diff --git a/WEGA/contrib/ingres/dbs/src/gutil/ftoa.c b/WEGA/contrib/ingres/dbs/src/gutil/ftoa.c
new file mode 100644
index 0000000..dacca78
--- /dev/null
+++ b/WEGA/contrib/ingres/dbs/src/gutil/ftoa.c
@@ -0,0 +1,187 @@
+# define	MAXDIG		25
+
+/*
+**  FLOATING POINT TO ASCII CONVERSION
+**
+**	'Value' is converted to an ascii character string and stored
+**	into 'ascii'.  Ascii should have room for at least 'width' + 1
+**	characters.  'Width' is the width of the output field (max).
+**	'Prec' is the number of characters to put after the decimal
+**	point.  The format of the output string is controlled by
+**	'format'.
+**
+**	'Format' can be:
+**		e or E: "E" format output
+**		f or F:  "F" format output
+**		g or G:  "F" format output if it will fit, otherwise
+**			use "E" format.
+**		n or N:  same as G, but decimal points will not always
+**			be aligned.
+**
+**	If 'format' is upper case, the "E" comes out in upper case;
+**	otherwise it comes out in lower case.
+**
+**	When the field width is not big enough, it fills the field with
+**	stars ("*****") and returns zero.  Normal return is the width
+**	of the output field (sometimes shorter than 'width').
+*/
+AA_ftoa(value, ascii, width, prec, format)
+double	value;
+char	*ascii;
+int	width;
+int	prec;
+char	format;
+{
+	register int	avail;
+	register char	*a;
+	register char	*p;
+	register int	lowercase;
+	int		expon;
+	int		sign;
+	extern char	*fcvt();
+	extern char	*ecvt();
+
+# ifdef EBCDIC
+	if (format < 'A')
+# else
+	if (format >= 'a')
+# endif
+		format -= lowercase = 'a' - 'A';
+	else
+		lowercase = 0;
+
+	if (format != 'E')
+	{
+		/* try 'F' style output */
+		p = fcvt(value, prec, &expon, &sign);
+		avail = width;
+		a = ascii;
+
+		/* output sign */
+		if (sign)
+		{
+			avail--;
+			*a++ = '-';
+		}
+
+		/* output '0' before the decimal point */
+		if (expon <= 0)
+		{
+			*a++ = '0';
+			avail--;
+		}
+
+		/* compute space length left after dec pt and fraction */
+		if (prec > 0)
+			avail -= prec + 1;
+		if (format == 'G')
+			avail -= 4;
+
+		if (avail >= expon)
+		{
+
+			/* it fits.  output */
+			while (expon > 0)
+			{
+				/* output left of dp */
+				expon--;
+				if (*p)
+				{
+					*a++ = *p++;
+				}
+				else
+					*a++ = '0';
+			}
+
+			/* output fraction (right of dec pt) */
+			avail = expon;
+			goto frac_out;
+		}
+		/* won't fit; let's hope for G format */
+	}
+
+	if (format != 'F')
+	{
+		/* try to do E style output */
+		p = ecvt(value, prec + 1, &expon, &sign);
+		avail = width - 5;
+		a = ascii;
+
+		/* output the sign */
+		if (sign)
+		{
+			*a++ = '-';
+			avail--;
+		}
+	}
+
+	/* check for field too small */
+	if (format == 'F' || avail < prec)
+	{
+		/* sorry joker, you lose */
+		a = ascii;
+		for (avail = width; avail > 0; avail--)
+			*a++ = '*';
+		*a = 0;
+		return (0);
+	}
+
+	/* it fits; output the number */
+	format = 'E';
+
+	/* output the LHS single digit */
+	*a++ = *p++;
+	expon--;
+
+	/* output the rhs */
+	avail = 1;
+
+  frac_out:
+	if (prec > 0)
+		*a++ = '.';
+	while (prec > 0)
+	{
+		prec--;
+		if (avail < 0)
+		{
+			avail++;
+			*a++ = '0';
+		}
+		else
+		{
+			if (*p)
+				*a++ = *p++;
+			else
+				*a++ = '0';
+		}
+	}
+
+	/* output the exponent */
+	if (format == 'E')
+	{
+		*a++ = 'E' + lowercase;
+		if (expon < 0)
+		{
+			*a++ = '-';
+			expon = -expon;
+		}
+		else
+			*a++ = '+';
+		*a++ = (expon / 10) % 10 + '0';
+		*a++ = expon % 10 + '0';
+	}
+
+	/* output spaces on the end in G format */
+	if (format == 'G')
+	{
+		*a++ = ' ';
+		*a++ = ' ';
+		*a++ = ' ';
+		*a++ = ' ';
+	}
+
+	/* finally, we can return */
+	*a = 0;
+	avail = a - ascii;
+	return (avail);
+}