Newsgroups: php.internals Path: news.php.net Xref: news.php.net php.internals:66628 Return-Path: Mailing-List: contact internals-help@lists.php.net; run by ezmlm Delivered-To: mailing list internals@lists.php.net Received: (qmail 6414 invoked from network); 14 Mar 2013 22:16:01 -0000 Received: from unknown (HELO lists.php.net) (127.0.0.1) by localhost with SMTP; 14 Mar 2013 22:16:01 -0000 Authentication-Results: pb1.pair.com smtp.mail=v-mafick@outlook.com; spf=pass; sender-id=pass Authentication-Results: pb1.pair.com header.from=v-mafick@outlook.com; sender-id=pass Received-SPF: pass (pb1.pair.com: domain outlook.com designates 65.54.190.89 as permitted sender) X-PHP-List-Original-Sender: v-mafick@outlook.com X-Host-Fingerprint: 65.54.190.89 bay0-omc2-s14.bay0.hotmail.com Windows 2000 SP4, XP SP1 Received: from [65.54.190.89] ([65.54.190.89:12722] helo=bay0-omc2-s14.bay0.hotmail.com) by pb1.pair.com (ecelerity 2.1.1.9-wez r(12769M)) with ESMTP id 02/54-14708-D1C42415 for ; Thu, 14 Mar 2013 17:15:59 -0500 Received: from BAY176-W38 ([65.54.190.124]) by bay0-omc2-s14.bay0.hotmail.com with Microsoft SMTPSVC(6.0.3790.4675); Thu, 14 Mar 2013 15:15:54 -0700 X-EIP: [O2kd6ZudYSO0TxX7SEy2LaksF1MCpwrH] X-Originating-Email: [v-mafick@outlook.com] Message-ID: Content-Type: multipart/mixed; boundary="_42f211d3-7c8d-499c-8f0b-08a02d81f4af_" To: PHP Internals Date: Thu, 14 Mar 2013 15:15:54 -0700 Importance: Normal In-Reply-To: References: ,<51314835.6060805@oracle.com>,,,, MIME-Version: 1.0 X-OriginalArrivalTime: 14 Mar 2013 22:15:54.0889 (UTC) FILETIME=[7EF36F90:01CE2101] Subject: RE: [PHP-DEV] Current Status of O+ on Windows From: v-mafick@outlook.com (Matt Ficken) --_42f211d3-7c8d-499c-8f0b-08a02d81f4af_ Content-Type: multipart/alternative; boundary="_5c47c875-3ccb-4921-97f0-bba460f0171e_" --_5c47c875-3ccb-4921-97f0-bba460f0171e_ Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable =0A= =0A= Earlier I had posted about testing O+ with Apache using Symfony's PhpUnit t= ests... I had an issue where it seems that test methods (which PhpUnit accesses usi= ng ReflectionMethod) were getting optimized out and replaced with an empty = method (so all PhpUnit tests passed). I have now written a PHPT to try to reproduce this issue (attached in case = it helps)=2C which passes/doesn=92t have this=0A= issue.=0A= =0A= =0A= =0A= I don=92t think this issue is full=0A= resolved/closed=2C but it=92s a low-priority edge case that can be fixed la= ter. =0A= =0A= Symfony runs with O+(see O+=0A= README for proper config)=85. This issue doesn=92t affect a normal Symfony = app=0A= running on Apache with O+... users aren't likely to have this problem. If users do have this problem=2C I have found that turning off all optimiza= tion passes except pass 3 and pass 10 (using the optimization_level directi= ve) seems to fix this problem. =0A= =0A= The reason I found this problem is that I=0A= run multiple PhpUnit tests on each Apache instance=2C one test is run at a = time=2C=0A= but multiple tests are run against each instance/process during its lifetim= e.=0A= This is a little more accurate (apache doesn=92t reboot between each HTTP= =0A= request) and makes testing much faster=2C otherwise I have to start and sto= p=0A= apache between each test=2C which is really slow.=0A= =0A= =0A= =0A= I have confirmed this by=0A= restarting apache between each PhpUnit test=2C and I get about=0A= the same errors and failures as I do without O+. Additionally=2C I get a bu= nch of=0A= crashes=2C which I will be investigating further.=0A= =0A= =0A= =0A= Regardless of the crashes=2C this=0A= issue is not critical/low priority now since all the PhpUnit tests actually= now=0A= run (reflectionmethods aren=92t getting optimized out).=0A= =0A= =0A= =0A= =0A= =0A= Regards=0A= =0A= -Matt=0A= =0A= > Date: Sat=2C 2 Mar 2013 09:43:34 +0100 > From: tyra3l@gmail.com > To: zeev@zend.com > CC: internals@lists.php.net=3B dmitry@zend.com > Subject: Re: [PHP-DEV] Current Status of O+ on Windows >=20 > On Sat=2C Mar 2=2C 2013 at 9:39 AM=2C Zeev Suraski wrote: >=20 > > > -----Original Message----- > > > From: Ferenc Kovacs [mailto:tyra3l@gmail.com] > > > Sent: Saturday=2C March 02=2C 2013 10:15 AM > > > To: Pierre Joye=3B Dmitry Stogov > > > Cc: Christopher Jones=3B Matt Ficken=3B internals@lists.php.net > > > Subject: Re: [PHP-DEV] Current Status of O+ on Windows > > > > > > > > Did you experiment with any options? > > > > > Putting the date of your O+ snapshot would also be handy. > > > > > > > > Latest from here are used: > > > > http://windows.php.net/downloads/pecl/snaps/Optimizer/7.0.0-dev/ > > > > > > > > Dates are included. It would be nice to have it in the report as we= ll=2C > > > > but we always use latest revision. It would however be much easier = if > > > > there were PECL releases. > > > > > > > > > > > +1 > > > Dmitry=2C is there any objection against creating a pecl release? > > > Could you tag the first version? > > > > The current vote that's going on right now deals with putting the exten= sion > > into PHP itself. If that happens (which seems awfully likely at this > > point)=2C why do we need it in PECL? We'd very much rather invest our = very > > limited cycles into the code itself. We're probably roughly a week awa= y > > from having builds as a part of the PHP 5.5 snaps. > > > > Zeev > > >=20 > I see. > so no O+ for <5.5? > having a pecl release would be a small amount of work=2C which I would gl= ad > to help with. >=20 > --=20 > Ferenc Kov=E1cs > @Tyr43l - http://tyrael.hu = --_5c47c875-3ccb-4921-97f0-bba460f0171e_ Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable
=0A= =0A=

Earlier I had posted a= bout testing O+ with Apache using Symfony's PhpUnit tests...


I had an issue where it seems tha= t test methods (which PhpUnit accesses using ReflectionMethod) were getting= optimized out and replaced with an empty method (so all PhpUnit tests pass= ed).


I have now wri= tten a PHPT to try to reproduce this issue (attached in case it helps)=2C w= hich passes/doesn=92t have this=0A= issue.

=0A= =0A=

 =3B

=0A= =0A=

I don=92t think this i= ssue is full=0A= resolved/closed=2C but it=92s a low-priority edge case that can be fixed la= ter.

 =3B=0A= =0A=

Symfony runs with O+(s= ee O+=0A= README for proper config)=85. This issue doesn=92t affect a normal Symfony = app=0A= running on Apache with O+... users aren't likely to have this problem.


If users do hav= e this problem=2C I have found that turning off all optimization passes exc= ept pass 3 and pass 10 (using the optimization_level directive) seems to fi= x this problem.


 =3B

=0A= =0A=

The reason I found thi= s problem is that I=0A= run multiple PhpUnit tests on each Apache instance=2C one test is run at a = time=2C=0A= but multiple tests are run against each instance/process during its lifetim= e.=0A= This is a little more accurate (apache doesn=92t reboot between each HTTP= =0A= request) and makes testing much faster=2C otherwise I have to start and sto= p=0A= apache between each test=2C which is really slow.

=0A= =0A=

 =3B

=0A= =0A=

I have confirmed this = by=0A= restarting apache between each PhpUnit test=2C and I get about=0A= the same errors and failures as I do without O+. Additionally=2C I get a bu= nch of=0A= crashes=2C which I will be investigating further.

=0A= =0A=

 =3B

=0A= =0A=

Regardless of the cras= hes=2C this=0A= issue is not critical/low priority now since all the PhpUnit tests actually= now=0A= run (reflectionmethods aren=92t getting optimized out).

=0A= =0A=

 =3B

=0A= =0A=

 =3B

=0A= =0A=

Regards

=0A= =0A=

-Matt

=0A= =0A=

>=3B Date:= Sat=2C 2 Mar 2013 09:43:34 +0100
>=3B From: tyra3l@gmail.com
>= =3B To: zeev@zend.com
>=3B CC: internals@lists.php.net=3B dmitry@zend.= com
>=3B Subject: Re: [PHP-DEV] Current Status of O+ on Windows
>= =3B
>=3B On Sat=2C Mar 2=2C 2013 at 9:39 AM=2C Zeev Suraski <=3Bzee= v@zend.com>=3B wrote:
>=3B
>=3B >=3B >=3B -----Original Me= ssage-----
>=3B >=3B >=3B From: Ferenc Kovacs [mailto:tyra3l@gmail= .com]
>=3B >=3B >=3B Sent: Saturday=2C March 02=2C 2013 10:15 AM>=3B >=3B >=3B To: Pierre Joye=3B Dmitry Stogov
>=3B >=3B &g= t=3B Cc: Christopher Jones=3B Matt Ficken=3B internals@lists.php.net
>= =3B >=3B >=3B Subject: Re: [PHP-DEV] Current Status of O+ on Windows>=3B >=3B >=3B
>=3B >=3B >=3B >=3B >=3B Did you experim= ent with any options?
>=3B >=3B >=3B >=3B >=3B Putting the dat= e of your O+ snapshot would also be handy.
>=3B >=3B >=3B >=3B>=3B >=3B >=3B >=3B Latest from here are used:
>=3B >=3B &= gt=3B >=3B http://windows.php.net/downloads/pecl/snaps/Optimizer/7.0.0-de= v/
>=3B >=3B >=3B >=3B
>=3B >=3B >=3B >=3B Dates are = included. It would be nice to have it in the report as well=2C
>=3B &g= t=3B >=3B >=3B but we always use latest revision. It would however be m= uch easier if
>=3B >=3B >=3B >=3B there were PECL releases.
&= gt=3B >=3B >=3B >=3B
>=3B >=3B >=3B >=3B
>=3B >=3B = >=3B +1
>=3B >=3B >=3B Dmitry=2C is there any objection against = creating a pecl release?
>=3B >=3B >=3B Could you tag the first ve= rsion?
>=3B >=3B
>=3B >=3B The current vote that's going on r= ight now deals with putting the extension
>=3B >=3B into PHP itself.= If that happens (which seems awfully likely at this
>=3B >=3B poin= t)=2C why do we need it in PECL? We'd very much rather invest our very
= >=3B >=3B limited cycles into the code itself. We're probably roughly = a week away
>=3B >=3B from having builds as a part of the PHP 5.5 sn= aps.
>=3B >=3B
>=3B >=3B Zeev
>=3B >=3B
>=3B
= >=3B I see.
>=3B so no O+ for <=3B5.5?
>=3B having a pecl rel= ease would be a small amount of work=2C which I would glad
>=3B to hel= p with.
>=3B
>=3B --
>=3B Ferenc Kov=E1cs
>=3B @Tyr43= l - http://tyrael.hu
= --_5c47c875-3ccb-4921-97f0-bba460f0171e_-- --_42f211d3-7c8d-499c-8f0b-08a02d81f4af_--