Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:54209
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain nebm.ist.utl.pt from 193.136.128.21 cause and error)
Content-Type: text/plain; charset=utf-8; format=flowed; delsp=yes
To: "internals@lists.php.net" <internals@lists.php.net>
References: <CABBJUpeqWM3_Gw=dzNMknE80dPe0fQCDXO9DPL6J+Egv6ktACg@mail.gmail.com> <CABBJUpd1Xssx+oa-f8__bTqmnavkN6d=sX19N_vvsvx8JXtFmQ@mail.gmail.com> <op.vy7x9iv0idpuyk@clk-0081.mshome.net> <CABBJUpeTgyZKGg9XxVMaaHiLJPs+HWhwAjFYUeqN0C-Ut+0Ydw@mail.gmail.com>
Date: Tue, 26 Jul 2011 12:31:07 +0100
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Organization: =?utf-8?Q?N=C3=BAcleo_de_Eng=2E_Biom=C3=A9di?=
 =?utf-8?Q?ca_do_I=2ES=2ET=2E?=
Message-ID: <op.vy74x5zmidpuyk@clk-0081.mshome.net>
In-Reply-To: <CABBJUpeTgyZKGg9XxVMaaHiLJPs+HWhwAjFYUeqN0C-Ut+0Ydw@mail.gmail.com>
User-Agent: Opera Mail/11.50 (Win32)
Subject: Re: [PHP-DEV] Re: [Patches]#36944, #54556, #55211(and svn karma
 applying)
From: glopes@nebm.ist.utl.pt ("Gustavo Lopes")

Em Tue, 26 Jul 2011 12:14:15 +0100, Laruence <laruence@php.net> escreveu:

> 2011/7/26 Gustavo Lopes <glopes@nebm.ist.utl.pt>:
>> Em Tue, 26 Jul 2011 06:42:50 +0100, Laruence <laruence@php.net>  
>> escreveu:
>>
>>> 2011/7/25 Laruence <laruence@php.net>:
>>>>   #54556  array access to empty var does not trigger a notice
>>>> https://bugs.php.net/bug.php?id=54556
>>
>> Won't this be affected?
>>
>> <?php
>> $foo[] = 'bar';
>>
> no, this situation won't be affected

Ah, if this is only for read situations, a notice seems appropriate.

>>>   #55128 SplFixedArray::fromArray() does not use extended class
>>
>> This doesn't look like a good idea. The constructor is not called (and  
>> in general can't be, since we can't know how the subclass constructor  
>> looks
>> like), so the new object can be in an illegal state.
>>
> After much thought, I think there maybe no reason for fixedarray
> extended class  to define it's own constructor, since we provide
> fixedarrary::__construct and fromArray already, and extended class
> should only get an instance through calling them.
>
> so I think we can make fixedarray::__construct final to make this  
> feature works.

I disagree. Besides being a BC break, there's no good reason to make the  
constructor final and it has the disadvantage of severely limiting the  
subclasses.

Instead of using LSB, I think a better option would be to provided a  
protected final instance method that would do this special form of  
initialization from an array. Subclasses could then write their own static  
method (possibly with the same name) for initializing from an array or  
even accept an array as a constructor argument. This is a bit more work  
and less magic than LSB, but much more versatile.

-- 
Gustavo Lopes