PHP程序运行流程：词法分析（Lexing，Tokenizing，Scanning）

在不开启 Opcache 的情况下，PHP解释器在解释PHP脚本的时候，首先会经过词法分析(Lexing)，而词法分析的具体实现就是将PHP代码转换成 Tokens，此过程成为 Lexing / Tokenizing / Scanning 。

那么 Tokens 是啥样的呢，Lex就是一个词法分析的依据表。 Zend/zend_language_scanner.c会根据Zend/zend_language_scanner.l (Lex文件),来输入的 PHP代码进行词法分析，从而得到一个一个的“词”，PHP4.2开始提供了一个函数叫token_get_all这个函数就可以将一段PHP代码 Scanning成Tokens；

token_get_all ( string $source ) : array

token_get_all() 解析提供的 source 源码字符，然后使用 Zend 引擎的语法分析器获取源码中的 PHP 语言的解析器代号

解析器代号列表见：https://www.php.net/manual/zh/tokens.php

token 代号是在zend引擎中定义的，php的词法分析都是基于开源的 lex 工具。

或者使用 翻译获取这个代号的字符串表示。

示例：

<?php
$str = 
'<?php 
$a = "a";
$b = "b";
$c = "c";

echo $a, $b, $c;'
;

$tokens = token_get_all($str);
print_r($tokens);

打印结果

Array
(
    [0] => Array
        (
            [0] => 379
            [1] => <?php
            [2] => 1
        )

    [1] => Array
        (
            [0] => 382
            [1] =>

            [2] => 1
        )

    [2] => Array
        (
            [0] => 320
            [1] => $a
            [2] => 2
        )

    [3] => Array
        (
            [0] => 382
            [1] =>
            [2] => 2
        )

    [4] => =
    [5] => Array
        (
            [0] => 382
            [1] =>
            [2] => 2
        )

    [6] => Array
        (
            [0] => 323
            [1] => "a"
            [2] => 2
        )

    [7] => ;
    [8] => Array
        (
            [0] => 382
            [1] =>

            [2] => 2
        )

    [9] => Array
        (
            [0] => 320
            [1] => $b
            [2] => 3
        )

    [10] => Array
        (
            [0] => 382
            [1] =>
            [2] => 3
        )

    [11] => =
    [12] => Array
        (
            [0] => 382
            [1] =>
            [2] => 3
        )

    [13] => Array
        (
            [0] => 323
            [1] => "b"
            [2] => 3
        )

    [14] => ;
    [15] => Array
        (
            [0] => 382
            [1] =>

            [2] => 3
        )

    [16] => Array
        (
            [0] => 320
            [1] => $c
            [2] => 4
        )

    [17] => Array
        (
            [0] => 382
            [1] =>
            [2] => 4
        )

    [18] => =
    [19] => Array
        (
            [0] => 382
            [1] =>
            [2] => 4
        )

    [20] => Array
        (
            [0] => 323
            [1] => "c"
            [2] => 4
        )

    [21] => ;
    [22] => Array
        (
            [0] => 382
            [1] =>


            [2] => 4
        )

    [23] => Array
        (
            [0] => 328
            [1] => echo
            [2] => 6
        )

    [24] => Array
        (
            [0] => 382
            [1] =>
            [2] => 6
        )

    [25] => Array
        (
            [0] => 320
            [1] => $a
            [2] => 6
        )

    [26] => ,
    [27] => Array
        (
            [0] => 382
            [1] =>
            [2] => 6
        )

    [28] => Array
        (
            [0] => 320
            [1] => $b
            [2] => 6
        )

    [29] => ,
    [30] => Array
        (
            [0] => 382
            [1] =>
            [2] => 6
        )

    [31] => Array
        (
            [0] => 320
            [1] => $c
            [2] => 6
        )

    [32] => ;
)

表示这段代码可以解析出 33 个元素，每一个元素有三个属性，第一个为解析器代号，第二个为字面值，第三个为行号。

比如我们看到的下标为 31 的数组，其代号为 320， 使用 token_name(320)得到其代号名称为T_VARIABLE，去代号列表查看：