Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:30195
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain typo3.org from 217.72.131.73 cause and error)
Mime-Version: 1.0 (Apple Message framework v752.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-ID: <7A58222C-FA15-47D5-ACFC-0B14CF4D12A7@typo3.org>
Content-Transfer-Encoding: 7bit
Date: Fri, 15 Jun 2007 12:25:07 +0200
To: internals@lists.php.net
Subject: PHP6 an unicode in practice - wanted behaviour for array indices?
From: robert@typo3.org (Robert Lemke)

Hi folks,

I am currently developing the next major version of TYPO3, a PHP  
based content management system. As we take advantage of the new  
Unicode features (among other things) it will be completely based on  
and rely on PHP6 with unicode.semantics turned on. (Sidenode: My veto  
would be that unicode.semantics would be turned on by default, but I  
trust on that you take the right decision and then we'll adapt to  
it). We bypassed all backward compatibility issues by just starting  
from scratch again ...

Anyway, using the latest CVS version, I stumbled over a behaviour I  
didn't expect and I'd like to ask you if that is a wanted behaviour  
or a bug.

Consider the following code (unicode semantics = on):

    preg_match('/(?P<character>\w),/', 'a,b,c,d', $matches);
    echo (isset($matches['character']) ? 'yes ' : 'no ');

The expected output is "yes", but it returns "no". The reason is that  
the index "character" in the
array returned by preg_match seemst to be a binary string, not  
unicode as I expected:

    preg_match('/(?P<character>\w),/', 'a,b,c,d', $matches);
    echo (isset($matches[(binary)'character']) ? 'yes ' : 'no ');
    echo (isset($matches[(unicode)'character']) ? 'yes ' : 'no ');

This code outputs "yes no".

A bit worse even:

preg_match('/(?P<character>\w),/', 'a,b,c,d', $matches);
echo (isset($matches[(unicode)'character']) ? 'yes ' : 'no ');
extract($matches);
echo(gettype($character) . ' ');
$matches = compact($character);
echo (isset($matches[(unicode)'character']) ? 'yes ' : 'no ');

The output is "no unicode no".

Sorry if I understood something wrong here, but to me it looks like a  
little inconsistency.

Cheers,
Robert