Tiny语言编译器开发之语法分析（利…

采用自底向上的LR分析方法，Yacc输入文件yyparser.y如下：

%{
#define YYPARSER

#include "globals.h"
#include "util.h"
#include "scan.h"
#include "parse.h"

#define YYSTYPE TreeNode *
static char * savedName;
static int savedLineNo; 
static TreeNode * savedTree;

%}

%token IF THEN ELSE END REPEAT UNTIL READ WRITE
%token ID NUM
%token ASSIGN EQ LT PLUS MINUS TIMES OVER LPAREN RPAREN SEMI
%token ERROR

%%

program     : stmt_seq
                 { savedTree = $1;}
            ;
stmt_seq    : stmt_seq SEMI stmt
                 { YYSTYPE t = $1;
                   if (t != NULL)
                   { while (t->sibling != NULL)
                        t = t->sibling;
                     t->sibling = $3;
                     $$ = $1; }
                     else $$ = $3;
                 }
            | stmt  { $$ = $1; }
            ;
stmt        : if_stmt { $$ = $1; }
            | repeat_stmt { $$ = $1; }
            | assign_stmt { $$ = $1; }
            | read_stmt { $$ = $1; }
            | write_stmt { $$ = $1; }
            ;
if_stmt     : IF exp THEN stmt_seq END
                 { $$ = newStmtNode(IfK);
                   $$->child[0] = $2;
                   $$->child[1] = $4;
                 }
            | IF exp THEN stmt_seq ELSE stmt_seq END
                 { $$ = newStmtNode(IfK);
                   $$->child[0] = $2;
                   $$->child[1] = $4;
                   $$->child[2] = $6;
                 }
            ;
repeat_stmt : REPEAT stmt_seq UNTIL exp
                 { $$ = newStmtNode(RepeatK);
                   $$->child[0] = $2;
                   $$->child[1] = $4;
                 }
            ;
assign_stmt : ID { savedName = copyString(tokenString);
                   savedLineNo = lineno; }
              ASSIGN exp
                 { $$ = newStmtNode(AssignK);
                   $$->child[0] = $4;
                   $$->attr.name = savedName;
                   $$->lineno = savedLineNo;
                 }
            ;
read_stmt   : READ ID
                 { $$ = newStmtNode(ReadK);
                   $$->attr.name =
                     copyString(tokenString);
                 }
            ;
write_stmt  : WRITE exp
                 { $$ = newStmtNode(WriteK);
                   $$->child[0] = $2;
                 }
            ;
exp         : simple_exp LT simple_exp
                 { $$ = newExpNode(opK);
                   $$->child[0] = $1;
                   $$->child[1] = $3;
                   $$->attr.op = LT;
                 }
            | simple_exp EQ simple_exp
                 { $$ = newExpNode(opK);
                   $$->child[0] = $1;
                   $$->child[1] = $3;
                   $$->attr.op = EQ;
                 }
            | simple_exp { $$ = $1; }
            ;
simple_exp  : simple_exp PLUS term
                 { $$ = newExpNode(opK);
                   $$->child[0] = $1;
                   $$->child[1] = $3;
                   $$->attr.op = PLUS;
                 }
            | simple_exp MINUS term
                 { $$ = newExpNode(opK);
                   $$->child[0] = $1;
                   $$->child[1] = $3;
                   $$->attr.op = MINUS;
                 }
            | term { $$ = $1; }
            ;
term        : term TIMES factor
                 { $$ = newExpNode(opK);
                   $$->child[0] = $1;
                   $$->child[1] = $3;
                   $$->attr.op = TIMES;
                 }
            | term OVER factor
                 { $$ = newExpNode(opK);
                   $$->child[0] = $1;
                   $$->child[1] = $3;
                   $$->attr.op = OVER;
                 }
            | factor { $$ = $1; }
            ;
factor      : LPAREN exp RPAREN
                 { $$ = $2; }
            | NUM
                 { $$ = newExpNode(ConstK);
                   $$->attr.val = atoi(tokenString);
                 }
            | ID { $$ = newExpNode(IdK);
                   $$->attr.name =
                         copyString(tokenString);
                 }
            ;

%%

 int yylex(void)
{ return getToken(); }

TreeNode * parse(void)
{ yyparse();
  return savedTree;
}

生成yyparser.h和yyparser.c文件，加入到Tiny编译器工程中，生成的语法树结构如下：

Syntax tree:
----Read
----If
--------Op: <
------------const: 0
------------ID: x
--------Assign To
------------const: 1
--------Repeat
------------Assign To
----------------Op: *
--------------------ID: fact
--------------------ID: x
------------Assign To
----------------Op: -
--------------------ID: x
--------------------const: 1
------------Op: =
----------------ID: x
----------------const: 0
--------Write
------------ID: fact
  可以看到，和手工编写的递归向下语法分析器的结果是一样的。