Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:31240
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain lerdorf.com from 204.11.219.139 cause and error)
Message-ID: <46A6205A.6040001@lerdorf.com>
Date: Tue, 24 Jul 2007 08:52:58 -0700
User-Agent: Thunderbird 2.0.0.5 (Macintosh/20070716)
MIME-Version: 1.0
To: Matthew Kavanagh <matthew@teh.ath.cx>
CC: Stanislav Malyshev <stas@zend.com>,
        Andrew Minerd <andrew.minerd@sellingsource.com>,
        "'PHP Internals'" <internals@lists.php.net>
References: <46A31D95.2080407@fischer.name>	<E996AAF4A714DF4D9BD59E9418957A2128CD19@peak8sbs.peak8solutions.local>	<46A4661A.6080202@zend.com>	<200707231140.55535.pstradomski@gmail.com>	<46A4BA4F.6090506@dealnews.com>	<40CA6572-030F-43E2-8706-64D1167C527C@bitxtender.com>	<46A4C090.50402@php.net>	<3d1a63d10707230942y2b2504d1p5f995f01d28a622b@mail.gmail.com>	<46A4DF33.1030403@zend.com> <20070723110303.0c033048@h4x0r.tss> <46A4EEBA.7040308@zend.com> <46A53A50.1020205@lerdorf.com> <46A5FA0B.7040708@teh.ath.cx>
In-Reply-To: <46A5FA0B.7040708@teh.ath.cx>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] Question about Namespace patch
From: rasmus@lerdorf.com (Rasmus Lerdorf)

Matthew Kavanagh wrote:
> Rasmus Lerdorf wrote:
>> Stanislav Malyshev wrote:
>>  
>>> Compile-time resolution means you don't get performance penalty for
>>> namespaces when you are not using it, and have very low costs when you
>>> do use it. Allowing blanket imports means we don't know what "new Foo()"
>>> means until it is executed - meaning we need to make extra lookups on
>>> each class fetch, even if the code doesn't use namespaces at all. That's
>>> only one problem with blanket imports.
>>>     
>>
>> And a fatal one in my book.  For the folks arguing so passionately about
>> this namespace implementation, try to consider the compiler vs. executor
>> issues involved.  We really need to stop the trend of pushing everything
>> down into the executor.
>>
>> -Rasmus
>>
>>   
> The compile time vs. execution time argument may be a good one against
> global import and against other crazy ideas in my head (like allowing
> statements in namespaces).
> 
> But why not allow importing namespaced functions and constants into the
> global namespace?
> 
> If this were added then "blanket import" could be left to the intrepid
> and foolhardy programmer to do manually. It'd just end up being
> boilerplate and if someone really objects to some extra lines they could
> add an explicit fold in their editor to hide it.

As long as it doesn't go into the global namespace directly.  As in:

namespace foo;
class bar1 {
...
}
class bar2 {
...
}

import foo:* as myfoo

gives you:

myfoo:bar1 and myfoo:bar2
and not bar1 and bar2 directly in the global namespace.  The latter is a
performance killer across the board but I see no issue with the former.

-Rasmus