lexer.java

编译原理大作业---tiger编译器 包括semant,translate,mipsframe,regalloc等所有phase 懂的人自会知道

package java_cup;

import java_cup.runtime.Symbol;
import java.util.Hashtable;

/** This class implements a small scanner (aka lexical analyzer or lexer) for
 *  the JavaCup specification.  This scanner reads characters from standard 
 *  input (System.in) and returns integers corresponding to the terminal 
 *  number of the next Symbol. Once end of input is reached the EOF Symbol is 
 *  returned on every subsequent call.<p>
 *  Symbols currently returned include: <pre>
 *    Symbol        Constant Returned     Symbol        Constant Returned
 *    ------        -----------------     ------        -----------------
 *    "package"     PACKAGE               "import"      IMPORT 
 *    "code"        CODE                  "action"      ACTION 
 *    "parser"      PARSER                "terminal"    TERMINAL
 *    "non"         NON                   "init"        INIT 
 *    "scan"        SCAN                  "with"        WITH
 *    "start"       START                 "precedence"  PRECEDENCE
 *    "left"        LEFT		  "right"       RIGHT
 *    "nonassoc"    NONASSOC		  "%prec        PRECENT_PREC  
 *      [           LBRACK                  ]           RBRACK
 *      ;           SEMI 
 *      ,           COMMA                   *           STAR 
 *      .           DOT                     :           COLON
 *      ::=         COLON_COLON_EQUALS      |           BAR
 *    identifier    ID                    {:...:}       CODE_STRING
 *  </pre>
 *  All symbol constants are defined in sym.java which is generated by 
 *  JavaCup from parser.cup.<p>
 * 
 *  In addition to the scanner proper (called first via init() then with
 *  next_token() to get each Symbol) this class provides simple error and 
 *  warning routines and keeps a count of errors and warnings that is 
 *  publicly accessible.<p>
 *  
 *  This class is "static" (i.e., it has only static members and methods).
 *
 * @version last updated: 7/3/96
 * @author  Frank Flannery
 */
public class lexer {

  /*-----------------------------------------------------------*/
  /*--- Constructor(s) ----------------------------------------*/
  /*-----------------------------------------------------------*/

  /** The only constructor is private, so no instances can be created. */
  private lexer() { }

  /*-----------------------------------------------------------*/
  /*--- Static (Class) Variables ------------------------------*/
  /*-----------------------------------------------------------*/

  /** First character of lookahead. */
  protected static int next_char; 

  /** Second character of lookahead. */
  protected static int next_char2;

  /** Second character of lookahead. */
  protected static int next_char3;

  /** Second character of lookahead. */
  protected static int next_char4;

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** EOF constant. */
  protected static final int EOF_CHAR = -1;

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Table of keywords.  Keywords are initially treated as identifiers.
   *  Just before they are returned we look them up in this table to see if
   *  they match one of the keywords.  The string of the name is the key here,
   *  which indexes Integer objects holding the symbol number. 
   */
  protected static Hashtable keywords = new Hashtable(23);

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Table of single character symbols.  For ease of implementation, we 
   *  store all unambiguous single character Symbols in this table of Integer
   *  objects keyed by Integer objects with the numerical value of the 
   *  appropriate char (currently Character objects have a bug which precludes
   *  their use in tables).
   */
  protected static Hashtable char_symbols = new Hashtable(11);

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Current line number for use in error messages. */
  protected static int current_line = 1;

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Character position in current line. */
  protected static int current_position = 1;

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Character position in current line. */
  protected static int absolute_position = 1;

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Count of total errors detected so far. */
  public static int error_count = 0;

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Count of warnings issued so far */
  public static int warning_count = 0;

  /*-----------------------------------------------------------*/
  /*--- Static Methods ----------------------------------------*/
  /*-----------------------------------------------------------*/

  /** Initialize the scanner.  This sets up the keywords and char_symbols
    * tables and reads the first two characters of lookahead.  
    */
  public static void init() throws java.io.IOException
    {
      /* set up the keyword table */
      keywords.put("package",    new Integer(sym.PACKAGE));
      keywords.put("import",     new Integer(sym.IMPORT));
      keywords.put("code",       new Integer(sym.CODE));
      keywords.put("action",     new Integer(sym.ACTION));
      keywords.put("parser",     new Integer(sym.PARSER));
      keywords.put("terminal",   new Integer(sym.TERMINAL));
      keywords.put("non",        new Integer(sym.NON));
      keywords.put("init",       new Integer(sym.INIT));
      keywords.put("scan",       new Integer(sym.SCAN));
      keywords.put("with",       new Integer(sym.WITH));
      keywords.put("start",      new Integer(sym.START));
      keywords.put("precedence", new Integer(sym.PRECEDENCE));
      keywords.put("left",       new Integer(sym.LEFT));
      keywords.put("right",      new Integer(sym.RIGHT));
      keywords.put("nonassoc",   new Integer(sym.NONASSOC));

      /* set up the table of single character symbols */
      char_symbols.put(new Integer(';'), new Integer(sym.SEMI));
      char_symbols.put(new Integer(','), new Integer(sym.COMMA));
      char_symbols.put(new Integer('*'), new Integer(sym.STAR));
      char_symbols.put(new Integer('.'), new Integer(sym.DOT));
      char_symbols.put(new Integer('|'), new Integer(sym.BAR));
      char_symbols.put(new Integer('['), new Integer(sym.LBRACK));
      char_symbols.put(new Integer(']'), new Integer(sym.RBRACK));

      /* read two characters of lookahead */
      next_char = System.in.read();
      if (next_char == EOF_CHAR) {
	next_char2 = EOF_CHAR;
        next_char3 = EOF_CHAR;
        next_char4 = EOF_CHAR;
      } else {
	next_char2 = System.in.read();
	if (next_char2 == EOF_CHAR) {
	  next_char3 = EOF_CHAR;
	  next_char4 = EOF_CHAR;
	} else {
	  next_char3 = System.in.read();
	  if (next_char3 == EOF_CHAR) {
	    next_char4 = EOF_CHAR;
	  } else {
	    next_char4 = System.in.read();
	  }
	}
      }
    }

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Advance the scanner one character in the input stream.  This moves
   * next_char2 to next_char and then reads a new next_char2.  
   */
  protected static void advance() throws java.io.IOException
    {
      int old_char;

      old_char = next_char;
      next_char = next_char2;
      if (next_char == EOF_CHAR) {
	next_char2 = EOF_CHAR;
        next_char3 = EOF_CHAR;
	next_char4 = EOF_CHAR;
      } else {
	next_char2 = next_char3;
	if (next_char2 == EOF_CHAR) {
	  next_char3 = EOF_CHAR;
	  next_char4 = EOF_CHAR;
	} else {
	  next_char3 = next_char4;
	  if (next_char3 == EOF_CHAR) {
	    next_char4 = EOF_CHAR;
	  } else {
	    next_char4 = System.in.read();
	  }
	}
      }

      /* count this */
      absolute_position++;
      current_position++;
      if (old_char == '\n')
	{
	  current_line++;
	  current_position = 1;
	}
    }

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Emit an error message.  The message will be marked with both the 
   *  current line number and the position in the line.  Error messages
   *  are printed on standard error (System.err).
   * @param message the message to print.
   */
  public static void emit_error(String message)
    {
      System.err.println("Error at " + current_line + "(" + current_position +
			 "): " + message);
      error_count++;
    }

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Emit a warning message.  The message will be marked with both the 
   *  current line number and the position in the line.  Messages are 
   *  printed on standard error (System.err).
   * @param message the message to print.
   */
  public static void emit_warn(String message)
    {
      System.err.println("Warning at " + current_line + "(" + current_position +
			 "): " + message);
      warning_count++;
    }

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Determine if a character is ok to start an id. 
   * @param ch the character in question.
   */
  protected static boolean id_start_char(int ch)
    {
      /* allow for % in identifiers.  a hack to allow my
	 %prec in.  Should eventually make lex spec for this 
	 frankf */
      return (ch >= 'a' &&  ch <= 'z') || (ch >= 'A' && ch <= 'Z') || 
	     (ch == '_');

      // later need to deal with non-8-bit chars here
    }

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Determine if a character is ok for the middle of an id.
   * @param ch the character in question. 
   */
  protected static boolean id_char(int ch)
    {
      return id_start_char(ch) || (ch >= '0' && ch <= '9');
    }

  /*. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*/

  /** Try to look up a single character symbol, returns -1 for not found. 
   * @param ch the character in question.
   */