深入理解PHP原理之foreach


·        作者Laruence(   )

·        本文地址http://www.laruence.com/2008/11/20/630.html

·        转载请注明出处

foreachPHP中很常用的一个用作数组循环的控制语句。
因为它的方便和易用,自然也就在后端隐藏着很复杂的具体实现方式(对用户透明)
今天,我们就来一起分析分析,foreach是如何实现数组(对象)的遍历的。
本节内容涉及到较多编译原理(lexand yacc)的知识,所以如果您觉得看不太懂,可以先找相关的资料看看。

我们知道PHP是一个脚本语言,也就是说,用户编写的PHP代码最终都是会被PHP解释器解释执行,
特别的,对于PHP来说,所有的用户编写的PHP代码,都会被翻译成PHP的虚拟机ZE的虚拟指令(OPCODES)来执行(参看:深入理解PHP原理之Opcodes).

不论细节的话,就是说,我们所编写的任何PHP脚本,都会最终被翻译成一条条的指令,从而根据指令,由相应的C编写的函数来执行。

那么foreach会被翻译成什么样子呢?

1.       foreach($arras$key => $val){

2.            echo$key . '=>' . $val . "\n";

3.       }

在词法分析阶段,foreach会被识别为一个TOKENT_FOREACH
在语法分析阶段,会被规则:

1.         unticked_statement:  //没有被绑定ticks的语句

2.            //有省略

3.           |   T_FOREACH '(' variable T_AS

4.               { zend_do_foreach_begin(&$1, &$2, &$3, &$4,1 TSRMLS_CC); }

5.               foreach_variableforeach_optional_arg ')' { zend_do_foreach_cont(&$1, &$2, &$4, &$6, &$7 TSRMLS_CC); }

6.               foreach_statement{ zend_do_foreach_end(&$1, &$4 TSRMLS_CC); }

7.           |   T_FOREACH '(' expr_without_variable T_AS

8.               { zend_do_foreach_begin(&$1, &$2, &$3, &$4,0 TSRMLS_CC); }

9.               variableforeach_optional_arg ')' { zend_check_writable_variable(&$6);zend_do_foreach_cont(&$1, &$2, &$4, &$6, &$7 TSRMLS_CC); }

10.            foreach_statement{ zend_do_foreach_end(&$1, &$4 TSRMLS_CC); }

11.         //有省略

12.    ;

仔细分析这段语法规则,我们可以发现,对于:
foreach($arr as $key => $val){
echo $key . ‘=>’ . $val .”\n”;
}

会被分析为:

1.            T_FOREACH '(' variable T_AS   { zend_do_foreach_begin('foreach','(',$arr,'as',1 TSRMLS_CC); }

2.           foreach_variable  foreach_optional_arg(T_DOUBLE_ARROW foreach_variable)  ')' { zend_do_foreach_cont('foreach','(','as',$key,$val TSRMLS_CC); }

3.           foreach_satement {zend_do_foreach_end('foreach','as');}

然后,让我们来看看foreach_statement
它其实就是一个代码块,体现了我们的 echo $key . ‘=>’ . $val .”\n”;
T_ECHO expr;

显然,实现foreach的核心就是如下3个函数:
zend_do_foreach_begin
zend_do_foreach_cont
zend_do_foreach_end

其中,zend_do_foreach_begin (代码太长,直接写伪码) 主要做了:
1.
记录当前的opline行数(为以后跳转而记录)
2.
对数组进行RESET(讲内部指针指向第一个元素)
3.
获取临时变量$val
4.
设置获取变量的OPCODE FE_FETCH,结果存第3步的临时变量
4.
记录获取变量的OPCODES的行数

而对于 zend_do_foreach_cont来说:
1.
根据foreach_variableu.EA.type来判断是否引用
2.
根据是否引用来调整zend_do_foreach_begin中生成的FE_FETCH方式
3.
根据zend_do_foreach_begin中记录的取变量的OPCODES的行数,来初始化循环(主要处理在循环内部的循环:do_begin_loop)

最后zend_do_foreach_end:
1.
根据zend_do_foreach_begin中记录的行数信息,设置ZEND_JMPOPCODES
2.
根据当前行数,设置循环体下一条opline, 用以跳出循环
3.
结束循环(处理循环内循环:do_end_loop)
4.
清理临时变量

当然,zend_do_foreach_contzend_do_foreach_end之间会在语法分析阶段被填充foreach_satement的语句代码。

这样,就实现了foreachOPCODES line
比如对于我们开头的实例代码,最终生成的OPCODES是:

1.       filename:       /home/huixinchen/foreach.php

2.       function name:  (null)

3.       number of ops:  17

4.       compiled vars:  !0 = $arr, !1 = $key, !2 = $val

5.       line     # op                          fetch          ext  return operands

6.       -------------------------------------------------------------------------------

7.          2     0  SEND_VAL                                                 1

8.                1  SEND_VAL                                                100

9.                2  DO_FCALL                                      2          'range'

10.             3  ASSIGN                                                  !0,$0

11.       3     4  FE_RESET                                         $2      !0, ->14

12.             5  FE_FETCH                                         $3      $2, ->14

13.             6  ZEND_OP_DATA                                     ~5

14.             7  ASSIGN                                                   !2,$3

15.             8  ASSIGN                                                  !1, ~5

16.       4     9  CONCAT                                           ~7      !1,'-'

17.            10  CONCAT                                           ~8      ~7, !2

18.            11  CONCAT                                           ~9      ~8,'%0A'

19.            12  ECHO                                                    ~9

20.       5    13  JMP                                                     ->5

21.            14  SWITCH_FREE                                              $2

22.       7    15  RETURN                                                  1

23.            16* ZEND_HANDLE_EXCEPTION

我们注意到FE_FETCHop2的操作数是14,也就是JMP后一条opline,也就是说,在获取完最后一个数组元素以后,FE_FETCH失败的情况下,会跳到第14opline,从而实现了循环的结束。
15oplineop1的操作数是指向了FE_FETCH,也就是无条件跳转到第5opline,从而实现了循环。

附录:

1.       void zend_do_foreach_begin(znode *foreach_token, znode *open_brackets_token, znode *array, znode *as_token, int variable TSRMLS_DC)

2.       {

3.           zend_op *opline;

4.           zend_bool is_variable;

5.           zend_bool push_container = 0;

6.           zend_op dummy_opline;

7.        

8.           if (variable) {

9.                 //是否是匿名数组

10.            if (zend_is_function_or_method_call(array)) {

11.                   //是否是函数返回值

12.                is_variable= 0;

13.            } else {

14.                is_variable= 1;

15.            }

16.            /* 使用括号记录FE_RESETopline行数 */

17.            open_brackets_token->u.opline_num = get_next_op_number(CG(active_op_array));

18.            zend_do_end_variable_parse(BP_VAR_W, 0 TSRMLS_CC); //获取数组/对象和zend_do_begin_variable_parse对应

19.            if (CG(active_op_array)->last > 0 &&

20.                CG(active_op_array)->opcodes[CG(active_op_array)->last-1].opcode == ZEND_FETCH_OBJ_W) {

21.                /* Only lock the container if we are fetchingfrom a real container and not $this */

22.                if (CG(active_op_array)->opcodes[CG(active_op_array)->last-1].op1.op_type == IS_VAR) {

23.                    CG(active_op_array)->opcodes[CG(active_op_array)->last-1].extended_value |= ZEND_FETCH_ADD_LOCK;

24.                    push_container= 1;

25.                }

26.            }

27.        } else {

28.            is_variable= 0;

29.            open_brackets_token->u.opline_num = get_next_op_number(CG(active_op_array));

30.        }

31.     

32.        foreach_token->u.opline_num = get_next_op_number(CG(active_op_array)); //记录数组Reset Opline number

33.     

34.        opline = get_next_op(CG(active_op_array) TSRMLS_CC); //生成Reset数组Opcode

35.     

36.        opline->opcode = ZEND_FE_RESET;

37.        opline->result.op_type = IS_VAR;

38.        opline->result.u.var = get_temporary_variable(CG(active_op_array));

39.        opline->op1 = *array;

40.        SET_UNUSED(opline->op2);

41.        opline->extended_value = is_variable ? ZEND_FE_RESET_VARIABLE: 0;

42.     

43.        dummy_opline.result = opline->result;

44.        if (push_container) {

45.            dummy_opline.op1 = CG(active_op_array)->opcodes[CG(active_op_array)->last-2].op1;

46.        } else {

47.            znode tmp;

48.     

49.            tmp.op_type = IS_UNUSED;

50.            dummy_opline.op1 = tmp;

51.        }

52.        zend_stack_push(&CG(foreach_copy_stack), (void *) &dummy_opline, sizeof(zend_op));

53.     

54.        as_token->u.opline_num = get_next_op_number(CG(active_op_array)); //记录循环起始点

55.     

56.        opline = get_next_op(CG(active_op_array) TSRMLS_CC);

57.        opline->opcode = ZEND_FE_FETCH;

58.        opline->result.op_type = IS_VAR;

59.        opline->result.u.var = get_temporary_variable(CG(active_op_array));

60.        opline->op1 = dummy_opline.result;    //被操作数组

61.        opline->extended_value = 0;

62.        SET_UNUSED(opline->op2);

63.     

64.        opline = get_next_op(CG(active_op_array) TSRMLS_CC);

65.        opline->opcode = ZEND_OP_DATA; //当使用key的时候附属操作数,foreach中不包含key时忽略

66.        SET_UNUSED(opline->op1);

67.        SET_UNUSED(opline->op2);

68.        SET_UNUSED(opline->result);

69.    }

1.       void zend_do_foreach_cont(znode *foreach_token, const znode *open_brackets_token, const znode *as_token, znode *value, znode *key TSRMLS_DC)

2.       {

3.           zend_op *opline;

4.           znode dummy, value_node;

5.           zend_bool assign_by_ref=0;

6.        

7.           opline = &CG(active_op_array)->opcodes[as_token->u.opline_num]; //获取FE_FETCH Opline

8.           if (key->op_type != IS_UNUSED) {

9.               znode *tmp;//交换keyval

10.     

11.            tmp = key;

12.            key = value;

13.            value = tmp;

14.     

15.            opline->extended_value |=ZEND_FE_FETCH_WITH_KEY; //表明需要同时获取keyval

16.        }

17.     

18.        if ((key->op_type != IS_UNUSED) && (key->u.EA.type & ZEND_PARSED_REFERENCE_VARIABLE)){

19.              //key不能以引用方式获取

20.            zend_error(E_COMPILE_ERROR, "Key element cannot be a reference");

21.        }

22.     

23.        if (value->u.EA.type &ZEND_PARSED_REFERENCE_VARIABLE) {

24.              //以引用方式获取值

25.            assign_by_ref= 1;

26.            if (!(opline-1)->extended_value) {

27.                   //根据FE_FETCH的上一条Opline也就是获取数组的扩展值来判断数组是否是匿名数组

28.                zend_error(E_COMPILE_ERROR, "Cannot create references to elements ofa temporary array expression");

29.            }

30.     

31.            opline->extended_value |= ZEND_FE_FETCH_BYREF; //指明按引用取

32.            CG(active_op_array)->opcodes[foreach_token->u.opline_num].extended_value |= ZEND_FE_RESET_REFERENCE; //重置原数组

33.        } else {

34.            zend_op *foreach_copy;

35.            zend_op *fetch = &CG(active_op_array)->opcodes[foreach_token->u.opline_num];

36.            zend_op *end = &CG(active_op_array)->opcodes[open_brackets_token->u.opline_num];

37.     

38.            /* Change "write context" into"read context" */

39.            fetch->extended_value = 0;  /* reset ZEND_FE_RESET_VARIABLE */

40.            while (fetch != end) {

41.                --fetch;

42.                if (fetch->opcode == ZEND_FETCH_DIM_W && fetch->op2.op_type == IS_UNUSED) {

43.                    zend_error(E_COMPILE_ERROR, "Cannot use [] for reading");

44.                }

45.                fetch->opcode -= 3;/* FETCH_W -> FETCH_R */

46.            }

47.     

48.            /* prevent double SWITCH_FREE */

49.            zend_stack_top(&CG(foreach_copy_stack), (void **) &foreach_copy);

50.            foreach_copy->op1.op_type = IS_UNUSED;

51.        }

52.     

53.        value_node = opline->result;

54.     

55.        if (assign_by_ref) {

56.            zend_do_end_variable_parse(value, BP_VAR_W, 0 TSRMLS_CC); //获取值(引用)

57.            zend_do_assign_ref(NULL, value, &value_node TSRMLS_CC);//指明value nodetypeIS_VAR

58.        } else {

59.            zend_do_assign(&dummy, value, &value_node TSRMLS_CC); //获取copy

60.            zend_do_free(&dummy TSRMLS_CC);

61.        }

62.     

63.        if (key->op_type != IS_UNUSED) {

64.            znodekey_node;

65.     

66.            opline = &CG(active_op_array)->opcodes[as_token->u.opline_num+1];

67.            opline->result.op_type = IS_TMP_VAR;

68.            opline->result.u.EA.type = 0;

69.            opline->result.u.opline_num = get_temporary_variable(CG(active_op_array));

70.            key_node = opline->result;

71.     

72.            zend_do_assign(&dummy, key, &key_node TSRMLS_CC);

73.            zend_do_free(&dummy TSRMLS_CC);

74.        }

75.     

76.        do_begin_loop(TSRMLS_C);

77.        INC_BPC(CG(active_op_array));

78.    }

1.       void zend_do_foreach_end(znode *foreach_token, znode *as_token TSRMLS_DC)

2.       {

3.           zend_op *container_ptr;

4.           zend_op *opline = get_next_op(CG(active_op_array) TSRMLS_CC); //生成JMP opcode

5.        

6.           opline->opcode = ZEND_JMP;

7.           opline->op1.u.opline_num = as_token->u.opline_num; //设置JMPFE_FETCH opline

8.           SET_UNUSED(opline->op1);

9.           SET_UNUSED(opline->op2);

10.     

11.        CG(active_op_array)->opcodes[foreach_token->u.opline_num].op2.u.opline_num = get_next_op_number(CG(active_op_array)); //设置跳出循环的opline

12.        CG(active_op_array)->opcodes[as_token->u.opline_num].op2.u.opline_num = get_next_op_number(CG(active_op_array)); //同上

13.     

14.        do_end_loop(as_token->u.opline_num, 1 TSRMLS_CC); //为循环嵌套而设置

15.     

16.        zend_stack_top(&CG(foreach_copy_stack), (void **) &container_ptr);

17.        generate_free_foreach_copy(container_ptr TSRMLS_CC);

18.        zend_stack_del_top(&CG(foreach_copy_stack));

19.     

20.        DEC_BPC(CG(active_op_array)); //PHP interactive模式而设置

21.    }

 


  • 1
    点赞
  • 3
    收藏
    觉得还不错? 一键收藏
  • 0
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值