1.cpp

用C写的简易C编译器

    s1->group='G';
    s1->line=line;
    sos.push(s1);
    s1=new Symbol;
    s1->group=')';
    s1->line=line;
    sos.push(s1);
    s1=new Symbol;
    s1->group='S';
    s1->line=line;
    sos.push(s1);
    continue;
   case 31:
    err(7,line);
    it=ip;
    ip=new Symbol;
    ip->group=';';
    ip->line=line;
    continue;
   case 32:
    err(9,line);
    s1=new Symbol;
    s1->line=line;
    s1->group='S';
    sos.push(s1);
    break;
   default:
    err(4,ip->line);
    goto label1;
   }//switch
   r=ss.top();
   iq=sos.top();
   r=Goto::lookUp(iq->group,r);
   if(r)
    ss.push(r);
   else
    err(4,iq->line);
  }//else
  else if(t==0&&!hasError){
   log<<endl;
   log<<"代码通过语法检查,未发现语法错误."<<endl;
  }
 }//while

label1:


 in.close();
 log<<"语法分析完毕..."<<endl;
 log<<"*******************************************"<<endl;
}

Compiler::preProcess()
{
 string tmpstr;
 ifstream ins((fileName+".crr").c_str(),ios::in);
 log<<"*******************************************"<<endl;
 log<<"程序预处理开始"<<endl;
 if(ins.is_open()){
  hasFile=1;
  log<<"打开文件 \'"<<fileName<<".crr\' 成功!"<<endl;
 }
 else{
  hasFile=0;
  log<<"打开文件失改!"<<endl;
 }
 while(getline(ins,tmpstr)){
  if(tmpstr.substr(0,18)=="#import system.out"){
   needOutSuppose=1;
   break;
  }
 }
 if(needOutSuppose)
  log<<"需要引入系统函数\'System.out\'."<<endl;
 else
  log<<"不需要引入系统函数\'System.out\'."<<endl;
 log<<"程序预处理完成!"<<endl;
 log<<"*******************************************"<<endl;
 ins.close();
}

char Compiler::nextChar()
{
 char ch;
 if(in.get(ch)){
  currentChar=ch;
  return ch;
 }
 else{
  currentChar=ch;
  return '$';
 }
}

void Compiler::emitter()
{
 if(hasError){
  log<<endl;
  log<<"*******************************************"<<endl;
  log<<"源代码中有错误,不能生成目标代码!"<<endl;
  log<<"*******************************************"<<endl;
  return;
 }
 log<<endl;
 log<<"*******************************************"<<endl;
 log<<"开始生成目标代码..."<<endl;
 out.open((fileName + ".asm").c_str(),ios::out);
 if(out.is_open()){
 log<<"创建文件"<<(fileName + ".asm").c_str()<<"成功!"<<endl;
 out<<"\t.model small"<<endl;
 out<<"\t.stack"<<endl;
 out<<"\t.data"<<endl;
 while(!symbolList.empty())
 {
  out<<symbolList.front().c_str()<<"\tdw ?"<<endl;
  symbolList.pop_front();
 }
 out<<"__end\tdb 13,10,\'Press Any Key To Continue...\',\'$\'"<<endl;
 out<<"\t.code"<<endl;
 out<<endl;
 if(needOutSuppose){
  log<<"写入被引入系统函数代码."<<endl;
  out<<"outint proc near"<<endl;
  out<<"\tmov cx,0"<<endl;
  out<<"\tmov bx,10"<<endl;
  out<<"again:\tmov dx,0"<<endl;
  out<<"\tdiv bx"<<endl;
  out<<"\tpush dx"<<endl;
  out<<"\tinc cx"<<endl;
  out<<"\tcmp ax,0"<<endl;
  out<<"\tjne again"<<endl;
  out<<"print:\tpop dx"<<endl;
  out<<"\tadd dl,30h"<<endl;
  out<<"\tmov ah,02h"<<endl;
  out<<"\tint 21h"<<endl;
  out<<"\tloop print"<<endl;
  out<<"\tmov dl,20h"<<endl;
  out<<"\tmov ah,02h"<<endl;
  out<<"\tint 21h"<<endl;
  out<<"\tret"<<endl;
  out<<"outint\tendp"<<endl;
  out<<endl;
 }
 out<<"main\tproc near"<<endl;
 out<<"\t;***CODE START***"<<endl;
 log<<"开始优化代码."<<endl;
 optimize();
 log<<"优化代码完毕."<<endl;
 out<<code;
 log<<"写入目标代码."<<endl;
 out<<"\t;***CODE END***"<<endl;
 out<<"\tret"<<endl;
 out<<"main endp"<<endl;
 out<<endl;
 out<<"start:"<<endl;
 out<<"\tmov ax,@data"<<endl;
 out<<"\tmov ds,ax"<<endl;
 out<<"\tcall main"<<endl;
 out<<"\tlea dx,__end"<<endl;
 out<<"\tmov ah,09h"<<endl;
 out<<"\tint 21h"<<endl;
 out<<"\tmov ah,08h"<<endl;
 out<<"\tint 21h"<<endl;
 out<<"\tmov ah,4ch"<<endl;
 out<<"\tint 21h"<<endl;
 out<<"\tend start"<<endl;
 out.close();
 log<<endl;
 log<<"代码生成完成!"<<endl;
 log<<"*******************************************"<<endl;
 }

 else
 {
  log<<endl;
  log<<"无法创建文件!"<<endl;
  log<<"*******************************************"<<endl;

 }
// Label *q;//测试Label
// for(int i=0;i<10;i++){
//   q=new Label;
//   log<<q->text.c_str()<<endl;
//  delete q;
// }
 return;
}

Compiler::err(int no,int line)
{
 string errText;
 hasError=1;
 switch(no){
 case 1:
  errText="出现无法识别的符号.";
  break;
 case 2:
  errText="错误的标识符.";
  break;
 case 3:
  errText="运算符\'&&\'(\'||\')写成\'&\'(\'|\')了.";
  break;
 case 4:
  errText="语法错误,编译停止!";
  break;
 case 5:
  errText="找不到main()函数.";
  break;
 case 6:
  errText="所使用的函数未定义";
  break;
 case 7:
  errText="语句后缺少\';\'.";
  break;
 case 8:
  errText="else没有与之匹配的if.";
  break;
 case 9:
  errText="缺少语句.";
  break;
 case 10:
  errText="变量未定义.";
  break;
 case 11:
  errText="变量以定义过,不能重新定义.";
  break;
 default:
  errText="未知错误!";
 }
 log<<"发生错误:文件"<<fileName<<".crr第"<<line<<"行:"<<errText<<endl;
}


int Compiler::lookup(string m)
{
 list<string>::iterator i;
 for (i =  symbolList.begin(); i != symbolList.end(); ++i)
  if(*i==m)
   return 1;
 return 0;

}

 

 

Symbol::Symbol()
{

}

Symbol::Symbol(const Symbol &b)
{
 group=b.group;
 line=b.line;
 word=b.word;
}

Symbol::~Symbol()
{

}

Symbol::operator =(const Symbol &b)
{
 group=b.group;
 line=b.line;
 word=b.word;
}


Label::Label()
{
 n=next();
 char buffer[6];
 text=_itoa(n,buffer,10);
 text="L"+text;
}

Label::~Label()
{

}

int Label::next()
{
 return ++_label;
}

int Label::_label=0;

Action::Action()
{

}

Action::~Action()
{

}

int Action::lookUp(char v,int s)
{
 int n=vs.find_first_of(v,0);
 return Table[s-1][n];
}

string Action::vs ="+*>&=!@n(){};,ziew$";

int Action::Table[54][19]={
-40,-40,-40,-40,-40,-40,3,-40,-40,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state1
-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-30,-40,0,//state2
-40,-40,-40,-40,-40,-40,-40,-40,4,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state3
-40,-40,-40,-40,-40,-40,-40,-40,-40,5,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state4
-40,-40,-40,-40,-40,-40,7,-40,-40,-40,8,-40,-40,-40,11,9,-30,10,-40,//state5
-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-30,-40,-1,//state6
-40,-40,-40,-40,12,-40,-40,-40,13,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state7
-40,-40,-40,-40,-40,-40,7,-40,-40,-40,8,-32,-32,-40,11,9,-30,10,-32,//state8
-40,-40,-40,-40,-40,-40,-40,-40,13,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state9
-40,-40,-40,-40,-40,-40,-40,-40,13,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state10
-40,-40,-40,-40,-40,-40,17,-40,-40,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state11
-40,-40,-40,-40,-40,-40,22,23,21,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state12
-40,-40,-40,-40,-40,27,22,23,28,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state13
-40,-40,-40,-40,-40,-40,7,-40,-40,-40,8,29,-40,-40,11,9,-30,10,-40,//state14
-40,-40,-40,-40,-40,-40,-10,-40,-40,-40,-10,-10,-40,-40,-10,-10,-30,-10,-40,//state15
-40,-40,-40,-40,-40,-40,-31,-40,-40,-40,-31,-31,31,32,-31,-31,-30,-31,-31,//state16
-40,-40,-40,-40,-40,-40,-31,-40,-40,-40,-31,-31,-11,-11,-31,-31,-30,-31,-31,//state17
35,-40,-40,-40,-40,-40,-31,-40,-40,34,-31,-31,33,-40,-31,-31,-30,-31,-31,//state18
-19,36,-19,-19,-40,-40,-31,-40,-40,-19,-31,-31,-19,-40,-31,-31,-30,-31,-31,//state19
-21,-21,-21,-21,-40,-40,-31,-40,-40,-21,-31,-31,-21,-40,-31,-31,-30,-31,-31,//state20
-40,-40,-40,-40,-40,-40,22,23,21,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state21
-23,-23,-23,-23,-40,-40,-31,-40,-40,-23,-31,-31,-23,-40,-31,-31,-30,-31,-31,//state22
-24,-24,-24,-24,-40,-40,-31,-40,-40,-24,-31,-31,-24,-40,-31,-31,-30,-31,-31,//state23
-40,-40,-40,38,-40,-40,-40,-40,-40,34,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state24
35,-40,39,-40,-40,-40,-31,-40,-40,34,-31,-31,33,-40,-31,-31,-30,-31,-31,//state25
-40,-40,-40,-14,-40,-40,-40,-40,-40,-14,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state26
-40,-40,-40,-40,-40,27,22,23,28,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state27
-40,-40,-40,-40,-40,27,22,23,28,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state28
-40,-40,-40,-40,-40,-40,-3,-40,-40,-40,-3,-3,-40,-40,-3,-3,-3,-3,-3,//state29
-40,-40,-40,-40,-40,-40,-9,-40,-40,-40,-9,-9,-40,-40,-9,-9,-9,-9,-40,//state30
-40,-40,-40,-40,-40,-40,-7,-40,-40,-40,-7,-7,-40,-40,-7,-7,-7,-7,-7,//state31
-40,-40,-40,-40,-40,-40,44,-40,-40,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state32
-40,-40,-40,-40,-40,-40,-2,-40,-40,-40,-2,-2,-40,-40,-2,-2,-2,-2,-2,//state33
-40,-40,-40,-40,-40,-40,7,-40,-40,-40,8,-32,46,-40,11,9,-30,10,-32,//state34
-40,-40,-40,-40,-40,-40,22,23,21,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state35
-40,-40,-40,-40,-40,-40,22,23,21,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state36
35,-40,-40,-40,-40,-40,-40,-40,-40,49,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state37
-40,-40,-40,-40,-40,27,22,23,28,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state38
-40,-40,-40,-40,-40,-40,22,23,21,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state39
-40,-40,-40,-16,-40,-40,-40,-40,-40,-16,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state40
35,-40,39,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state41
-40,-40,-40,38,-40,-40,-40,-40,-40,52,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state42
35,-40,39,-40,-40,-40,-31,-40,-40,49,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state43
-40,-40,-40,-40,-40,-40,-31,-40,-40,-40,-31,-31,-12,-12,-31,-31,-30,-31,-31,//state44
-40,-40,-40,-40,-40,-40,-4,-40,-40,-40,-4,-4,-40,-40,-4,-4,53,-4,-4,//state45
-40,-40,-40,-40,-40,-40,-8,-40,-40,-40,-8,-8,-40,-40,-8,-8,-30,-8,-8,//state46
-18,36,-18,-18,-40,-40,-31,-40,-40,-18,-31,-31,-18,-40,-31,-31,-30,-31,-31,//state47
-20,-20,-20,-20,-40,-40,-31,-40,-40,-20,-31,-31,-20,-40,-31,-31,-30,-31,-31,//state48
-22,-22,-22,-22,-40,-40,-31,-40,-40,-22,-31,-31,-22,-40,-31,-31,-30,-31,-31,//state49
-40,-40,-40,-13,-40,-40,-40,-40,-40,-13,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state50
35,-40,-40,-15,-40,-40,-40,-40,-40,-15,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state51
-40,-40,-40,-17,-40,-40,-40,-40,-40,-17,-40,-40,-40,-40,-40,-40,-30,-40,-40,//state52
-40,-40,-40,-40,-40,-40,7,-40,-40,-40,8,-32,-32,-40,11,9,-30,10,-32,//state53
-40,-40,-40,-40,-40,-40,-5,-40,-40,-40,-5,-5,-40,-40,-5,-5,-30,-5,-5};//state54


Goto::Goto()
{

}

Goto::~Goto()
{

}

int Goto::lookUp(char v,int s)
{
 int n=vs.find_first_of(v,0);
 return Table[s-1][n];
}

string Goto::vs="DSWLGMETF";

int Goto::Table[54][9]={
2,0,0,0,0,0,0,0,0,//state1
0,0,0,0,0,0,0,0,0,//state2
0,0,0,0,0,0,0,0,0,//state3
0,0,0,0,0,0,0,0,0,//state4
0,6,0,0,0,0,0,0,0,//state5
0,0,0,0,0,0,0,0,0,//state6
0,0,0,0,0,0,0,0,0,//state7
0,15,14,0,0,0,0,0,0,//state8
0,0,0,0,0,0,0,0,0,//state9
0,0,0,0,0,0,0,0,0,//state10
0,0,0,16,0,0,0,0,0,//state11
0,0,0,0,0,0,18,19,20,//state12
0,0,0,0,24,26,25,19,20,//state13
0,30,0,0,0,0,0,0,0,//state14
0,0,0,0,0,0,0,0,0,//state15
0,0,0,0,0,0,0,0,0,//state16
0,0,0,0,0,0,0,0,0,//state17
0,0,0,0,0,0,0,0,0,//state18
0,0,0,0,0,0,0,0,0,//state19
0,0,0,0,0,0,0,0,0,//state20
0,0,0,0,0,0,37,19,20,//state21
0,0,0,0,0,0,0,0,0,//state22
0,0,0,0,0,0,0,0,0,//state23
0,0,0,0,0,0,0,0,0,//state24
0,0,0,0,0,0,0,0,0,//state25
0,0,0,0,0,0,0,0,0,//state26
0,0,0,0,0,40,41,19,20,//state27
0,0,0,0,42,26,43,19,20,//state28
0,0,0,0,0,0,0,0,0,//state29
0,0,0,0,0,0,0,0,0,//state30
0,0,0,0,0,0,0,0,0,//state31
0,0,0,0,0,0,0,0,0,//state32
0,0,0,0,0,0,0,0,0,//state33
0,45,0,0,0,0,0,0,0,//state34
0,0,0,0,0,0,0,47,20,//state35
0,0,0,0,0,0,0,0,48,//state36
0,0,0,0,0,0,0,0,0,//state37
0,0,0,0,0,50,41,19,20,//state38
0,0,0,0,0,0,51,19,20,//state39
0,0,0,0,0,0,0,0,0,//state40
0,0,0,0,0,0,0,0,0,//state41
0,0,0,0,0,0,0,0,0,//state42
0,0,0,0,0,0,0,0,0,//state43
0,0,0,0,0,0,0,0,0,//state44
0,0,0,0,0,0,0,0,0,//state45
0,0,0,0,0,0,0,0,0,//state46
0,0,0,0,0,0,0,0,0,//state47
0,0,0,0,0,0,0,0,0,//state48
0,0,0,0,0,0,0,0,0,//state49
0,0,0,0,0,0,0,0,0,//state50
0,0,0,0,0,0,0,0,0,//state51
0,0,0,0,0,0,0,0,0,//state52
0,54,0,0,0,0,0,0,0,//state53
0,0,0,0,0,0,0,0,0//state54
};

Compiler::optimize()
{
 int i,j,f;
 string t="\tpush ax\n\tpop ax\n";
 do{
  f=0;
  j=code.length()-17;
  for(i=0;i<j;i++){
   if(code.substr(i,17)==t){
    f=1;
    code.replace(i,17,"");
    break;
   }
  }
 }while(f);
}

 

int APIENTRY WinMain(HINSTANCE hInstance,
                     HINSTANCE hPrevInstance,
                     LPSTR     lpCmdLine,
                     int       nCmdShow)
{
 string file(lpCmdLine);
 file=file.substr(0,file.find_last_of('.'));
 if(!file.empty()){
  Compiler cc(file);
  cc.preProcess();
  cc.parser();
  cc.emitter();
  ofstream bat;
  bat.open((file+".bat").c_str(),ios::out);
  bat<<"masm "<<file<<".asm"<<endl;
  bat<<"link "<<file<<".obj"<<endl;
  bat.close();
 }
 return 0;
}