Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:73184
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.215.46 as permitted sender)
MIME-Version: 1.0
Sender: yohgaki@gmail.com
In-Reply-To: <CAGa2bXZdrxRP1558hBpoi_hDDWQtjGXuegiSJtPXpF5CLG8mdQ@mail.gmail.com>
References: <CAPhkiZx=0Vr5p=7VZJD4PeD0ub-khSVMw3G03QhO8CC4xxm11A@mail.gmail.com>
 <CAGa2bXZTU5HzM4oXyiu0PnS=W2Oz1uLP_BnyWBUXyYUTO0tMYg@mail.gmail.com>
 <CAPhkiZz_d_wFPgwwYXJD-A+tz0Uf9KQ6uXNi1HhZGWFF1=W91A@mail.gmail.com>
 <CAGa2bXY86ri7b=rquo8Vbbh5ZoHOkwGBRoPWPup-SVPMu0WEgg@mail.gmail.com>
 <CAPhkiZyEqcx+UTPzpyYjz9pxWJnCNDO9pUMQNxvTpnrxmez9nw@mail.gmail.com>
 <CAGa2bXb5rLU26NDDHkWDw-FYf=vh2J=LAGFrqU3A3JJpOTY_dA@mail.gmail.com>
 <CAPhkiZyyXuqqybs+q8rnp_eJZDU0De2wMqwfv5mWiihr9GsueQ@mail.gmail.com>
 <CAGa2bXboo3CGOP9x+fW0Q98aniBcXejkGSJv5=qygL9ia67FbA@mail.gmail.com>
 <CAPhkiZzaTCE6sR=Rt=4VBmVOCo+uSQMREpf-fg6hWBrbCW+RXQ@mail.gmail.com>
 <CAGa2bXYWVoX_05BdS0M-V5G+B=4v-W-DPbWT8-vPrtd=gJtcCg@mail.gmail.com>
 <CAJ26g5TfiwjECNH_GbU6OCwCJdCCRJxKLUBL6wtw89XNo8P2eg@mail.gmail.com>
 <CAGa2bXbeucsEK2X1_wvtY_zMXs6guQAWxEMFhpjhyyOopq3sog@mail.gmail.com>
 <CAEZPtU6J01hP2+xjRUhTY9+rq9dTZyPOv9Mes_Xt0jiogsCE9w@mail.gmail.com>
 <CAPhkiZx3iUo=4cJYkB7KPcD8TMs79kvLBrnBdGUZ4YqTq-+h-Q@mail.gmail.com>
 <CAPhkiZx_ztbVXFm10_4MSQyqkBDYVqBiCmNBdF4k7N9fev5uLQ@mail.gmail.com>
 <CAGa2bXafxzFTMRfn16JNDgWm2tPQeuYE8cdqg9BbRsKSy3S1QA@mail.gmail.com>
 <CAPhkiZxO44Cj7h_NP-5_ZtixLuFoWJrtdPsRabJHP5yTOrCf5g@mail.gmail.com>
 <CAGa2bXbtMQ2cQoe-X1S=Ab05iS7yph9x6Kw5DjfsHdrNEjGF4w@mail.gmail.com>
 <CAPhkiZxdUFjWrRm7SAtagLmtZXPDbsXp9Cv13u2ChKLsuuxtgA@mail.gmail.com>
 <CAGa2bXZ458+diX3FFh3tw_W28y9MZpipmWrPfQ8obo-Z3n9S8w@mail.gmail.com>
 <CAPhkiZz8aH0H9-rJQxkQAzC_LgAPgy4c75uS74xiDZ_30H7DAA@mail.gmail.com>
 <CAJ26g5STp=DeJ7u9k2Z97HG2+qG+knXOJ_7L1srtU03Ca3u_1Q@mail.gmail.com>
 <CAPhkiZxRdPOJyWtktvcwN-=Bp2M=fBE9i8jG4WgTnZYysdvibg@mail.gmail.com>
 <CAGa2bXbJhBu9htc1PY0Rr-_GW0Zp6=z6=Y=jqpee3O5KEsF1wg@mail.gmail.com>
 <CAPhkiZy69Z=15TNGHrnVmoJ8C+9CtAPf=dTaOkphqOb-V0n+4A@mail.gmail.com>
 <CAGa2bXYCdAwU-LUjHk7gKcQ5fnoqm_q3sSn=uG-591=9sCqYqA@mail.gmail.com>
 <CAGa2bXYBqD_5UW6fBahnR0ad7BQsdUnTiiQVBtcXefOH0BHv_w@mail.gmail.com>
 <CAPhkiZyG+i8j6mCa-K7PkwaxLBkmBuLZ9UNqznfewzUj4yRtOg@mail.gmail.com>
 <CAGa2bXbmzr+r3h9KhU4J+A8xONoPmAk9qvKttw5wKQ9F=ZWP5g@mail.gmail.com>
 <CAPhkiZxJGw-1Wj=-GEdgKEx9rw13UpX9_q7_mHA-b6k30Qd-UA@mail.gmail.com> <CAGa2bXZdrxRP1558hBpoi_hDDWQtjGXuegiSJtPXpF5CLG8mdQ@mail.gmail.com>
Date: Sun, 16 Mar 2014 09:57:51 +0900
Message-ID: <CAGa2bXap5gsKGp7ynz0L=sksdx03j5gmB6xv4Bjr8YAQ25F7-g@mail.gmail.com>
To: Andrey Andreev <narf@devilix.net>
Cc: Patrick Schaaf <bof@bof.de>, internals <internals@lists.php.net>, 
	Pierre Joye <pierre.php@gmail.com>
Content-Type: multipart/alternative; boundary=089e0112bf4445a86b04f4aecda1
Subject: Re: [PHP-DEV] Re: Revert session_serializer_name(), session_gc()
From: yohgaki@ohgaki.net (Yasuo Ohgaki)

--089e0112bf4445a86b04f4aecda1
Content-Type: text/plain; charset=UTF-8

Hi Andrey,

On Sun, Mar 16, 2014 at 9:45 AM, Yasuo Ohgaki <yohgaki@ohgaki.net> wrote:

> On Sun, Mar 16, 2014 at 3:33 AM, Andrey Andreev <narf@devilix.net> wrote:
>
>> >> > "write short circuit" omits "write" when $_SESSION hasn't change.
>> There
>> >> > is
>> >> > no point calling write API and writing to storage for the same data.
>> >>
>> >> "write short circuit" as I understand it, is an exact copy of the
>> >> 'lazy_write' option. This will be addressed in the previously
>> >> mentioned RFC that I'll post later today.
>> >
>> >
>> > No. It's not.
>> >
>> > "lazy_write" does not lock session data while "write short circuit"
>> does.
>> > In other words, "lazy_write" changes session behavior, but "write short
>> > circuit" does not.
>> > i.e. With locked session data, "write short circuit" would not change
>> how
>> > session manager works.
>>
>> I can't stress enough how confusing and potentially unsafe that is
>
>
> It does not make difference since many users use session_commit()
> to get more concurrency already. Simple documentation would be enough to
> understand what it does.
>
> There are freedom to shoot their own foot by themselves in any languages.
> We have better performance/more complex operations/convenience in return.
>

One more comment for this.

"write short cut" is designed to work with old(I mean existing) save
handlers.
Behavior does _not_ change at all. New save handlers that support new save
handler API work much faster. That's it.

Save handlers like memcahe/memcached updates last access time stamp
by reading data. These save handler may omit writes at all when session data
hasn't changed. I don't think explanation is needed why this gives us better
performance.

Regards,

--
Yasuo Ohgaki
yohgaki@ohgaki.net

--089e0112bf4445a86b04f4aecda1--