/*
 * FormatCmd.java
 *
 * Copyright (c) 1997 Sun Microsystems, Inc.
 *
 * See the file "license.terms" for information on usage and
 * redistribution of this file, and for a DISCLAIMER OF ALL
 * WARRANTIES.
 * 
 * RCS: @(#) $Id: FormatCmd.java,v 1.1.1.1 1998/10/14 21:09:19 cvsadmin Exp $
 *
 */

package tcl.lang;
import java.util.*;

/**
 * This class implements the built-in "format" command in Tcl.
 */

class FormatCmd implements Command {

    private static final int LEFT_JUSTIFY  = 1;
    private static final int SHOW_SIGN     = 2;
    private static final int SPACE_OR_SIGN = 4;
    private static final int PAD_W_ZERO    = 8;
    private static final int ALT_OUTPUT    = 16;
    private static final int SIGNED_VALUE  = 32;
    private static final int RADIX   = 1;  /* Integer types.  %d, %x, %o */
    private static final int FLOAT   = 2;  /* Floating point.  %f */
    private static final int EXP     = 3;  /* Exponentional. %e and %E */
    private static final int GENERIC = 4;  /* Floating or exponential, 
					    * depending on exponent. %g */  
    
    /**
     * This procedure is invoked to process the "format" Tcl command.
     * See the user documentation for details on what it does.
     *
     * The first argument to the cmdProc is the formatString.  The cmdProc
     * simply copies all the chars into the sbuf until a '%' is found.  At
     * this point the cmdProc parces the formatString and determines the 
     * format parameters.  The parcing of the formatString can be broken into
     * six possible phases:
     * 
     * Phase 0 - Simply Print:            If the next char is %
     * Phase 1 - XPG3 Position Specifier: If the format [1-n]$ is used
     * Phase 2 - A Set of Flags:          One or more of the following + - 
     *                                           [space] 0 # 
     * Phase 3 - A Minimun Field Width    Either [integer] or *
     * Phase 4 - A Precision              If the format .[integer] or .*
     * Phase 5 - A Length Modifier        If h is present
     * Phase 6 - A Conversion Character   If one of the following is used
     *                                    d u i o x X c s f E g G
     * 
     * Any phase can skip ahead one or more phases, but are not allowed 
     * to move back to previous phases.  Once the parameters are determined
     * the cmdProc calls one of three private methods that returns a fully
     * formatted string.  This loop occurs for ever '%' in the formatString.
     */

    public void cmdProc(Interp interp, TclObject argv[])
            throws TclException {
	

	StringBuffer sbuf;     /* Stores the return value of the parced
			        * format string */  
	StrtoulResult stoul;   /* A return object to the strtoul call */
	char[]  format;        /* The format argument is converted to a char 
			        * array and manipulated as such */
	int     phase;         /* Stores the current phase of the parcing */
	int     width;         /* Minimum field width */
	int     precision;     /* Field precision from field specifier */
	int     fmtFlags;      /* Used to store the format flags ( #,+,etc) */
	int     argIndex;      /* Index of argument to substitute next. */
	int     fmtIndex;      /* Used to locate end of the format fields. */
	int     endIndex;      /* Used to locate end of numerical fields. */
	int     intValue;      /* Generic storage variable */
	long    lngValue;      /* Store the TclInteger.get() result */
	double  dblValue;      /* Store the TclDouble.get() result */
	boolean noPercent;     /* Special case for speed:  indicates there's
				* no field specifier, just a string to copy. */
	boolean xpgSet;        /* Indicates that xpg has been used for the 
				* particular format of the main while loop */
	boolean gotXpg;	       /* True means that an XPG3 %n$-style
				* specifier has been seen. */
	boolean gotSequential; /* True means that a regular sequential
				* (non-XPG3) conversion specifier has
				* been seen. */
	boolean useShort;      /* Value to be printed is short 
				* (half word). */
	boolean precisionSet;  /* Used for f, e, and E conversions */
	boolean cont;          /* Used for phase 3 */
	
	if (argv.length < 2) {
	    throw new TclNumArgsException(interp, 1, argv, 
		    "formatString ?arg arg ...?");
	}

	argIndex = 2;
	fmtIndex = 0;
	gotXpg   = gotSequential = false;
	format   = argv[1].toString().toCharArray();
	sbuf     = new StringBuffer();
	
       /*
	* So, what happens here is to scan the format string one % group
	* at a time, making many individual appends to the StringBuffer.
	*/

	while (fmtIndex < format.length) {
	    fmtFlags  = phase = width = 0;
	    noPercent = true;
	    xpgSet    = precisionSet = useShort = false;
	    precision = -1;

	    /*
	     * Append all characters to sbuf that are not used for the
	     * format specifier.
	     */

	    if (format[fmtIndex] != '%') {
	        int i;
	        for (i = fmtIndex; (i < format.length); i++) {
		    if (format[i] == '%'){
		        noPercent = false;
			break;
		    }
		}
		sbuf.append(new String(format, fmtIndex, i-fmtIndex));
		fmtIndex = i;
		if (noPercent) {
		    break;
		}
	    }
	    
	    /*
	     * If true, then a % has been indicated but we are at the end
	     * of the format string.  Call function to throw exception.
	     */

	    if (fmtIndex+1 >= format.length) {
	        errorEndMiddle(interp);
	    }

	    /*
	     * Phase 0:
	     * Check for %%.  If true then simply write a single '%'
	     * to the list.
	     */

	    checkOverFlow(interp, format, fmtIndex+1);
	    if (format[fmtIndex+1] == '%') {
		sbuf.append("%");
		fmtIndex += 2;
		/* 
		 * Re-enter the loop 
		 */

		continue;      
	    }

	    fmtIndex++;
	    checkOverFlow(interp, format, fmtIndex);
	    if (Character.isDigit(format[fmtIndex])) {
	        /*
		 * Parce the format array looking for the end of
		 * the number. 
		 */

		stoul = strtoul(format, fmtIndex);
	        intValue = (int)stoul.value;
		endIndex = stoul.index;
		
		if (format[endIndex] == '$') {
		    if (intValue == 0) {
		        errorBadIndex(interp, true);
		    }

		    /*
		     * Phase 1:
		     * Check for an XPG3-style %n$ specification.  
		     * Note: there must not be a mixture of XPG3
		     * specs and non-XPG3 specs in the same format string.
		     */

		    if (gotSequential) {
		        errorMixedXPG(interp);
		    }
		    gotXpg   = true;
		    xpgSet   = true;
		    phase    = 2;
		    fmtIndex = endIndex+1;
		    argIndex = intValue+1;
		    if ((argIndex < 2) || (argIndex >= argv.length)) {
		        errorBadIndex(interp, gotXpg);
		    }
			
		} else {
		    /*
		     * Phase 3:
		     * Format jumped straight to phase 3; Setting 
		     * width field.  Again, verify that all format
		     * specifiers are sequential.
		     */		    

		    if (gotXpg) {
		        errorMixedXPG(interp);
		    }
		    gotSequential = true;
		    if (format[fmtIndex] != '0') {
		        fmtIndex = endIndex;
			width = intValue;
			phase = 4;
		    }
		}
	    } else {
	        if (gotXpg) {
		    errorMixedXPG(interp);
		}
		gotSequential = true;
	    }

	    /*
	     * Phase 2:
	     * Setting the Format Flags.  At this point the phase value
	     * can be either zero or three.  Anything greater is an
	     * incorrect format.
	     */

	    if (phase < 3) {
	        checkOverFlow(interp, format, fmtIndex);
	        char ch = format[fmtIndex];
		cont = true;
		while (cont) {
		    switch (ch) {
		        case '-': {
			    fmtFlags |= LEFT_JUSTIFY;
			    break;
			}
		        case '#': {
		            fmtFlags |= ALT_OUTPUT;
			    break;
		        }
		        case '0': {
		            fmtFlags |= PAD_W_ZERO;
			    break;
		        }
		        case ' ': {
		            fmtFlags |= SPACE_OR_SIGN;
			    break;
		        }
		        case '+': {
		            fmtFlags |= SHOW_SIGN;
			    break;
		        }
		        default: {
			    cont = false;
			}
		    }
		    if (cont) {
		        fmtIndex++;
			checkOverFlow(interp, format, fmtIndex);
			ch = format[fmtIndex];
		    }		
		}
		phase = 3;
	    }

	    /*
	     * Phase 3:
	     * Setting width field.  Partially redundant code from the
	     * Phase 1 if/else statement, but this is made to run fast.
	     */

	    checkOverFlow(interp, format, fmtIndex);
	    if (Character.isDigit(format[fmtIndex])) {
		stoul = strtoul(format, fmtIndex);
	        width = (int)stoul.value;
		fmtIndex = stoul.index;

	    } else if (format[fmtIndex] == '*') {
	        if (argv.length > argIndex) {
	            width = TclInteger.get(interp, argv[argIndex]);
		    argIndex++;
		    fmtIndex++;
		}
	    }

	    /*
	     * Phase 4:
	     * Setting the precision field.
	     */

	    checkOverFlow(interp, format, fmtIndex);
	    if (format[fmtIndex] == '.') {
	        fmtIndex++;
		checkOverFlow(interp, format, fmtIndex);
		if (Character.isDigit(format[fmtIndex])) {
		    precisionSet = true;

		    stoul = strtoul(format, fmtIndex);
		    precision = (int)stoul.value;
		    fmtIndex = stoul.index;
		} else if (format[fmtIndex] == '*') {
		    if (argv.length > argIndex) {
		        precisionSet = true;
			precision = TclInteger.get(interp, argv[argIndex]);
			argIndex++;
			fmtIndex++;
			checkOverFlow(interp, format, fmtIndex);
		    }
		} else {
 		  /*
		   * Format field had a '.' without an integer or '*'
		   * preceeding it  (eg  %2.d  or %2.-5d)
		   */

		  errorBadField(interp, format[fmtIndex]);
		}
	    }
		
	    /*
	     * Phase 5:
	     * Setting the length modifier.
	     */

	    if (format[fmtIndex] == 'h') { 
	        fmtIndex++;
		checkOverFlow(interp, format, fmtIndex);
	        useShort = true;
	    } else if (format[fmtIndex] == 'l') { 
	        fmtIndex++;
		checkOverFlow(interp, format, fmtIndex);

		/*
		 * 'l' is ignored, but should still be processed.
		 */
	    }

	    if ((argIndex < 2) || (argIndex >= argv.length)) {
	        errorBadIndex(interp, gotXpg);
	    }

	    /*
	     * Phase 6:
	     * Setting conversion field.
	     * At this point, variables are initialized as follows:
	     *
	     * width               The specified field width.  This is always
	     *                         non-negative.  Zero is the default.
	     * precision           The specified precision.  The default
	     *                         is -1.
	     * argIndex            The argument index from the argv array 
	     *                         for the appropriate arg.
	     * fmtFlags            The format flags are set via bitwise 
	     *                         operations.  Below are the bits
	     *                         and their meanings.

	     *     ALT_OUTPUT          set if a '#' is present.
	     *     SHOW_SIGN           set if a '+' is present.
	     *     SPACE_OR_SIGN       set if a ' ' is present.
	     *     LEFT_JUSTIFY        set if a '-' is present or if the
	     *                           field width was negative.
	     *     PAD_W_ZERO          set if a '0' is present
	     */

	    String strValue = "";
	    char index = format[fmtIndex];

 	    switch (index) {
 	        case 'u': 
 	        case 'd': 
 	        case 'o': 
 	        case 'x': 
 	        case 'X': 
 	        case 'i': {
		    if (index == 'u') {
		        /*
			 * Since Java dosent provide unsigned ints we need to 
			 * make our own.  If the value is negative we need to 
			 * clear out all of the leading bits from the 33rd bit
			 * and on.  The result is a long value equal to that
			 * of an unsigned int.
			 */

		        lngValue = (long)TclInteger.get(interp, 
				 argv[argIndex]);
			if (lngValue < 0) {
			    lngValue = (lngValue << 32);
			    lngValue = (lngValue >>> 32);
			}
		    } else {
		        fmtFlags |= SIGNED_VALUE;
		        lngValue = (long)TclInteger.get(interp, 
				  argv[argIndex]);
		    }

		    /*
		     * If the useShort option has been selected, we need
		     * to clear all but the first 16 bits.
		     */

		    if (useShort) {
		        lngValue = (lngValue << 48);
		        lngValue = (lngValue >> 48);
		    }

		    if ( index == 'o' ) {
		        sbuf.append(cvtLngToStr(lngValue, width, precision, 
                                fmtFlags, 8,"01234567".toCharArray() , "0"));
		    } else if ( index == 'x' ) {
		        sbuf.append(cvtLngToStr(lngValue, width, precision, 
                                fmtFlags, 16,"0123456789abcdef".toCharArray(),
                                "0x"));
		    } else if ( index == 'X' ) {
		        sbuf.append(cvtLngToStr(lngValue, width, precision, 
                                fmtFlags, 16,"0123456789ABCDEF".toCharArray(),
				"0X"));
		    } else {
		        sbuf.append(cvtLngToStr(lngValue, width, precision, 
                                fmtFlags, 10,"0123456789".toCharArray() , ""));
		    }
 		    break;		  
 		}
 	        case 'c': {
 		    intValue = 0;
 		    char arr[] = {(char)TclInteger.get(interp, 
                            argv[argIndex])};
 		    strValue = new String(arr);
 		    sbuf.append(cvtStrToStr(strValue, width, precision, 
			    fmtFlags));
 		    break;		  
 		}
 	        case 's': {
 		    strValue = argv[argIndex].toString();
 		    sbuf.append(cvtStrToStr(strValue, width, precision, 
			    fmtFlags));
 		    break;		  
 		}
	        case 'f': {
	            dblValue = TclDouble.get(interp, argv[argIndex]);
		    sbuf.append(cvtDblToStr(dblValue, width, precision, 
                                fmtFlags, 10,"0123456789".toCharArray() , 
				"", FLOAT));
		    break;		  
		}
	        case 'e': {
 		    dblValue = TclDouble.get(interp, argv[argIndex]);
		    sbuf.append(cvtDblToStr(dblValue, width, precision, 
                                fmtFlags, 10,"e".toCharArray() , 
				"", EXP));
		    break;		  
		}
 	        case 'E': {
 		    dblValue = TclDouble.get(interp, argv[argIndex]);
		    sbuf.append(cvtDblToStr(dblValue, width, precision, 
                                fmtFlags, 10,"E".toCharArray() , 
				"", EXP));
		    break;		  
		}
 	        case 'g': {
 		    dblValue = TclDouble.get(interp, argv[argIndex]);
		    sbuf.append(cvtDblToStr(dblValue, width, precision, 
                                fmtFlags, 10,"e".toCharArray() , 
				"", GENERIC));
		    break;		  
		}
 	        case 'G': {
 		    dblValue = TclDouble.get(interp, argv[argIndex]);
		    sbuf.append(cvtDblToStr(dblValue, width, precision, 
                                fmtFlags, 10,"E".toCharArray() , 
				"", GENERIC));	
		    break;		  
		}
 	        default: {
 		    errorBadField(interp, format[fmtIndex]);
 		}
 	    }
 	    fmtIndex++;
	    argIndex++;
   	}
 	interp.setResult(sbuf.toString());
    }


    /**
     * This procedure is invoked in "phase 6" od the Format cmdProc.  It 
     * converts the lngValue to a string with a specified format determined by
     * the other input variables.
     * @param lngValue  - Is the value of the argument input
     * @param width     - The minimum width of the string.
     * @param precision - The minimum width if the integer.  If the string len 
     *                        is less than precision, leading 0 are appended.
     * @param flags     - Specifies various formatting to the string 
     *                        representation (-, +, space, 0, #)
     * @param base      - The base of the integer (8, 10, 16)
     * @param charSet   - The char set to use for the conversion to ascii OR
     *                        The char used for sci notation.
     * @param altPrefix - If not empty, str to append on the beginnig of the 
     *                        resulting string  (eg 0 or 0x or 0X ).  
     * @return String representation of the long.
     */

    private String cvtLngToStr(long lngValue, int width, 
            int precision, int flags, int base, char[] charSet, 
            String altPrefix) {
        StringBuffer sbuf   = new StringBuffer(100);
        StringBuffer tmpBuf = new StringBuffer(0).append("");

	int  i;
	int  length;
	int  nspace;
	int  prefixSize = 0;
	char prefix     = 0;
	
	/*
	 * For the format %#x, the value zero is printed "0" not "0x0".
	 * I think this is stupid. 
	 */

	if (lngValue==0) {
	    flags = (flags | ALT_OUTPUT);
	}

       
	if ((flags & SIGNED_VALUE) != 0) {
            if (lngValue < 0){
	        if (altPrefix.length() > 0) {
		    lngValue = (lngValue << 32);
		    lngValue = (lngValue >>> 32);
		} else {        
		    lngValue = -lngValue;
		    prefix = '-';
		    prefixSize = 1;
		}
	    } else if ((flags & SHOW_SIGN) != 0) {
	        prefix = '+';
		prefixSize = 1;
	    } else if ((flags & SPACE_OR_SIGN) != 0) {
	        prefix = ' ';
		prefixSize = 1;
	    }
	}

        if (((PAD_W_ZERO & flags) != 0) && (precision<width-prefixSize) ){
            precision = width-prefixSize;
	}

	/* 
	 * Convert to ascii 
	 */

	do{                                           
	    sbuf.insert(0, charSet[(int)(lngValue%base)]);
            lngValue = lngValue/base;
	} while (lngValue > 0);
       
        length = sbuf.length();
	for (i = (precision-length); i > 0; i--) {
            sbuf.insert(0, '0');
	}
        if (prefix != 0) {
            sbuf.insert(0, prefix);
	}
        if ((flags & ALT_OUTPUT) != 0) {
            if ((altPrefix.length() > 0)
                    && (sbuf.charAt(0) != altPrefix.charAt(0))) {
		sbuf.insert(0, altPrefix);
	    }
        }

	/*
	 * The text of the conversion is pointed to by "bufpt" and is
	 * "length" characters long.  The field width is "width".  Do
	 * the output.
	 */

        nspace = width-sbuf.length();
	if (nspace > 0) {
	    tmpBuf = new StringBuffer(nspace);
	    for (i = 0; i < nspace; i++) {
	        tmpBuf.append(" ");
	    }
	}

	if ((LEFT_JUSTIFY & flags) != 0) {    
	    /* 
	     * left justified 
	     */

	    return sbuf.toString() + tmpBuf.toString();
	} else {                              
	    /* 
	     * right justified 
	     */

	    return tmpBuf.toString() + sbuf.toString() ;
	}
   }

    static String toString(double dblValue, int precision, int base) {
	return cvtDblToStr(dblValue, 0, precision, 0, base, 
		"e".toCharArray(),  null, GENERIC);
    } 

    /**
     * This procedure is invoked in "phase 6" od the Format cmdProc.  It 
     * converts the lngValue to a string with a specified format determined 
     * by the other input variables.
     * @param dblValue  - Is the value of the argument input
     * @param width     - The minimum width of the string.
     * @param precision - The minimum width if the integer.  If the string len 
     *                        is less than precision, leading 0 are appended.
     * @param flags     - Specifies various formatting to the string 
     *                        representation (-, +, space, 0, #)
     * @param base      - The base of the integer (8, 10, 16)
     * @param charSet   - The char set to use for the conversion to ascii OR
     *                        The char used for sci notation.
     * @param altPrefix - If not empty, str to append on the beginnig of the 
     *                        resulting string  (eg 0 or 0x or 0X ).  
     * @param xtype     - Either FLOAT, EXP, or GENERIC depending on the 
     *                        format specifier.
     * @return String representation of the long.
     */

    private static String cvtDblToStr(double dblValue, int width, 
            int precision, int flags, int base, char[] charSet, 
            String altPrefix, int xtype) {
        StringBuffer sbuf = new StringBuffer(100);
	int     i;
	int     exp;
	int     length;
	int     count;
	int     digit;
	int     prefixSize = 0;
	char    prefix     = 0;
	double  rounder;
	boolean flag_exp   = false;   /* Flag for exponential representation */
	boolean flag_rtz   = true;    /* Flag for "remove trailing zeros" */
	boolean flag_dp    = true;    /* Flag for remove "decimal point" */
      
	/* 
	 * If precision < 0 (eg -1) then the precision defaults
	 */

	if (precision < 0) {
	    precision = 6;         
	}

	if (dblValue < 0.0) {
	    dblValue = -dblValue;
	    prefix = '-';
	    prefixSize = 1;
	} else if ((flags & SHOW_SIGN) != 0) {
	    prefix = '+';
	    prefixSize = 1;
	} else if ((flags & SPACE_OR_SIGN) != 0) {
	    prefix = ' ';
	    prefixSize = 1;
	}

	/* 
	 * For GENERIC xtypes the precision includes the ones digit
	 * so just decrement to get the correct precision.
	 */

	if (xtype==GENERIC && precision > 0) {
	    precision--;
	}

	/* 
	 * Rounding works like BSD when the constant 0.4999 is used.  Wierd! 
	 */

	for (i = precision, rounder = 0.4999; i > 0; i--, rounder*=0.1);

	if (xtype == FLOAT) { 
	    dblValue += rounder;
	}

	/* 
	 * Normalize dblValue to within 10.0 > dblValue >= 1.0 
	 */

	exp = 0;
	if (dblValue>0.0) {
	    int k = 0;
	    while ((dblValue >= 1e8) && (k++ < 100)) {
	        dblValue *= 1e-8; 
		exp+=8; 
	    }
	    while ((dblValue >= 10.0) && (k++ < 100)) {
	        dblValue *= 0.1; 
		exp++; 
	    }
	    while ((dblValue < 1e-8) && (k++ < 100)) {
	        dblValue *= 1e8; 
		exp-=8; 
	    }
	    while ((dblValue < 1.0) && (k++ < 100)) {
	        dblValue *= 10.0; 
		exp--; 
	    }
	    if (k >= 100){
	        return "NaN";
	    }
	}

        /*
	 * If the field type is GENERIC, then convert to either EXP
	 * or FLOAT, as appropriate.
	 */

        flag_exp = xtype==EXP;
        if (xtype!=FLOAT) {
            dblValue += rounder;
	    if (dblValue>=10.0) {
	        dblValue *= 0.1; 
		exp++; 
	    }
        }
        if (xtype==GENERIC) {
            flag_rtz = !((flags & ALT_OUTPUT) !=0);
            if ((exp < -4) || (exp > precision)) {
                xtype = EXP;
	    }else{
                precision = (precision - exp);
		xtype = FLOAT;
	    }
	} else {
            flag_rtz = false;
	}

        /*
	 * The "exp+precision" test causes output to be of type EXP if
	 * the precision is too large to fit in buf[].
	 */

        count = 0;
        if (xtype == FLOAT) {
            flag_dp = ((precision > 0) || ((flags & ALT_OUTPUT) != 0));
	    if (prefixSize > 0) {
	        /* 
		 * Sign 
		 */

	        sbuf.append(prefix);         
	    }
	    if (exp < 0) {
	        /* 
		 * Digits before "."
		 */

	        sbuf.append('0');            
	    } 
	    for (  ; exp>=0; exp--) {
	        if (count++ >= 16) {
		    sbuf.append('0');
		} else {
		    digit = (int)dblValue;
		    dblValue = (dblValue - digit)*10.0;
		    sbuf.append(digit);
		}
	    }
	    if (flag_dp) {
	        sbuf.append('.');            
	    }
	    for (exp++; (exp < 0) && (precision > 0); precision--, exp++) {
	        sbuf.append('0');
	    }
	    while ((precision--) > 0) {
	        if (count++ >= 16) {
		    sbuf.append('0');
		} else {
		    digit = (int)dblValue;
		    dblValue = (dblValue - digit)*10.0;
		    sbuf.append(digit);
		}
	    }
	    
	    if (flag_rtz && flag_dp) {
	        /* 
		 * Remove trailing zeros and "." 
		 */

	        int len, index;
		len = index = 0;
		for (len = (sbuf.length() - 1), index = 0;
		        (len >= 0) && (sbuf.charAt(len) == '0');
		        len--, index++);

		if ((len >= 0) && (sbuf.charAt(len)=='.')) {
		    index++;
		}
		
		if (index > 0) {
		    sbuf = new StringBuffer(sbuf.toString().substring(0,
                            sbuf.length()-index));
		}
	    }
	} else {    
	    /* 
	     * EXP or GENERIC 
	     */

            flag_dp = ((precision > 0) || ((flags & ALT_OUTPUT) != 0));

	    if (prefixSize > 0) {
	        sbuf.append(prefix);   
	    }
	    digit = (int)dblValue;
	    dblValue = (dblValue - digit)*10.0;
	    sbuf.append(digit);        
	    if (flag_dp) {
	        sbuf.append('.');
	    }
	    while (precision-- > 0) {
	        if (count++ >= 16) {
		    sbuf.append('0');
		} else {
		    digit = (int)dblValue;
		    dblValue = (dblValue - digit)*10.0;
		    sbuf.append(digit);
		}
	    }

	    if (flag_rtz && flag_dp) {
	        /* 
		 * Remove trailing zeros and "." 
		 */

		for (i = 0, length = (sbuf.length() - 1);
		        (length >= 0) && (sbuf.charAt(length) == '0');
		        length--, i++);

		if ((length >= 0) && (sbuf.charAt(length)=='.')) {
		    i++;
		}
		    
		if (i > 0) {
		    sbuf = new StringBuffer(sbuf.toString().substring(0,
                            sbuf.length()-i));
		}
	    }
	    if ((exp != 0) || flag_exp) {
	        sbuf.append(charSet[0]);
		if (exp < 0) {
		   sbuf.append('-');
		   exp = -exp;
		} else {	
		   sbuf.append('+');
		}
		if (exp>=100) {
		    sbuf.append((exp/100));
		    exp %= 100;
		}
		sbuf.append(exp / 10);
		sbuf.append(exp % 10);
	    }
	}

        /* 
	 * The converted number is in sbuf. Output it.
	 * Note that the number is in the usual order, not reversed as with
	 * integer conversions. 
	 */

	length = sbuf.length();

	/* 
	 * Special case:  Add leading zeros if the PAD_W_ZERO flag is
	 * set and we are not left justified 
	 */

        if (((PAD_W_ZERO & flags) != 0) && ((LEFT_JUSTIFY & flags) == 0)) {
	    int nPad = width - length;
            i = prefixSize;
	    while ((nPad--) > 0) {
	        sbuf.insert(prefixSize,'0');
	    }
	    length = width;
        }

	/*
	 * Count the number of spaces remaining and creat a StringBuffer
	 * (tmpBuf) with the correct number of spaces.
	 */

	int nspace = width-length;
	StringBuffer tmpBuf = new StringBuffer(0).append("");
	if (nspace > 0) {
	    tmpBuf = new StringBuffer(nspace);
	    for (i=0; i < nspace; i++) {
	        tmpBuf.append(" ");
	    }
	}

	if ((LEFT_JUSTIFY & flags) != 0) {    
	    /* 
	     * left justified 
	     */

	    return sbuf.toString() + tmpBuf.toString();
	} else {                              
	    /* 
	     * right justified 
	     */

	    return tmpBuf.toString() + sbuf.toString() ;
	}

    }

    /**
     * This procedure is invoked in "phase 6" od the Format cmdProc.  It 
     * converts the strValue to a string with a specified format determined 
     * by the other input variables.
     * @param strValue  - Is the String w/o formatting.
     * @param width     - The minimum width of the string.
     * @param precision - The minimum width if the integer.  If the string 
     *                        len is less than precision, leading 0 are 
     *                        appended.
     * @param flags     - Specifies various formatting to the string 
     *                        representation (-, +, space, 0, #)
     * @return String representation of the integer.
     */

    private static String cvtStrToStr(String strValue, int width, 
            int precision, int flags) {
        String left  = "";
	String right = "";
	
	if (precision < 0) {
	    precision = 0;
	}
	
	if ((precision != 0) && (precision < strValue.length())) {
	    strValue = strValue.substring(0, precision);
	}
	
	if (width > strValue.length()) {
	    StringBuffer sbuf = new StringBuffer();
	    int index = (width - strValue.length());
	    for (int i = 0; i < index; i++) {
	        sbuf.append(' ');
	    }
	    if ((LEFT_JUSTIFY & flags) != 0) {
	        right = sbuf.toString();
	    } else {
	        left = sbuf.toString();
	    }
	}
	
	return(left + strValue + right );
    }


    /**
     * Search through the array while the current char is a digit.  When end 
     * of array occurs or the char is not a digit, stop the loop, convert the
     * sub-array into a long.  At this point return a StrtoulResult object
     * that contains the new long value and the current pointer to the array.
     *
     * @param arr - the array that contains a string representation of an int.
     * @param endIndex - the arr index where the numeric value begins.
     * @return StrtoResult containing the value and new index/
     */

    private StrtoulResult strtoul(char[] arr, int endIndex) {
        int curIndex; 
	int orgIndex;

	orgIndex = endIndex;
	for ( ; endIndex < arr.length; endIndex++) {
	    if (!Character.isDigit(arr[endIndex])) {
	        break;
	    }
	}
	return (new StrtoulResult(Long.parseLong(new String(arr, orgIndex, 
                endIndex-orgIndex)), endIndex, 0));
    }


    /*
     *
     *  Error routines:
     *
     */


    /**
     * Called whenever the fmtIndex in the cmdProc is changed.  It verifies
     * the the array index is still within the bounds of the array.  If no
     * throw error.
     * @param interp  - The TclInterp which called the cmdProc method .
     * @param arr     - The array to be checked.
     * @param index   - The new value for the array index.
     */

    private static void checkOverFlow(Interp interp, char[] arr, int index) 
            throws TclException {      
        if ((index >= arr.length) || (index < 0)) {
	    throw new TclException(interp,
                    "\"%n$\" argument index out of range");
	}    
    }


    /**
     * Called whenever Sequential format specifiers are interlaced with 
     * XPG format specifiers in one call to cmdProc.
     *
     * @param interp  - The TclInterp which called the cmdProc method .
     */

    private static void errorMixedXPG(Interp interp) 
            throws TclException {      
        throw new TclException(interp,
	        "cannot mix \"%\" and \"%n$\" conversion specifiers");
    }

    
    /**
     * Called whenever the argIndex access outside the argv array.  If the
     * type is an XPG then the error message is different.
     *
     * @param interp  - The TclInterp which called the cmdProc method .
     * @param gotXpg  - Boolean the determines if the current format is of a
     *                      XPG type or Sequential
     */

    private static void errorBadIndex(Interp interp, boolean gotXpg) 
            throws TclException {      
        if (gotXpg) {
	    throw new TclException(interp,
                    "\"%n$\" argument index out of range");
	} else {
	    throw new TclException(interp,
		    "not enough arguments for all format specifiers");
	}
    }


    /**
     * Called whenever the current char in the format array is erroneous
     *
     * @param interp  - The TclInterp which called the cmdProc method .
     * @param fieldSpecifier  - The erroneous character
     */

    private static void errorBadField(Interp interp, char fieldSpecifier) 
            throws TclException {      
        throw new TclException(interp, "bad field specifier \""
	        + fieldSpecifier + "\"");
    }


    /**
     * Called whenever the a '%' is found but then the format string ends.
     *
     * @param interp  - The TclInterp which called the cmdProc method .
     */

    private static void errorEndMiddle(Interp interp) 
            throws TclException {      
        throw new TclException(interp, 
                "format string ended in middle of field specifier");
    }

}