A brief explanation of PHP garbage collection mechanism_PHP tutorial

Although I am also a learner of PHP, I really didn’t know much about the garbage collection process inside PHP before. We just used unset, null, mysql_close, __destruct and other functions in our code to release objects to prevent memory overflow, so I found the following instructions under GG on the Internet and made a record: "PHP can automatically manage memory and clear objects that are no longer needed. PHP uses a simple garbage collection mechanism called reference counting. Every time Each object contains a reference counter, and each reference is connected to the object. When the reference leaves the living space or is set to NULL, the counter is decremented by 1. When the reference counter of an object is zero, PHP knows that you will. This object is no longer needed and the memory space it occupies is released.”

众所周知, PHP 引擎本身是用 C 写的,提到 C 不能不提的就是 GC(垃圾回收).通过 PHP 手册 我们了解到, PHP 引擎会自动进行 GC 动作.那么我们不禁要问,到底它是怎么回收的, & 引用操作是不是指针, unset() 了一个变量时它是不是真的被回收了呢?这些看似手册有提及的问题,如果仔细分析会发现,远没有那么简单泛泛.也许有人会跳出来说:看 PHP 源码不就知道了.是的,等你通读了 PHP 源码后这个问题肯定不在话下了,然本篇要仅从 PHP 本身来分析这些看似平常却被忽视的小细节,当然了,其中难免水平所限,有所疏漏,热烈欢迎广大 phper 来共同讨论.

首先咱先看到例子,最简单不过的执行流程了:
Example 1: gc.php
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

echo $b ." ";
?>

不用说 % php -f gc.php 输出结果非常明了:
hy0kl% php -f gc.php
I am test.

好,下一个:
Example 2:
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

$b = 'I will change?';

echo $a ." ";
echo $b ." ";
?>
执行结果依然很明显:
hy0kl% php -f gc.php
I will change?
I will change?

君请看:
Example 3:
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

unset($a);

echo $a ." ";
echo $b ." ";
?>
是不是得想一下下呢?
hy0kl% php -f gc.php
Notice: Undefined variable: a in /usr/local/www/apache22/data/test/gc.php on line 8
I am test.
有点犯迷糊了吗?

君再看:
Example 4:
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

unset($b);

echo $a ." ";
echo $b ." ";
?>
其实如果 Example 3 理解了,这个与之异曲同工.
hy0kl% php -f gc.php
I am test.
Notice: Undefined variable: b in /usr/local/www/apache22/data/test/gc.php on line 9

君且看:
Example 5:
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

$a = null;

echo '$a = '. $a ." ";
echo '$b = '. $b ." ";
?>
猛的第一感觉是什么样的?
hy0kl% php -f gc.php
$a =
$b =
没错,这就是输出结果,对 PHP GC 已有深入理解的 phper 不会觉得有什么奇怪,说实话,当我第一次运行这段代码时很意外,却让我对 PHP GC 有更深刻的理解了.那么下面与之同工的例子自然好理解了.

Example 6:
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

$b = null;

echo '$a = '. $a ." ";
echo '$b = '. $b ." ";
?>

OK,如果上面的例子的结果对看官来说无任何细节可言,那您可关闭本窗口了,欢迎有空再来!

下面我们来详细分析 GC 与引用.
1. 所有例子中,创建了一个变量,这个过程通俗一点讲:是在内存中开辟了一块空间,在里面存放了一个字符串 I am test. . PHP 内部有个符号表,用来记录各块内存引用计数,那么此时会将这块内存的引用计数 加 1,并且用一个名为 $a 的标签(变量)指向这块内存,方便依标签名来操作内存.

2. 对变量 $a 进行 & 操作,我的理解是找到 $a 所指向的内存,并为 $b 建立同样的一引用指向,并将存放字符串 I am test. 的内存块在符号表中引用计数 加 1.换言之,我们的脚本执行到这一行的时候,存放字符串 I am test. 的那块内存被引用了两次.这里要强调的是, & 操作是建立了引用指向,而不是指针, PHP 没有指针的概念!同时有人提出说类似于 UNIX 的文件软链接.可以在一定程度上这么理解: 存放字符 I am test. 的那块内存是我们的一个真实的文件,而变量 $a 与 $b 是针对真实文件建立的软链接,但它们指向的是同一个真实文件. So, 我们看到,在 Example 2 中给 $b 赋值的同时, $a 的值也跟着变化了.与通过某一软链操作了文件类似.

3. 在 Example 3 与 4 中,进行了 unset() 操作.根据实际的执行结果,可以看出: unset() 只是断开这个变量对它原先指向的内存的引用,使变量本身成为没有定义过空引用,所在调用时发出了 Notice ,并且使那块内存在符号表中引用计数 减 1,并没有影响到其他指向这块内存的变量.换言之,只有当一块内存在符号表中的引用计数为 0 时, PHP 引擎才会将这块内存回收.
PHP 手册
4.0.0 unset() became an expression. (In PHP 3, unset() would always return 1).
这意味着什么?
看看下面的代码与其结果:
error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

unset($a);
unset($a);
unset($a);

echo '$a = '. $a ." ";
echo '$b = '. $b ." ";
?>
hy0kl% php -f gc.php

Notice: Undefined variable: a in /usr/local/www/ apache22/data/test/gc.php on line 10
$a =
$b = I am test.
The first unset() operation has disconnected the pointer, so subsequent operations cannot It will affect the reference count of any memory in the symbol table.

4. Through Example 5 & 6, it can be clearly concluded that the assignment null operation is quite violent, it will directly point to the variable If the reference count of the memory in the symbol is set to 0, then this memory will naturally be reclaimed by the engine. It is unknown when it will be used again. It may be used to store other information immediately, or it may never be used again. However, In any case, all the original variables pointing to that memory will no longer be able to operate on the reclaimed memory, and any variable trying to call it will return null.

error_reporting(E_ALL);
$a = 'I am test.';
$b = & $a;

$b = null;

echo '$a = '. $a . " ";
echo '$b = '. $b ." ";

if (null === $a)
{
echo '$a is null.';
} else
{
echo 'The type of $a is unknown.';
}
?>
hy0kl% php -f gc.php
$a =
$b =
$a is null.

To sum up, it fully explains why when we look at the source code of open source products, we often see some relatively large temporary variables, or Reused information that is no longer called after use will be concentrated or displayed as null. It is equivalent to directly killing the real file in UNIX, and all soft links pointing to it will naturally become empty links.