Re: RFC - "class underloading" -or- "ancestor overloading"

From: Date: Fri, 19 Mar 2010 15:19:27 +0000
Subject: Re: RFC - "class underloading" -or- "ancestor overloading"
References: 1 2  Groups: php.internals 
Request: Send a blank email to internals+get-47428@lists.php.net to get a copy of this message
Hi,

The feature specified by Chris remind me the "metaclass" extension system of
Groovy language (I am not a Groovy Expert, but properties seems similar).

+1 for this kind of feature at a language level, IMHO it will really help
for Aspect Oriented Programming, Legacy Code Refactoring, Unit Testing
(Mocking mecanism).


> Gwynne already referred to Runkit, which gives you all the tools you need.
>

IMHO Language-level features are much more appreciated in facts, and
Professionals pratices-required features (AOP, unit-test...) should be
language-level (in order for pratices to become more popular)

However, since monkey-patching isn't something generally appreciate, I
> wouldn't go with promoting it to the language level. Instead, some Runkit
> functionality could be OO-ified and integrated into the Reflection API
> (somewhere along the lines of [2] perhaps).
>

+1


> A technical issue with your proposal you do not address are conflicts.
> What happens when two classes overload the same class?
>

Why "including/parsing" order should not be taken into account ?

Beside that, the best thing you could to for the moment is to convince the
> Zend framework guys to read up about Open Implementations[3].
>

Thanks for the link(s) !

--
Olivier Hoareau

PHP Best Pratices Evanglist
Bordeaux, France


Thread (4 messages)

« previous php.internals (#47428) next »