Everything

php

reference issue

php.internals

Lukas Smith

21 years ago

Hi,

I have encountered yet another reference issue. I tested this on PHP
4.3.10 and 4.3.11RC1 on both windows and linux.

Here is a script that reproduces the issue:
http://www.backendmedia.com/PHP/reference_bug.phps

In the comments at the top you can also see my "hack" which fixes the
problem for me.

I have also tried Derick's patch on the 4.3.11RC1 install on linux, but
this didnt change the situation at all.

regards,
Lukas

Moriyoshi Koizumi

21 years ago

That's expected behaviour. Have a look at http://bugs.php.net/20993 .

The problem can also be prevented by the following hack:

// modifying and assining branch causes unexpexted results to
$this->tree
$metatree[$object_id] = (array)unserialize(serialize($branch));

Regards,
Moriyoshi


On 2005/03/17, at 21:04, Lukas Smith wrote:

Derick Rethans

21 years ago

On Thu, 17 Mar 2005, Moriyoshi Koizumi wrote:

> That's expected behaviour. Have a look at http://bugs.php.net/20993 .

Sorry, but this is not expected - it's a bug.

Derick

--
Derick Rethans
http://derickrethans.nl | http://ez.no | http://xdebug.org

Lukas Smith

21 years ago

Derick Rethans wrote:

> On Thu, 17 Mar 2005, Moriyoshi Koizumi wrote:
>
>> That's expected behaviour. Have a look at http://bugs.php.net/20993 .
>
>
> Sorry, but this is not expected - it's a bug.

I would also like to disagree. I think this is a huge inconsistency that
should be addressed even if that costs performance.

If however the core php developers do think this behavior should stick,
then there should be a function that allows people to clean these
references without such hacks (even if internally these hacks are still
going on).

regards,
Lukas

Moriyoshi Koizumi

21 years ago

On 2005/03/17, at 21:52, Lukas Smith wrote:

> Derick Rethans wrote:
>> On Thu, 17 Mar 2005, Moriyoshi Koizumi wrote:
>>> That's expected behaviour. Have a look at http://bugs.php.net/20993 .
>> Sorry, but this is not expected - it's a bug.
>
> I would also like to disagree. I think this is a huge inconsistency
> that should be addressed even if that costs performance.

I agree that it's not intuitive at all, though this issue was discussed
ending up with the following words.

http://marc.theaimsgroup.com/?l=php-dev&m=104018832405963&w=2

> If however the core php developers do think this behavior should
> stick, then there should be a function that allows people to clean
> these references without such hacks (even if internally these hacks
> are still going on).

I think it's possible to add such a function to the 5.x branches.

Moriyoshi