Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:66591
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: unknown (pb1.pair.com: domain zend.com does not designate 209.85.219.53 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <CAF+90c9oX_cr4Ah33BF5sYTbAicuZM6ibtXnWv4QCYW=6GGp=w@mail.gmail.com>
References: <CAF+90c-E7n-WZcqNg28bX+G7zgbzhKoNCwR6hf++tfAgSzmtTw@mail.gmail.com>
	<CAF+90c-wSqeff_2oY8HeY+gDn-ZnwJeXKQ+cz-1W6xcEBjs+xw@mail.gmail.com>
	<CA+9eiLtiOcFuxRXhpGXe5b9SpyVPB1_ixJ4t1GQ3J1bvBURMKA@mail.gmail.com>
	<CAF+90c92AwjX7Niro_tT5vOv+vHeY-AmgfizL5f9i1FYhKTz2A@mail.gmail.com>
	<CA+9eiLsWNV7H1zAr1U-wPafz9SzB=eR_C33RQKgqy1Tx1+X8gw@mail.gmail.com>
	<CAF+90c_Nt=Q-vTHO3iqC24mkwULMvV+KxpYpy1ug5=7em-x1eg@mail.gmail.com>
	<CA+9eiLuV_af9wAS_1zd_NYw=NXo1kJs990s5+bb_ZQ6h8=yW4Q@mail.gmail.com>
	<CA+9eiLvnmHA7fQonaLj-oRWp_d7HmEbtoejfqWrKQxK-DOTjMg@mail.gmail.com>
	<CAF+90c9oX_cr4Ah33BF5sYTbAicuZM6ibtXnWv4QCYW=6GGp=w@mail.gmail.com>
Date: Mon, 11 Mar 2013 23:42:18 +0400
Message-ID: <CA+9eiLs3hWwqLqBD+ozRLuOmfTZ8Zv-RRZqXJfUMe1zr=w1GTw@mail.gmail.com>
To: Nikita Popov <nikita.ppv@gmail.com>
Cc: PHP internals <internals@lists.php.net>, Zeev Suraski <zeev@zend.com>, Andi Gutmans <andi@zend.com>
Content-Type: multipart/alternative; boundary=f46d044788fded810204d7ab5e9f
Subject: Re: [PHP-DEV] Re: [VOTE] Allow non-scalar keys in foreach
From: dmitry@zend.com (Dmitry Stogov)

--f46d044788fded810204d7ab5e9f
Content-Type: text/plain; charset=UTF-8

On Mon, Mar 11, 2013 at 10:27 PM, Nikita Popov <nikita.ppv@gmail.com> wrote:

> On Mon, Mar 11, 2013 at 9:50 AM, Dmitry Stogov <dmitry@zend.com> wrote:
>
>> Hi Nikita,
>>
>> The patch looks good. I have just few comments
>>
>> - In ZEND_FE_FETCH handler PLAIN_OBJECT may have only STRING keys. I
>> didn't get why you added unreachable code for INT and NULL.
>>
>
> You are right about the NULL case, that can indeed not occur. But INT is
> possible, e.g. consider this:
>
> <?php
> foreach ((object) ['x','y','z'] as $k => $v) {
>     var_dump($k);
> }
>
> this will give you keys int(0), int(1), int(2).
>

Agree. I missed it.


> I'll remove the check for NULL.
>
>
> - At first, I fought, that it might be a good idea to change
>> zend_user_it_get_current_key() to return SUCCESS/FAILURE instead of
>> returning IS_NULL that has a special meaning. But after looking into the
>> FE_FETCH code before the commit I understood where this NULL came from. I
>> know that NULL key may not appear for plain array and objects but I'm not
>> sure about iterators and generators. Now IS_NULL keys may mean that
>> iterator returned it directly IS_NULL or may be it was returned because of
>> some error conditions. Probably it's not a problem. What do you think?
>>
>
> In foreach IS_NULL can't occur in an error condition, because the loop is
> already aborted when get_current_data provides NULL. IS_NULL can only
> happen when its explicitly provided (or handlers are incorrectly coded). I
> think the IS_NULL fallback is mainly important when the iterator is used
> for other things (like a dual it), to be sure that it'll always work. I
> don't think that it's important to distinguish between explicit NULL and
> failure NULL.
>

Agree as well.


> I have one more question: Right now I added the
> zend_hash_get_current_key_zval API in zend_hash.c/h, which is a bit ugly
> because it has a zval dependency (unlike all the other code in there) and
> requires me to forward-declare the zval struct. Would it be better to move
> this somewhere else, e.g. zend_API.c/h? If so, can I still keep the same
> name (with the zend_hash_ prefix)? If I move it to zend_API, I won't be
> able to do the IS_CONSISTENT check anymore, is that a problem?
>

I think we can move zval forward declaration (typedef struct _zval_struct
zval;) from zend.h into zend_type.h.

Thanks. Dmitry.

--f46d044788fded810204d7ab5e9f--