lua解析脚本过程中的关键数据结构介绍

在这一篇文章中我先来介绍一下lua解析一个脚本文件时要用到的一些关键的数据结构，为将来的一系列代码分析打下一个良好的基础。在整个过程中，比较重要的几个源码文件分别是：llex.h，lparse.h、lobject.h和lopcode.h。

在llex.h中

typedef struct Token {
   int token;
   SemInfo seminfo;
 } Token;

Token代表了一个词法单元，其中token表示词法类型如TK_NAME、TK_NUMBER等如果不是这些类型则存放则词素的字符表示，例如分析的代码会这么判断词素单元：

switch (ls->t.token) {
    case '(': {
     //...
    }
    case TK_NAME: {
      //...
    }
    default: {
      //...
    }

在Token中SemInfo存放了一些语义相关的一些内容信息

1 typedef union {
2   lua_Number r;
3   TString *ts;
4 } SemInfo;  /* semantics information */

其中当token是数字是内容存放在r中，其他情况存放在ts指向的TString中。

下面是最重要的一个数据结构之一

typedef struct LexState {
  int current;  /* current character (charint) */
  int linenumber;  /* input line counter */
  int lastline;  /* line of last token `consumed' */
  Token t;  /* current token */
  Token lookahead;  /* look ahead token */
  struct FuncState *fs;  /* `FuncState' is private to the parser */
  struct lua_State *L;
  ZIO *z;  /* input stream */
  Mbuffer *buff;  /* buffer for tokens */
  TString *source;  /* current source name */
  char decpoint;  /* locale decimal point */
} LexState;

LexState不仅用于保存当前的词法分析状态信息，而且也保存了整个编译系统的全局状态。current指向了当前字符，t存放了当前的toekn，lookahead存放了向前看的token，由此我认为lua应该是ll(1)的~哈哈（不知道对不对）。fs指向了parser当前解析的函数的一些相关的信息，L指向了当前lua_State结构，z指向输入流，buff指向了token buffer，其他的看注释吧。

下面看看lparse.h文件中的重要结构：

typedef struct expdesc {
  expkind k;
  union {
    struct { int info, aux; } s;
    lua_Number nval;
  } u;
  int t;  /* patch list of `exit when true' */
  int f;  /* patch list of `exit when false' */
} expdesc;

expdesc是存放了表达式的相关描述信息，k是表达式的种类，u在不同的表达式中有不同的含义。

1 typedef struct upvaldesc {
2   lu_byte k;
3   lu_byte info;
4 } upvaldesc;

upvaldesc是存放了upval的相关描述信息。

最后是本文件中最重要的结构：

typedef struct FuncState {
  Proto *f;  /* current function header */
  Table *h;  /* table to find (and reuse) elements in `k' */
  struct FuncState *prev;  /* enclosing function */
  struct LexState *ls;  /* lexical state */
  struct lua_State *L;  /* copy of the Lua state */
  struct BlockCnt *bl;  /* chain of current blocks */
  int pc;  /* next position to code (equivalent to `ncode') */
  int lasttarget;   /* `pc' of last `jump target' */
  int jpc;  /* list of pending jumps to `pc' */
  int freereg;  /* first free register */
  int nk;  /* number of elements in `k' */
  int np;  /* number of elements in `p' */
  short nlocvars;  /* number of elements in `locvars' */
  lu_byte nactvar;  /* number of active local variables */
  upvaldesc upvalues[LUAI_MAXUPVALUES];  /* upvalues */
  unsigned short actvar[LUAI_MAXVARS];  /* declared-variable stack */
} FuncState;

在编译过程中，使用FuncState结构体来保存一个函数编译的状态数据。其中，f指向了本函数的协议描述结构体，prev指向了其父函数的FuncState描述，因为在lua中可以在一个函数中定义另一个函数，因此当parse到一个函数的内部函数的定义时会new一个FuncState来描述内部函数，同时开始parse这个内部函数，将这个FuncState的prev指向其外部函数的FuncState，prev变量用来引用外围函数的FuncState，使当前所有没有分析完成的FuncState形成一个栈结构。bl指向当前parse的block，在一个函数中会有很多block代码，lua会将这些同属于同一个函数的block用链表串联起来。jpc是一个OP_JMP指令的链表，因为lua是一遍过的parse，在开始的时候有一些跳转指令不能决定其跳转位置，因此jpc将这些pending
jmp指令串联起来，在以后能确定的时候回填，freereg为第一个空闲寄存器的下标，upvalues数组保存了当前函数的所有upvalue，nactvar是当前作用域的局部变量数。

在lparse.c中定义了BlockCnt

/*
** nodes for block list (list of active blocks)
*/
typedef struct BlockCnt {
  struct BlockCnt *previous;  /* chain */
  int breaklist;  /* list of jumps out of this loop */
  lu_byte nactvar;  /* # active locals outside the breakable structure */
  lu_byte upval;  /* true if some variable in the block is an upvalue */
  lu_byte isbreakable;  /* true if `block' is a loop */
} BlockCnt;

Lua使用BlockCnt来保存一个block的数据。与FuncState的分析方法类似，BlockCnt使用一个previous变量保存外围block的引用，形成一个栈结构。



下面介绍一些在lobject.h文件里面的数据结构

/*
** Function Prototypes
*/
typedef struct Proto {
  CommonHeader;
  TValue *k;  /* constants used by the function */
  Instruction *code;
  struct Proto **p;  /* functions defined inside the function */
  int *lineinfo;  /* map from opcodes to source lines */
  struct LocVar *locvars;  /* information about local variables */
  TString **upvalues;  /* upvalue names */
  TString  *source;
  int sizeupvalues;
  int sizek;  /* size of `k' */
  int sizecode;
  int sizelineinfo;
  int sizep;  /* size of `p' */
  int sizelocvars;
  int linedefined;
  int lastlinedefined;
  GCObject *gclist;
  lu_byte nups;  /* number of upvalues */
  lu_byte numparams;
  lu_byte is_vararg;
  lu_byte maxstacksize;
} Proto;

结构体Proto是lua函数协议的描述，在lua解析脚本时首先会将main chunk代码包裹为一个函数，用main proto描述，接着将里面定义的内部函数一一用Proto结构体描述，将这些Proto的关系用树来组合起来，例如有lua源码文件如下

a = 1
function f1()
-- ...
end
function f2()
    function f3()
    -- ...
    end
end

则parse完成后会有如图如下关系



在Proto结构体中，k指向一个const变量数组，存放则函数要用到的常量；code指向lua parse过程中生成的本函数的instruction集合；p就是指向本函数内部定义的函数的那些proto；locvars指向本函数局部变量数组；upvalues指向本函数upvalue变量数组；nups为upvalue的数量；numparams为函数参数的数量；is_vararg表示函数是否接收可变参数；maxstacksize为函数stack的max大小。

在编译期间lua使用Proto描述函数的，当lua vm开始运行vm时需要根据Proto生成相应的Closure来执行vm instructions。

1 typedef union Closure {
2   CClosure c;
3   LClosure l;
4 } Closure;

Closure要么代表了c函数，要么为lua函数，在这里我们只看lua函数的LClosure

#define ClosureHeader \
    CommonHeader; lu_byte isC; lu_byte nupvalues; GCObject *gclist; \
    struct Table *env
//... ...
typedef struct LClosure {
  ClosureHeader;
  struct Proto *p;
  UpVal *upvals[1];
} LClosure;

在LClousre中，p就是指向对应函数的Proto结构体啦，upvals顾名思义就是此closure的upvalue数组罗。在ClosureHeader宏中isC表示此closure是否是c函数，nupvalues为upvalue数目，env指向了此closue运行时的函数环境，在lua中可以用stefenv来改变当前函数的环境，就是改变env变量的指向啦。

最后，在文件lopcode.h中定义了lua vm的指令结构



下面是vm指令的一些定义与描述，我在相应vm指令的上方添加了一些注释

typedef enum {
/*----------------------------------------------------------------------
name        args    description
------------------------------------------------------------------------*/
OP_MOVE,/*    A B    R(A) := R(B)                    */
//Constants are usually numbers or strings. Each function has its own constant list, or pool.
OP_LOADK,/*    A Bx    R(A) := Kst(Bx)                    */
OP_LOADBOOL,/*    A B C    R(A) := (Bool)B; if (C) pc++            */
//The optimization rule is  a simple one: If no other instructions have been generated, 
//then a LOADNIL as the first instruction can be optimized away.
OP_LOADNIL,/*    A B    R(A) := ... := R(B) := nil            */

OP_GETUPVAL,/*    A B    R(A) := UpValue[B]                */
OP_GETGLOBAL,/*    A Bx    R(A) := Gbl[Kst(Bx)]                */
OP_GETTABLE,/*    A B C    R(A) := R(B)[RK(C)]                */

OP_SETGLOBAL,/*    A Bx    Gbl[Kst(Bx)] := R(A)                */
OP_SETUPVAL,/*    A B    UpValue[B] := R(A)                */
OP_SETTABLE,/*    A B C    R(A)[RK(B)] := RK(C)                */

OP_NEWTABLE,/*    A B C    R(A) := {} (size = B,C)                */

//This instruction is used for object-oriented programming. It is only generated for method calls that use the colon syntax.
//R(B) is the register holding the reference to the table with the method. 
OP_SELF,/*    A B C    R(A+1) := R(B); R(A) := R(B)[RK(C)]        */

//The optimization rule is simple: If both terms of a subexpression are numbers, 
//the subexpression will be evaluated at compile time.
OP_ADD,/*    A B C    R(A) := RK(B) + RK(C)                */
OP_SUB,/*    A B C    R(A) := RK(B) - RK(C)                */
OP_MUL,/*    A B C    R(A) := RK(B) * RK(C)                */
OP_DIV,/*    A B C    R(A) := RK(B) / RK(C)                */
OP_MOD,/*    A B C    R(A) := RK(B) % RK(C)                */
OP_POW,/*    A B C    R(A) := RK(B) ^ RK(C)                */
OP_UNM,/*    A B    R(A) := -R(B)                    */
OP_NOT,/*    A B    R(A) := not R(B)                */
//Returns the length of the object in R(B)
OP_LEN,/*    A B    R(A) := length of R(B)                */

//Performs concatenation of two or more strings.
//The source registers must be consecutive, and C must always be greater than B. 
OP_CONCAT,/*    A B C    R(A) := R(B).. ... ..R(C)            */

//if sBx is 0, the VM will proceed to the next instruction
OP_JMP,/*    sBx    pc+=sBx                    */

/*If the boolean result is not A, then skip the next instruction. 
Conversely, if the boolean result equals A, continue with the next instruction.*/
OP_EQ,/*    A B C    if ((RK(B) == RK(C)) ~= A) then pc++        */
OP_LT,/*    A B C    if ((RK(B) <  RK(C)) ~= A) then pc++          */
OP_LE,/*    A B C    if ((RK(B) <= RK(C)) ~= A) then pc++          */

OP_TEST,/*    A C    if not (R(A) <=> C) then pc++            */ 
//register R(B) is coerced into a boolean.
OP_TESTSET,/*    A B C    if (R(B) <=> C) then R(A) := R(B) else pc++    */ 

//If B is 0, parameters range from R(A+1) to the top of the stack.If B is 1, the function has no parameters.
//If C is 1, no return results are saved. If C is 0, then multiple return results are saved, depending on the called function
//CALL always updates the top of stack value.
OP_CALL,/*    A B C    R(A), ... ,R(A+C-2) := R(A)(R(A+1), ... ,R(A+B-1)) */
OP_TAILCALL,/*    A B C    return R(A)(R(A+1), ... ,R(A+B-1))        */
//If B is 1, there are no return values. If B is 0, the set of values from R(A) to the top of the stack is returned. 
OP_RETURN,/*    A B    return R(A), ... ,R(A+B-2)    (see note)    */

//FORPREP initializes a numeric for loop, while FORLOOP performs an iteration of a numeric for loop.
OP_FORLOOP,/*    A sBx    R(A)+=R(A+2);
            if R(A) <?= R(A+1) then { pc+=sBx; R(A+3)=R(A) }*/
OP_FORPREP,/*    A sBx    R(A)-=R(A+2); pc+=sBx                */

//Performs an iteration of a generic for loop. 
OP_TFORLOOP,/*    A C    R(A+3), ... ,R(A+2+C) := R(A)(R(A+1), R(A+2)); 
                        if R(A+3) ~= nil then R(A+2)=R(A+3) else pc++    */ 
//This instruction is used to initialize array elements in a table.
//If B is 0, the table is set with a variable number of array elements, from register R(A+1) up to the top of the stack. 
//If C is 0, the next instruction is cast as an integer, and used as the C value.
OP_SETLIST,/*    A B C    R(A)[(C-1)*FPF+i] := R(A+i), 1 <= i <= B    */

/*If a local is used as an upvalue, then the local variable need to be placed somewhere, 
other wise it will go out of scope and disappear when a lexicalblock enclosing the local variable ends. 
CLOSE performs this operation for all affected local variables for do end blocks or loop blocks. 
RETURN also does an implicit CLOSE when a function returns.*/
OP_CLOSE,/*    A     close all variables in the stack up to (>=) R(A)*/
/*Each upvalue corresponds to either a MOVE or a GETUPVAL pseudo-instruction. 
Only the B field on either of these pseudo-instructions are significant.*/
//MOVE pseudo-instructions corresponds to local variable R(B) in the current lexical block.
//GETUPVAL pseudo-instructions corresponds upvalue number B in the current lexical block.
OP_CLOSURE,/*    A Bx    R(A) := closure(KPROTO[Bx], R(A), ... ,R(A+n))    */

//If B is 0, VARARG copies as many values as it can based on the number of parameters passed.
//If a fixed number of values is required, B is a value greater than 1. 
OP_VARARG/*    A B    R(A), R(A+1), ..., R(A+B-1) = vararg        */
} OpCode;