date.java

gcc的组建

   * recognised by <code>parse()</code> (e.g. GMT, CET, etc.)
   * and may reflect the fact that daylight savings time is in
   * effect.  The empty string is used if there is no time zone
   * information.
   * </li>
   * <li>
   * <code>yyyy</code> -- the year as four decimal digits.
   * </li>
   * </ul>
   * <p>
   * The <code>DateFormat</code> class should now be 
   * preferred over using this method.
   * </p>
   *
   * @return A string of the form 'day mon dd hh:mm:ss zz yyyy'
   * @see #parse(String)
   * @see DateFormat
   */
  public String toString()
  {
    Calendar cal = Calendar.getInstance();
    cal.setTimeInMillis(time);
    String day = "0" + cal.get(Calendar.DATE);
    String hour = "0" + cal.get(Calendar.HOUR_OF_DAY);
    String min = "0" + cal.get(Calendar.MINUTE);
    String sec = "0" + cal.get(Calendar.SECOND);
    String year = "000" + cal.get(Calendar.YEAR);
    return weekNames[cal.get(Calendar.DAY_OF_WEEK) - 1] + " "
      + monthNames[cal.get(Calendar.MONTH)] + " "
      + day.substring(day.length() - 2) + " "
      + hour.substring(hour.length() - 2) + ":"
      + min.substring(min.length() - 2) + ":"
      + sec.substring(sec.length() - 2) + " "
      +
      cal.getTimeZone().getDisplayName(cal.getTimeZone().inDaylightTime(this),
				       TimeZone.SHORT) + " " +
      year.substring(year.length() - 4);
  }

  /** 
   * Returns a locale-dependent string representation of this
   * <code>Date</code> object.
   *
   * @deprecated Use DateFormat.format(Date)
   * @return A locale-dependent string representation.
   * @see #parse(String)
   * @see DateFormat
   */
  public String toLocaleString()
  {
    return java.text.DateFormat.getInstance().format(this);
  }

  /** 
   * <p>
   * Returns a string representation of this <code>Date</code>
   * object using GMT rather than the local timezone.
   * The following date format is used:
   * </p>
   * <p>
   * <code>d mon yyyy hh:mm:ss GMT</code>
   * </p>
   * <p>where the fields used here are:
   * <ul>
   * <li>
   * <code>d</code> -- the day of the month
   * as one or two decimal digits (1 to 31).
   * </li>
   * <li>
   * <code>mon</code> -- the month (Jan to Dec).
   * </li>
   * <li>
   * <code>yyyy</code> -- the year as four decimal digits.
   * </li>
   * <li>
   * <code>hh</code> -- the hour of the day
   * as two decimal digits in 24-hour clock notation
   * (01 to 23).
   * </li>
   * <li>
   * <code>mm</code> -- the minute of the day
   * as two decimal digits (01 to 59).
   * </li>
   * <li>
   * <code>ss</code> -- the second of the day
   * as two decimal digits (01 to 61).
   * </li>
   * <li>
   * <code>GMT</code> -- the literal string "GMT"
   * indicating Greenwich Mean Time as opposed to
   * the local timezone.
   * </li>
   * </ul>
   * 
   * @deprecated Use DateFormat.format(Date) with a GMT TimeZone.
   * @return A string of the form 'd mon yyyy hh:mm:ss GMT' using
   *         GMT as opposed to the local timezone.
   * @see #parse(String)
   * @see DateFormat
   */
  public String toGMTString()
  {
    java.text.DateFormat format = java.text.DateFormat.getInstance();
    format.setTimeZone(TimeZone.getTimeZone("GMT"));
    return format.format(this);
  }

  /**
   * Parses the time zone string.
   *
   * @param tok The token containing the time zone.
   * @param sign The sign (+ or -) used by the time zone.
   * @return An integer representing the number of minutes offset
   *         from GMT for the time zone.
   */
  private static int parseTz(String tok, char sign)
    throws IllegalArgumentException
  {
    int num;

    try
      {
	// parseInt doesn't handle '+' so strip off sign.
	num = Integer.parseInt(tok.substring(1));
      }
    catch (NumberFormatException ex)
      {
	throw new IllegalArgumentException(tok);
      }

    // Convert hours to minutes.
    if (num < 24)
      num *= 60;
    else
      num = (num / 100) * 60 + num % 100;

    return sign == '-' ? -num : num;
  }

  /**
   * Parses the month string.
   *
   * @param tok the token containing the month.
   * @return An integer between 0 and 11, representing
   *         a month from January (0) to December (11),
   *         or -1 if parsing failed.
   */
  private static int parseMonth(String tok)
  {
    // Initialize strings for month names.
    // We could possibly use the fields of DateFormatSymbols but that is
    // localized and thus might not match the English words specified.
    String months[] = { "JANUARY", "FEBRUARY", "MARCH", "APRIL", "MAY",
			"JUNE", "JULY", "AUGUST", "SEPTEMBER", "OCTOBER",
			"NOVEMBER", "DECEMBER" };

    int i;
    for (i = 0; i < 12; i++)
      if (months[i].startsWith(tok))
        return i;

    // Return -1 if not found.
    return -1;
  }

  /**
   * Parses the day of the week string.
   *
   * @param tok the token containing the day of the week.
   * @return true if the token was parsed successfully.
   */
  private static boolean parseDayOfWeek(String tok)
  {
    // Initialize strings for days of the week names.
    // We could possibly use the fields of DateFormatSymbols but that is
    // localized and thus might not match the English words specified.
    String daysOfWeek[] = { "SUNDAY", "MONDAY", "TUESDAY", "WEDNESDAY",
			    "THURSDAY", "FRIDAY", "SATURDAY" };

    int i;
    for (i = 0; i < 7; i++)
      if (daysOfWeek[i].startsWith(tok))
        return true;

    return false;
  }

  /** 
   * <p>
   * Parses a String and returns the time, in milliseconds since the
   * epoch, it represents.  Most syntaxes are handled, including
   * the IETF date standard "day, dd mon yyyy hh:mm:ss zz" (see
   * <code>toString()</code> for definitions of these fields).
   * Standard U.S. time zone abbreviations are recognised, in
   * addition to time zone offsets in positive or negative minutes.
   * If a time zone is specified, the specified time is assumed to
   * be in UTC and the appropriate conversion is applied, following
   * parsing, to convert this to the local time zone.  If no zone
   * is specified, the time is assumed to already be in the local
   * time zone.
   * </p>
   * <p>
   * The method parses the string progressively from left to right.
   * At the end of the parsing process, either a time is returned
   * or an <code>IllegalArgumentException</code> is thrown to signify
   * failure.  The ASCII characters A-Z, a-z, 0-9, and ',', '+', '-',
   * ':' and '/' are the only characters permitted within the string,
   * besides whitespace and characters enclosed within parantheses
   * '(' and ')'.  
   * </p>
   * <p>
   * A sequence of consecutive digits are recognised as a number,
   * and interpreted as follows:
   * <ul>
   * <li>
   * A number preceded by a sign (+ or -) is taken to be a time zone
   * offset.  The time zone offset can be specified in either hours
   * or minutes.  The former is assumed if the number is less than 24.
   * Otherwise, the offset is assumed to be in minutes.  A - indicates
   * a time zone west of GMT, while a + represents a time zone to the
   * east of GMT.  The time zones are always assumed to be relative
   * to GMT, and a (redundant) specification of this can be included
   * with the time zone.  For example, '-9', 'utc-9' and 'GMT-9' all
   * represent a time zone nine hours west of GMT.  Similarly,
   * '+4', 'ut+4' and 'UTC+4' all give 4 hours east of GMT.
   * </li>
   * <li>
   * A number equal to or greater than 70 is regarded as a year specification.
   * Values lower than 70 are only assumed to indicate a year if both the
   * day of the month and the month itself have already been recognised.
   * Year values less than 100 are interpreted as being relative to the current
   * century when the <code>Date</code> class is initialised..  Given a century,
   * x, the year is assumed to be within the range x - 80 to x + 19.  The value
   * itself is then used as a match against the two last digits of one of these
   * years.  For example, take x to be 2004.  A two-digit year is assumed to fall
   * within the range x - 80 (1924) and x + 19 (2023).  Thus, any intepreted value
   * between 0 and 23 is assumed to be 2000 to 2023 and values between 24 and 99
   * are taken as being 1924 to 1999.  This only applies for the case of 2004.
   * With a different year, the values will be interpreted differently. 2005
   * will used 0 to 24 as 2000 to 2024 and 25 to 99 as 1925 to 1999, for example.
   * This behaviour differs from that of <code>SimpleDateFormat</code> and is
   * time-dependent (a two-digit year will be interpreted differently depending
   * on the time the code is run).
   * </li>
   * <li>
   * Numbers followed by a colon are interpreted by first an hour, and then
   * as a minute, once an hour has been found.
   * </li>
   * <li>
   * <li>
   * Numbers followed by a slash are regarded first as a month, and then as
   * a day of the month once the month has been found.  This follows the
   * U.S. date format of mm/dd, rather than the European dd/mm.  Months
   * are converted to the recognised value - 1 before storage, in order
   * to put the number within the range 0 to 11.
   * </li>
   * <li>
   * Numbers followed by commas, whitespace, hyphens or the end of the string
   * are interpreted in the following order: hour, minute, second, day of month.
   * The first type not already recognised in the current string being parsed is
   * assumed.
   * </li>
   * </ul>
   * </p>
   * <p>
   * A sequence of consecutive alphabetic characters is recognised as a word,
   * and interpreted as follows, in a case-insentive fashion:
   * <ul>
   * <li>
   * The characters 'AM' or 'PM' restrict the hour value to a value between 0
   * and 12.  In the latter case, 12 is added to the hour value before storage.
   * </li>
   * <li>
   * Any words which match any prefix of one of the days of the week ('Monday',
   * 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday' and 'Sunday'),
   * are simply ignored.
   * </li>
   * <li>
   * Any words which match any prefix of one of the months of the year ('January',
   * 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September',
   * 'October', 'November', 'December') are recognised and interpreted as the
   * appropriate value between 0 and 11.  The first match made against a
   * month is the one used, in the order specified here.  For example, 'Ma' is
   * intepreted as 'March' (2) and not as 'May' (4).  Similarly, 'Ju' is 'June',
   * and not 'July'.
   * </li>
   * <li>
   * The words 'GMT', 'UT' and 'UTC' are interpreted as specifying UTC as the
   * time zone in use for this date.
   * </li>
   * <li>
   * The word pairs 'EST'/'EDT', 'CST'/'CDT', 'MST'/'MDT' and 'PST'/'PDT' are
   * interpreted as the appropriate U.S. time zone abbreviation.  Each pair
   * is the standard and daylight savings time zone specification, respectively,
   * for each zone within the U.S, these being Eastern Standard/Daylight Time
   * (-5), Central Standard/Daylight Time (-6), Mountain Standard/Daylight Time
   * (-7) and Pacific Standard/Daylight Time (-8).
   * </li>
   * </ul>
   *
   * @param string The String to parse.
   * @return The time in milliseconds since the epoch.
   * @throws IllegalArgumentException if the string fails to parse.
   * @deprecated Use DateFormat.parse(String)
   * @see #toString()
   * @see SimpleDateFormat
   */
  public static long parse(String string)
  {
    // Initialize date/time fields before parsing begins.
    int year = -1;
    int month = -1;
    int day = -1;
    int hour = -1;
    int minute = -1;
    int second = -1;
    int timezone = 0;
    boolean localTimezone = true;

    // Trim out any nested stuff in parentheses now to make parsing easier.
    StringBuffer buf = new StringBuffer();
    int parenNesting = 0;
    int len = string.length();
    for (int i = 0;  i < len;  i++)
      {
	char ch = string.charAt(i);
	if (ch >= 'a' && ch <= 'z')
	  ch -= 'a' - 'A';
	if (ch == '(')
	  parenNesting++;
	else if (parenNesting == 0)
	  buf.append(ch);
	else if (ch == ')')
	  parenNesting--;
      }
    int tmpMonth;

    // Make all chars upper case to simplify comparisons later.
    // Also ignore commas; treat them as delimiters.
    StringTokenizer strtok = new StringTokenizer(buf.toString(), " \t\n\r,");

    while (strtok.hasMoreTokens())
      {
	String tok = strtok.nextToken();
	char firstch = tok.charAt(0);
	if ((firstch == '+' || firstch == '-') && year >= 0)
	  {
	    timezone = parseTz(tok, firstch);
	    localTimezone = false;
	  }
	else if (firstch >= '0' && firstch <= '9')
	  {
	    while (tok != null && tok.length() > 0)
	      {
		int punctOffset = tok.length();
		int num = 0;
		int punct;
		for (int i = 0;  ;  i++)
		  {
		    if (i >= punctOffset)
		      {
			punct = -1;
			break;
		      }
		    else
		      {
			punct = tok.charAt(i);
			if (punct >= '0' && punct <= '9')
			  {
			    if (num > 999999999) // in case of overflow
			      throw new IllegalArgumentException(tok);
			    num = 10 * num + (punct - '0');
			  }
			else
			  {
			    punctOffset = i;
			    break;
			  }
		      }
		      
		  }

		if (punct == ':')
		  {
		    if (hour < 0)
		      hour = num;
		    else
		      minute = num;
		  }
	        else if ((num >= 70
			  && (punct == ' ' || punct == ','
			      || punct == '/' || punct < 0))
			 || (num < 70 && day >= 0 && month >= 0 && year < 0))
		  {
		    if (num >= 100)
		      year = num;
		    else
		      {
			int curYear = 1900 + new Date().getYear();
			int firstYear = curYear - 80;
			year = firstYear / 100 * 100 + num;
			if (year < firstYear)
			  year += 100;
		      }
		  }
		else if (punct == '/')
		  {
		    if (month < 0)
		      month = num - 1;
		    else
		      day = num;
		  }
		else if (hour >= 0 && minute < 0)
		  minute = num;
		else if (minute >= 0 && second < 0)
		  second = num;
		else if (day < 0)
		  day = num;
		else
		  throw new IllegalArgumentException(tok);

		// Advance string if there's more to process in this token.