sub.c

一个非常小的C语言编译器!!! !!! !!! !!!!

//=====================================================//
//   StringT, Token, SymbolT Implementation  "sub.C"   //
//=====================================================//
#include "sub.h"
#include "arm.h"
#include "mips.h"
//                                               Tan Watanabe, UEC
//========== StringT Class Implementation ==========//

//---- Globals and statics ----//
  String  str_null;            // Null string ("").
  StringP StringT::first;      // First   StringT instance.
  StringP StringT::current;    // Current StringT instance.

//--------------------//
void StringT::Initiation()     // Initiate the string block list.
{
  first    = this;
  current  = first;
  str_null = Append(""); 
} //---- Initiation end ----//

//--------------------//
String StringT:: Append(String pStr) // Append pStr to the current block.
{                                    // If no room, allocate the next block.
  StringP lNext;
  String  lStr;
  int     lLeng;
  
  lLeng = strlen(pStr);
  if ((current->top+lLeng) >= current->tail) { // There is no room.
    lNext = new StringT;                       // Allocate the next block
    current->nextBlk = lNext;
    current = lNext;
    if ((current->top+lLeng) >= current->tail) // pStr is too long.
      return (NULL);             // Error
  }
  strcpy(current->top, pStr);    // Append the string to the tail
  lStr = current->top;           //   of the current block.
  current->top += (lLeng+1);     // Add length + 1(for '\0')
  return lStr;
} //---- Append end ====//

//========== Token Class Implementation ==========//

//---- Global variables ----//
Tkn_kindT tkn_chKindT[384];  // Token kind guessed by 1st char.
Token     tkn_curr;          // Current token
String    tkn_kindName[] =
    {"blank  ","spec   ","compoun","eol    ","eof    ",
     "intC   ","key    ","id     ","var    ","param  ",
     "func   ","type   ","label  ","oper   ","reg    ","mrg    " };

//--------------------//
void Token::Initiation()     // Initiate the token class.
{
  int i;
  
  tknC   = TKN_eol;  
  leng   = 0;
  value  = 0;
  for (i = 0; i < TKN_lim; i++) 
    name[i] = ' ';
  for (i = 0; i < 384; i++)           // Initiate by special char kind.
    tkn_chKindT[i] = tknSpec;         // (tkn_chKindT[0..126] is unused)
  for (i = '0'; i <= '9'; i++)        // Numeric constant header.
    tkn_chKindT[i+128] = tknIntC;     // char is -128 ~ 127
  for (i = 'A'; i <= 'Z'; i++)        // Identifier initial.
    tkn_chKindT[i+128] = tknId; 
  for (i = 'a'; i <= 'z'; i++)        // Identifier initial.
    tkn_chKindT[i+128] = tknId; 
  tkn_chKindT['_'+128] = tknId;       // GIAN 011107
  tkn_chKindT[' '+128] = tknBlank;    // Blank
  tkn_chKindT['='+128] = tknCompound; // Compound operator initial.
  tkn_chKindT['>'+128] = tknCompound;
  tkn_chKindT['<'+128] = tknCompound;
  tkn_chKindT['!'+128] = tknCompound;
  tkn_chKindT[TKN_eol+128] = tknEol;  // End of line.
  tkn_chKindT[TKN_eof+128] = tknEof;  // End of file (EOF == -1).
} //---- Initiation end ----//

//--------------------//
void Token::Get()             // Get the next token from inFile.
{
  while ((tknC != TKN_eof)&&
         ((tknC == ' ')||(tknC == TKN_tab)||(tknC == TKN_eol)))
    tknC = NextC();              // Skip blanks, tabs, and EOL.
  leng = 0;
  kind = tkn_chKindT[tknC+128];  // Guess the token kind by 1st char.
  if (kind == tknId) {           // Identifier.
    do {
      if (leng < TKN_lim)        // Cut out the identifier.
        name[leng] = tknC;
      leng++;
      tknC = NextC();
    } while ((tkn_chKindT[tknC+128] == tknId)||
             (tkn_chKindT[tknC+128] == tknIntC));
    if (leng > TKN_lim)          // Truncate if too long.
      leng = TKN_lim;
  }else if (kind == tknIntC) {   // Numeric constant.
    value = 0;
    do {                         // Cut out the integer constant.
      value = value*10 + ((int)tknC - (int)'0');
      if (leng < TKN_lim)
        name[leng] = tknC;
      leng++;
      tknC = NextC();
    } while (tkn_chKindT[tknC+128] == tknIntC);
    if (leng > TKN_lim)          // Truncate if too long.
      leng = TKN_lim;
  }else {            // Special character or compound operator.
    name[0] = tknC;
    leng = 1;
    tknC = NextC();
    if (kind == tknCompound) {  // May be compound operator.
      if (((name[0] == '=')&&((tknC == '=')||(tknC == '>'))) ||
          ((name[0] == '!')&&(tknC == '=')) ||
          ((name[0] == '<')&&(tknC == '=')) ||
          ((name[0] == '<')&&(tknC == '<')) ||
          ((name[0] == '>')&&(tknC == '=')) ||
          ((name[0] == '>')&&(tknC == '>')) ||
          ((name[0] == '-')&&(tknC == '>')) ) { // Compound operator.
        name[1] = tknC;
        leng = 2;
        tknC = NextC();
      }
    }else if ((name[0] == '/')&&(tknC == '*')) {  // Comment start
      tknC = NextC();
      while ((tknC != TKN_eof)&&
             ((tknC != '*')||((tknC = NextC()) != '/'))) 
        tknC = NextC();
      tknC = NextC();
      Get();
    }
  }
  name[leng] = '\0';            // Attach end_of_string indication.
} //---- Get end ----//

//-------- Static Variables --------//
char Token::tknC;           // Current token character.

//========== SymbolTable Class Implementation ==========//

//-------- Global Variables --------//
  StringT sym_string;       // String table containing symbol names.
  SymbolP                   // Predefined symbols.
    sym_eof,    sym_lpar,   sym_rpar,  sym_lbracket,sym_rbracket,
    sym_lbrace, sym_rbrace, sym_comma, sym_scolon,  sym_period,
    sym_arrow,  sym_int,    sym_ptr,   sym_void,    sym_extern,  
    sym_if ,    sym_else,   sym_while, sym_return,  sym_add ,
    sym_sub,    sym_mult,   sym_div,   sym_assign,  sym_equal,
    sym_ne,     sym_gt,     sym_ge,    sym_lt,      sym_le,
    sym_or,     sym_and,    sym_shl,   sym_shr,     sym_mod;
    //    sym_main

  SymbolT sym_tbl;      // Pre-declared symbol table.
  SymbolP sym_curr;     // Current symbol.
  int     sym_tempVar;  // Generated temporal index.

//--------------------//
void SymbolT::Initiation()    
     // Initiate symbol table and string table.
     // (initiate hash table, set predefined symbols, and so on.) 
{
  char lEof[2];
  
  for (int i = 0; i < SYM_hashLim; i++)
    hash[i] = NULL;                     // Nullify the hash table.
  sym_string.Initiation();              // Initiate the string table.
  head         = this;                  // Set pointers.
  tail         = head;
  name         = str_null;
  lEof[0]      = TKN_eof;               // End file code
  lEof[1]      = '\0';                  // followed by end of string.
  sym_eof      = Set(lEof,  tknEof);    // Set availabel special characters
  sym_lpar     = Set("(",   tknSpec);   // in the symbol table.
  sym_rpar     = Set(")",   tknSpec);
  sym_lbracket = Set("[",   tknSpec);
  sym_rbracket = Set("]",   tknSpec);
  sym_lbrace   = Set("{",   tknSpec);
  sym_rbrace   = Set("}",   tknSpec);
  sym_comma    = Set(",",   tknSpec);
  sym_scolon   = Set(";",   tknSpec);
  sym_period   = Set(".",   tknSpec);
  sym_int      = Set("int",    tknType);  // Set key words in the symbol table.
  sym_void     = Set("void",   tknType);
  sym_extern   = Set("extern", tknType);
  sym_if       = Set("if",     tknKey);
  sym_else     = Set("else",   tknKey);
  sym_while    = Set("while",  tknKey);
  sym_return   = Set("return", tknKey);
  sym_add      = SetOp("+" ,tknOper, opAdd ); // Set operators in the 
  sym_sub      = SetOp("-" ,tknOper, opSub ); //   symbol table.
  sym_mult     = SetOp("*" ,tknOper, opMult);
  sym_div      = SetOp("/" ,tknOper, opDiv );
  sym_assign   = SetOp("=" ,tknOper, opStore);
  sym_equal    = SetOp("==",tknOper, opBrne); 
  sym_ne       = SetOp("!=",tknOper, opBreq);
  sym_gt       = SetOp(">" ,tknOper, opBrle);
  sym_ge       = SetOp(">=",tknOper, opBrlt);
  sym_lt       = SetOp("<" ,tknOper, opBrge);
  sym_le       = SetOp("<=",tknOper, opBrgt);
  //  sym_main     = Set("main",tknFunc);
  sym_int->SizeSet(4);
} //---- Initiation end ----//

//--------------------//
SymbolP SymbolT::SearchOrAdd( // Search given symbol in the symbol table
                          // and add it if not found.
  String    pString,      //   Name string of the symbol to be searched.
  int       pLeng,        //   Character length of pString.
  Tkn_kindT pKind)        //   Token kind of the symbol. For identifier,
                          //   give tknId if it is not yet declared.
                          // Return symbol table ptr of the symbol.
{
  int     lHash;
  SymbolP lSymp, lPrev, lNew;
  String  lName;
  
  lHash = HashKey(pString, pLeng);
  lSymp = hash[lHash];
  lPrev = NULL;
  while ((lSymp != NULL)&&strcmp(pString, lSymp->name)) { 
    lPrev = lSymp;                      // Not equal.
    lSymp = lSymp->hashChain;           // Advance to the next symbol.
  }