Nan and Infinity can be negative

Decimal point was missing for some exponential output
Only divide sign by 10 once per loop
Add a missing NUL terminator for one case
If there is no whole number portion, print a leading zero (e.g. 0.1 not .1)

src/lib/libc/stdlib/gcvt.c

@@ -1,4 +1,4 @@
-/*	$OpenBSD: gcvt.c,v 1.6 2005/08/08 08:05:36 espie Exp $	*/
+/*	$OpenBSD: gcvt.c,v 1.7 2006/01/10 02:08:28 millert Exp $	*/
 
 /*
  * Copyright (c) 2002, 2003 Todd C. Miller <Todd.Miller@courtesan.com>
@@ -40,7 +40,7 @@ gcvt(double value, int ndigit, char *buf)
 	digits = __dtoa(value, 2, ndigit, &decpt, &sign, NULL);
 	if (decpt == 9999) {
 		/* Infinity or NaN, assume buffer is at least ndigit long. */
-		strlcpy(buf, digits, ndigit + 1);
+		snprintf(buf, ndigit + 1, "%s%s", sign ? "-" : "", digits);
 		return (buf);
 	}
 
@@ -49,13 +49,16 @@ gcvt(double value, int ndigit, char *buf)
 		*dst++ = '-';
 
 	if (decpt < 0 || decpt > ndigit) {
-		/* exponential format */
+		/* exponential format (e.g. 1.2345e+13) */
 		if (--decpt < 0) {
 			sign = 1;
 			decpt = -decpt;
 		} else
 			sign = 0;
-		for (src = digits; *src != '\0'; )
+		src = digits;
+		*dst++ = *src++;
+		*dst++ = '.';		/* XXX - locale-specific */
+		while (*src != '\0')
 			*dst++ = *src++;
 		*dst++ = 'e';
 		if (sign)
@@ -69,7 +72,8 @@ gcvt(double value, int ndigit, char *buf)
 		} else {
 			/* XXX - optimize */
 			for (sign = decpt, i = 0; (sign /= 10) != 0; i++)
-				sign /= 10;
+				continue;
+			dst[i + 1] = '\0';
 			while (decpt != 0) {
 				dst[i--] = '0' + decpt % 10;
 				decpt /= 10;
@@ -84,7 +88,9 @@ gcvt(double value, int ndigit, char *buf)
 				*dst++ = '0';
 		}
 		if (*src != '\0') {
-			*dst++ = '.';		/* XXX - locale-specific (LC_NUMERIC) */
+			if (src == digits)
+				*dst++ = '0';	/* zero before decimal point */
+			*dst++ = '.';		/* XXX - locale-specific */
 			for (i = decpt; digits[i] != '\0'; i++) {
 				*dst++ = digits[i];
 			}