files_ext
save handler and lock INI.

session_unlock() is a simple new function and it does not require RFC, but
it is mentioned in it. However, unlocking session after start makes thing
more
complex for save handlers. It is better to specify when session is started.
Therefore, it would be better to pass new behavior options to
session_start().

session_start(array('lock'=3D>false, 'lazy_write'=3D>true,
'lazy_destroy'=3D>120));

It may also set other options for session like save handler, serializer,
expire, etc.

Since session.lazy_destroy is required for reliable session_regenerate_id()
operation, it remains but not exposed as INI option. Users has to pass
option
to session_start() to disable it.

In short, no new INI is added. Option parameter is added to session_start()=
.
I'll add new API for save handler to get start options.

I would like to change the RFC like this.
Any comments?

Regards,

--
Yasuo Ohgaki
yohgaki@ohgaki.net

--089e0160b79a48d5cf04f06ef325--