Edit report at http://bugs.php.net/bug.php?id=52741&edit=1
ID: 52741
User updated by: flyguy dot by at gmail dot com
Reported by: flyguy dot by at gmail dot com
Summary: OOP late static bindings bug
Status: Assigned
Type: Bug
Package: Scripting Engine problem
Operating System: *
PHP Version: 5.3.3
Assigned To: colder
Block user comment: N
New Comment:
Sory, my English is very bad :))
Previous Comments:
------------------------------------------------------------------------
[2010-09-13 22:46:23] flyguy dot by at gmail dot com
class my_parent {
public static $blabla='asd';
public static function set_blabla($val) {
static::$blabla=$val;
}
}
class my_child extends my_parent {
}
my_child::set_blabla('qwerty');
var_dump(my_child::$blabla);
var_dump(my_parent::$blabla);
--------
prints:
string 'qwerty' (length=6)
string 'qwerty' (length=6)
----
This is normal code too ?
Why changed extended propety on parent ?
By your logic inherited properties are not inherited properties.
Simply class-parent becomes part of the subclass.
------------------------------------------------------------------------
[2010-09-13 21:42:17] col...@php.net
parent:: forwards LSB information
In the initial case, it makes it perfectly normal for get_called_class
to return
test2, this is by design, if you don't want to pass the LSB info,
reference your
class by name, without keyword (i.e. test1::testing_method()).
---
Johannes:
on test3::testingmethod():
- parent::testing_method() will call parent (here test2), passing LSB
info
(test3)
- test2::testing_method falls back to test1::method, passing LSB info
test1::testing_method's get_called_class will be test3.
To me it sounds perfectly normal?
---
For the last comment:
static::$field is resolved to once my_child::$field and once
my_parent::$field,
my_child::$field falls back to my_parent::$field, and updates it.
=> normal again
Am I missing something here?
------------------------------------------------------------------------
[2010-09-13 00:13:13] flyguy dot by at gmail dot com
Ok. This situation is analogous to the one hand, but why then:
class my_parent {
public static $field;
public static function field_setup() {
static::$field='asd';
}
}
class my_child extends my_parent {
}
my_child::field_setup();
var_dump(my_child::$field);
var_dump(my_parent::$field);
----------
prints:
string 'asd' (length=3)
string 'asd' (length=3)
------------------------------------------------------------------------
[2010-08-30 18:35:08] johan...@php.net
Etienne, you are the LSB expert - what's your take on this? - In the
given example I can see the reporter's point. (While there is a "extends
test1" missing)
I wonder about this:
class test1 {
public static function testing_method() {
var_dump(get_called_class());
}
}
class test2 extends test1 {
}
class test3 extends test2 {
public static function testing_method() {
parent::testing_method();
}
}
Here parent might refer to test2, test2 inherits testing_method() from
test1 so the called class might be test2 ... some might argue that test1
is correct.
I think the most simple thing is to keep the current behavior and define
parent not to change the lsb scope.
------------------------------------------------------------------------
[2010-08-30 17:33:48] flyguy dot by at gmail dot com
Description:
------------
This is code prints "test2". Why ?
Test script:
---------------
class test1 {
public static function testing_method() {
var_dump(get_called_class());
}
}
class test2 {
public static function testing_method() {
parent::testing_method();
}
}
------------------------------------------------------------------------
--
Edit this bug report at http://bugs.php.net/bug.php?id=52741&edit=1
