Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:80969
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain zeyos.com designates 109.70.220.166 as permitted sender)
Date: Thu, 22 Jan 2015 11:38:04 +0100
To: Nikita Popov <nikita.ppv@gmail.com>
Cc: Dmitry Stogov <dmitry@zend.com>,
	Xinchen Hui <xinchen.h@zend.com>,
	"internals@lists.php.net" <internals@lists.php.net>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset="UTF-8"
Message-ID: <20150122103807.DD8255F8EE@mx.zeyon.net>
Subject: Re: [PHP-DEV] Improvements to for-each implementation
From: ben.coutu@zeyos.com (Benjamin Coutu)

Hi Nikita,=0A=0AI would suggest using the proposal for the by-value case an=
d sticking with the current behavior for the by-reference case as you sugge=
sted. Granted, we then cannot remove the internal pointer all together, but=
 we would just use it for the less common by-reference case as well as for =
the legacy reset/next functions, of course. This would still improve most f=
or-each traversals. At least we then wouldn't have to find a replacement so=
lution for rest/prev/next/end. Moreover we can then move the internal posit=
ion pointer out of the hashtable structure into zend_array because it is on=
ly used for userland arrays, not other hash tables (such as object properti=
es or internal structures that build upon the hashtable struct).=0A=0AThank=
s,=0A=0ABen=0A=0A=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Original =3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=0AFrom: Nikita Popov <nikita.ppv@gmail.com>=0ATo: Benjamin Cou=
tu <ben.coutu@zeyos.com>=0ADate: Thu, 22 Jan 2015 10:53:19 +0100=0ASubject:=
 Re: [PHP-DEV] Improvements to for-each implementation=0A=0ADoing this was =
the idea I had in mind as well, i.e. change the semantics of=0Aforeach to s=
ay that it will always work on a copy for by-value iteration=0A(which ironi=
cally avoids having to actually copy it). Note that this will=0Adiffer from=
 the current behavior in a number of ways. In particular it=0Ameans that ch=
anges to arrays that were references prior to iteration will=0Anot influenc=
e the iteration.=0A=0AThe real question is what we should do in the by-refe=
rence case. Given that=0Awe need to acquire references to elements of the o=
riginal array we can't=0Areasonably work with copy-semantics (at least I do=
n't see how). So would we=0Ajust stick with the previous behavior (using th=
e hash position hack) for=0Athat?=0A=0ANikita