Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:85982
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.213.180 as permitted sender)
MIME-Version: 1.0
Sender: morrison.levi@gmail.com
In-Reply-To: <1430169137.2886384.259330117.00E1734F@webmail.messagingengine.com>
References: <1430169137.2886384.259330117.00E1734F@webmail.messagingengine.com>
Date: Mon, 27 Apr 2015 22:14:46 -0600
Message-ID: <CAFMT4NqKtQmsNEqE+Mi0W8vb35r4TeAVn8xhL2PPr6pS86_F3g@mail.gmail.com>
To: php@lvl.fastmail.com
Cc: internals <internals@lists.php.net>
Content-Type: text/plain; charset=UTF-8
Subject: Re: [PHP-DEV] Resolution of static variables in the function scope
 not using forwarded calling information?
From: levim@php.net (Levi Morrison)

On Mon, Apr 27, 2015 at 3:12 PM,  <php@lvl.fastmail.com> wrote:
> Hello all,
>
> I ran into some very strange behavior while experimenting with static
> function variables. It looks like a bug to me, but I couldn't find any
> previous reports or discussions about the subject, so I wanted to gather
> your input before submitting it as such.
>
> Static variables in the function scope appear to be resolved using late
> static binding: the content of the variable in that function is shared
> amongst all instances of a class, though each subclass gets its own
> separate variable space.
>
> This is pretty logical and works fine. As an example, let's create
> classes Base => Animal => {Cat,Dog}. Base has a function that describes
> the class, and caches the result in a static function variable
> $description. Calling describe() on the three different classes
> correctly gives us three different results: http://3v4l.org/Qldve
>
> However, once you override the describe function in Animal and call
> parent::describe() from within, it looks like all calling information is
> lost: the resolution of $description in Base's function scope now always
> points to Animal's variable space. For example: http://3v4l.org/qWrvf
>
> Overriding the function in one of the childmost classes like Dog gives
> an even stranger result: $description is now correctly a separate value
> in Cat vs Dog, but calling Animal::describe() now uses the value left by
> its child class Cat?! See: http://3v4l.org/61P72

This last is the only one that seems to be a bug to me; I think if you
step through the others you'll see why they are fine. Maybe the last
isn't a bug either; I'll try to think about this more but it does
appear to be a bug.

> Any thoughts?
>
> Thanks,
> Laszlo