Reformated comments and long lines
[mmh] / sbr / snprintf.c
index efb6df0..1dd4a65 100644 (file)
@@ -1,70 +1,70 @@
 /*
- * snprintf.c -- formatted output to a string
- *
- * This is an implementation of snprintf() and vsnprintf()
- * taken from the Apache web server.  This is only used on
- * systems which do not have a native version.
- */
+** snprintf.c -- formatted output to a string
+**
+** This is an implementation of snprintf() and vsnprintf()
+** taken from the Apache web server.  This is only used on
+** systems which do not have a native version.
+*/
 
 /* ====================================================================
- * Copyright (c) 1995-1999 The Apache Group.  All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer. 
- *
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *
- * 3. All advertising materials mentioning features or use of this
- *    software must display the following acknowledgment:
- *    "This product includes software developed by the Apache Group
- *    for use in the Apache HTTP server project (http://www.apache.org/)."
- *
- * 4. The names "Apache Server" and "Apache Group" must not be used to
- *    endorse or promote products derived from this software without
- *    prior written permission. For written permission, please contact
- *    apache@apache.org.
- *
- * 5. Products derived from this software may not be called "Apache"
- *    nor may "Apache" appear in their names without prior written
- *    permission of the Apache Group.
- *
- * 6. Redistributions of any form whatsoever must retain the following
- *    acknowledgment:
- *    "This product includes software developed by the Apache Group
- *    for use in the Apache HTTP server project (http://www.apache.org/)."
- *
- * THIS SOFTWARE IS PROVIDED BY THE APACHE GROUP ``AS IS'' AND ANY
- * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
- * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE APACHE GROUP OR
- * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
- * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
- * OF THE POSSIBILITY OF SUCH DAMAGE.
- * ====================================================================
- *
- * This software consists of voluntary contributions made by many
- * individuals on behalf of the Apache Group and was originally based
- * on public domain software written at the National Center for
- * Supercomputing Applications, University of Illinois, Urbana-Champaign.
- * For more information on the Apache Group and the Apache HTTP server
- * project, please see <http://www.apache.org/>.
- *
- * This code is based on, and used with the permission of, the
- * SIO stdio-replacement strx_* functions by Panos Tsirigotis
- * <panos@alumni.cs.colorado.edu> for xinetd.
- */
+** Copyright (c) 1995-1999 The Apache Group.  All rights reserved.
+**
+** Redistribution and use in source and binary forms, with or without
+** modification, are permitted provided that the following conditions
+** are met:
+**
+** 1. Redistributions of source code must retain the above copyright
+**    notice, this list of conditions and the following disclaimer.
+**
+** 2. Redistributions in binary form must reproduce the above copyright
+**    notice, this list of conditions and the following disclaimer in
+**    the documentation and/or other materials provided with the
+**    distribution.
+**
+** 3. All advertising materials mentioning features or use of this
+**    software must display the following acknowledgment:
+**    "This product includes software developed by the Apache Group
+**    for use in the Apache HTTP server project (http://www.apache.org/)."
+**
+** 4. The names "Apache Server" and "Apache Group" must not be used to
+**    endorse or promote products derived from this software without
+**    prior written permission. For written permission, please contact
+**    apache@apache.org.
+**
+** 5. Products derived from this software may not be called "Apache"
+**    nor may "Apache" appear in their names without prior written
+**    permission of the Apache Group.
+**
+** 6. Redistributions of any form whatsoever must retain the following
+**    acknowledgment:
+**    "This product includes software developed by the Apache Group
+**    for use in the Apache HTTP server project (http://www.apache.org/)."
+**
+** THIS SOFTWARE IS PROVIDED BY THE APACHE GROUP ``AS IS'' AND ANY
+** EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+** IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+** PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE APACHE GROUP OR
+** ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+** SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+** NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+** LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+** HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+** STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+** ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+** OF THE POSSIBILITY OF SUCH DAMAGE.
+** ====================================================================
+**
+** This software consists of voluntary contributions made by many
+** individuals on behalf of the Apache Group and was originally based
+** on public domain software written at the National Center for
+** Supercomputing Applications, University of Illinois, Urbana-Champaign.
+** For more information on the Apache Group and the Apache HTTP server
+** project, please see <http://www.apache.org/>.
+**
+** This code is based on, and used with the permission of, the
+** SIO stdio-replacement strx_* functions by Panos Tsirigotis
+** <panos@alumni.cs.colorado.edu> for xinetd.
+*/
 
 #include <stdio.h>
 #include <ctype.h>
 #include <netinet/in.h>
 
 typedef enum {
-    NO = 0, YES = 1
+       NO = 0, YES = 1
 } boolean_e;
 
 #ifndef FALSE
-#define FALSE                  0
+# define FALSE 0
 #endif
 #ifndef TRUE
-#define TRUE                   1
+# define TRUE 1
 #endif
-#define NUL                    '\0'
-#define INT_NULL               ((int *)0)
-#define WIDE_INT               long
+#define NUL '\0'
+#define INT_NULL ((int *)0)
+#define WIDE_INT long
 
 typedef struct {
-    char *curpos;
-    char *endpos;
+       char *curpos;
+       char *endpos;
 } ap_vformatter_buff;
 
 typedef WIDE_INT wide_int;
 typedef unsigned WIDE_INT u_wide_int;
 typedef int bool_int;
 
-#define S_NULL                 "(null)"
-#define S_NULL_LEN             6
+#define S_NULL "(null)"
+#define S_NULL_LEN 6
 
-#define FLOAT_DIGITS           6
-#define EXPONENT_LENGTH                10
+#define FLOAT_DIGITS 6
+#define EXPONENT_LENGTH 10
 
-/* These macros allow correct support of 8-bit characters on systems which
- * support 8-bit characters.  Pretty dumb how the cast is required, but
- * that's legacy libc for ya.  These new macros do not support EOF like
- * the standard macros do.  Tough.
- */
+/*
+** These macros allow correct support of 8-bit characters on systems which
+** support 8-bit characters.  Pretty dumb how the cast is required, but
+** that's legacy libc for ya.  These new macros do not support EOF like
+** the standard macros do.  Tough.
+*/
 #define ap_isalpha(c) (isalpha(((unsigned char)(c))))
 #define ap_isdigit(c) (isdigit(((unsigned char)(c))))
 #define ap_islower(c) (islower(((unsigned char)(c))))
 
 /*
- * NUM_BUF_SIZE is the size of the buffer used for arithmetic conversions
- *
- * XXX: this is a magic number; do not decrease it
- */
-#define NUM_BUF_SIZE           512
+** NUM_BUF_SIZE is the size of the buffer used for arithmetic conversions
+**
+** XXX: this is a magic number; do not decrease it
+*/
+#define NUM_BUF_SIZE 512
 
 /*
- * cvt.c - IEEE floating point formatting routines for FreeBSD
- * from GNU libc-4.6.27.  Modified to be thread safe.
- */
+** cvt.c - IEEE floating point formatting routines for FreeBSD
+** from GNU libc-4.6.27.  Modified to be thread safe.
+*/
 
 /*
- *    ap_ecvt converts to decimal
- *      the number of digits is specified by ndigit
- *      decpt is set to the position of the decimal point
- *      sign is set to 0 for positive, 1 for negative
- */
+**    ap_ecvt converts to decimal
+**      the number of digits is specified by ndigit
+**      decpt is set to the position of the decimal point
+**      sign is set to 0 for positive, 1 for negative
+*/
 
-#define        NDIG    80
+#define NDIG 80
 
 /* buf must have at least NDIG bytes */
-static char *ap_cvt(double arg, int ndigits, int *decpt, int *sign, int eflag, char *buf)
+static char *
+ap_cvt(double arg, int ndigits, int *decpt, int *sign, int eflag, char *buf)
 {
-    register int r2;
-    double fi, fj;
-    register char *p, *p1;
-
-    if (ndigits >= NDIG - 1)
-       ndigits = NDIG - 2;
-    r2 = 0;
-    *sign = 0;
-    p = &buf[0];
-    if (arg < 0) {
-       *sign = 1;
-       arg = -arg;
-    }
-    arg = modf(arg, &fi);
-    p1 = &buf[NDIG];
-    /*
-     * Do integer part
-     */
-    if (fi != 0) {
+       register int r2;
+       double fi, fj;
+       register char *p, *p1;
+
+       if (ndigits >= NDIG - 1)
+               ndigits = NDIG - 2;
+       r2 = 0;
+       *sign = 0;
+       p = &buf[0];
+       if (arg < 0) {
+               *sign = 1;
+               arg = -arg;
+       }
+       arg = modf(arg, &fi);
        p1 = &buf[NDIG];
-       while (fi != 0) {
-           fj = modf(fi / 10, &fi);
-           *--p1 = (int) ((fj + .03) * 10) + '0';
-           r2++;
+       /*
+       ** Do integer part
+       */
+       if (fi != 0) {
+               p1 = &buf[NDIG];
+               while (fi != 0) {
+                       fj = modf(fi / 10, &fi);
+                       *--p1 = (int) ((fj + .03) * 10) + '0';
+                       r2++;
+               }
+               while (p1 < &buf[NDIG])
+                       *p++ = *p1++;
+       } else if (arg > 0) {
+               while ((fj = arg * 10) < 1) {
+                       arg = fj;
+                       r2--;
+               }
        }
-       while (p1 < &buf[NDIG])
-           *p++ = *p1++;
-    }
-    else if (arg > 0) {
-       while ((fj = arg * 10) < 1) {
-           arg = fj;
-           r2--;
+       p1 = &buf[ndigits];
+       if (eflag == 0)
+               p1 += r2;
+       *decpt = r2;
+       if (p1 < &buf[0]) {
+               buf[0] = '\0';
+               return (buf);
        }
-    }
-    p1 = &buf[ndigits];
-    if (eflag == 0)
-       p1 += r2;
-    *decpt = r2;
-    if (p1 < &buf[0]) {
-       buf[0] = '\0';
-       return (buf);
-    }
-    while (p <= p1 && p < &buf[NDIG]) {
-       arg *= 10;
-       arg = modf(arg, &fj);
-       *p++ = (int) fj + '0';
-    }
-    if (p1 >= &buf[NDIG]) {
-       buf[NDIG - 1] = '\0';
-       return (buf);
-    }
-    p = p1;
-    *p1 += 5;
-    while (*p1 > '9') {
-       *p1 = '0';
-       if (p1 > buf)
-           ++ * --p1;
-       else {
-           *p1 = '1';
-           (*decpt)++;
-           if (eflag == 0) {
-               if (p > buf)
-                   *p = '0';
-               p++;
-           }
+       while (p <= p1 && p < &buf[NDIG]) {
+               arg *= 10;
+               arg = modf(arg, &fj);
+               *p++ = (int) fj + '0';
+       }
+       if (p1 >= &buf[NDIG]) {
+               buf[NDIG - 1] = '\0';
+               return (buf);
+       }
+       p = p1;
+       *p1 += 5;
+       while (*p1 > '9') {
+               *p1 = '0';
+               if (p1 > buf)
+                       ++ * --p1;
+               else {
+                       *p1 = '1';
+                       (*decpt)++;
+                       if (eflag == 0) {
+                               if (p > buf)
+                                       *p = '0';
+                               p++;
+                       }
+               }
        }
-    }
-    *p = '\0';
-    return (buf);
+       *p = '\0';
+       return (buf);
 }
 
-static char *ap_ecvt(double arg, int ndigits, int *decpt, int *sign, char *buf)
+static char *
+ap_ecvt(double arg, int ndigits, int *decpt, int *sign, char *buf)
 {
-    return (ap_cvt(arg, ndigits, decpt, sign, 1, buf));
+       return (ap_cvt(arg, ndigits, decpt, sign, 1, buf));
 }
 
-static char *ap_fcvt(double arg, int ndigits, int *decpt, int *sign, char *buf)
+static char *
+ap_fcvt(double arg, int ndigits, int *decpt, int *sign, char *buf)
 {
-    return (ap_cvt(arg, ndigits, decpt, sign, 0, buf));
+       return (ap_cvt(arg, ndigits, decpt, sign, 0, buf));
 }
 
 /*
- * ap_gcvt  - Floating output conversion to
- * minimal length string
- */
+** ap_gcvt  - Floating output conversion to
+** minimal length string
+*/
 
-static char *ap_gcvt(double number, int ndigit, char *buf, boolean_e altform)
+static char *
+ap_gcvt(double number, int ndigit, char *buf, boolean_e altform)
 {
-    int sign, decpt;
-    register char *p1, *p2;
-    register int i;
-    char buf1[NDIG];
-
-    p1 = ap_ecvt(number, ndigit, &decpt, &sign, buf1);
-    p2 = buf;
-    if (sign)
-       *p2++ = '-';
-    for (i = ndigit - 1; i > 0 && p1[i] == '0'; i--)
-       ndigit--;
-    if ((decpt >= 0 && decpt - ndigit > 4)
-       || (decpt < 0 && decpt < -3)) {         /* use E-style */
-       decpt--;
-       *p2++ = *p1++;
-       *p2++ = '.';
-       for (i = 1; i < ndigit; i++)
-           *p2++ = *p1++;
-       *p2++ = 'e';
-       if (decpt < 0) {
-           decpt = -decpt;
-           *p2++ = '-';
-       }
-       else
-           *p2++ = '+';
-       if (decpt / 100 > 0)
-           *p2++ = decpt / 100 + '0';
-       if (decpt / 10 > 0)
-           *p2++ = (decpt % 100) / 10 + '0';
-       *p2++ = decpt % 10 + '0';
-    }
-    else {
-       if (decpt <= 0) {
-           if (*p1 != '0')
-               *p2++ = '.';
-           while (decpt < 0) {
-               decpt++;
-               *p2++ = '0';
-           }
-       }
-       for (i = 1; i <= ndigit; i++) {
-           *p2++ = *p1++;
-           if (i == decpt)
+       int sign, decpt;
+       register char *p1, *p2;
+       register int i;
+       char buf1[NDIG];
+
+       p1 = ap_ecvt(number, ndigit, &decpt, &sign, buf1);
+       p2 = buf;
+       if (sign)
+               *p2++ = '-';
+       for (i = ndigit - 1; i > 0 && p1[i] == '0'; i--)
+               ndigit--;
+       if ((decpt >= 0 && decpt - ndigit > 4)
+               || (decpt < 0 && decpt < -3)) {  /* use E-style */
+               decpt--;
+               *p2++ = *p1++;
                *p2++ = '.';
+               for (i = 1; i < ndigit; i++)
+                       *p2++ = *p1++;
+               *p2++ = 'e';
+               if (decpt < 0) {
+                       decpt = -decpt;
+                       *p2++ = '-';
+               } else
+                       *p2++ = '+';
+               if (decpt / 100 > 0)
+                       *p2++ = decpt / 100 + '0';
+               if (decpt / 10 > 0)
+                       *p2++ = (decpt % 100) / 10 + '0';
+               *p2++ = decpt % 10 + '0';
+       } else {
+               if (decpt <= 0) {
+                       if (*p1 != '0')
+                               *p2++ = '.';
+                       while (decpt < 0) {
+                               decpt++;
+                               *p2++ = '0';
+                       }
+               }
+               for (i = 1; i <= ndigit; i++) {
+                       *p2++ = *p1++;
+                       if (i == decpt)
+                               *p2++ = '.';
+               }
+               if (ndigit < decpt) {
+                       while (ndigit++ < decpt)
+                               *p2++ = '0';
+                       *p2++ = '.';
+               }
        }
-       if (ndigit < decpt) {
-           while (ndigit++ < decpt)
-               *p2++ = '0';
-           *p2++ = '.';
-       }
-    }
-    if (p2[-1] == '.' && !altform)
-       p2--;
-    *p2 = '\0';
-    return (buf);
+       if (p2[-1] == '.' && !altform)
+               p2--;
+       *p2 = '\0';
+       return (buf);
 }
 
 /*
- * The INS_CHAR macro inserts a character in the buffer and writes
- * the buffer back to disk if necessary
- * It uses the char pointers sp and bep:
- *      sp points to the next available character in the buffer
- *      bep points to the end-of-buffer+1
- * While using this macro, note that the nextb pointer is NOT updated.
- *
- * NOTE: Evaluation of the c argument should not have any side-effects
- */
-#define INS_CHAR(c, sp, bep, cc)                               \
-           {                                                   \
-               if (sp >= bep) {                                \
-                   vbuff->curpos = sp;                         \
-                   if (flush_func(vbuff))                      \
-                       return -1;                              \
-                   sp = vbuff->curpos;                         \
-                   bep = vbuff->endpos;                        \
-               }                                               \
-               *sp++ = (c);                                    \
-               cc++;                                           \
-           }
-
-#define NUM( c )                       ( c - '0' )
-
-#define STR_TO_DEC( str, num )         \
-    num = NUM( *str++ ) ;              \
-    while ( ap_isdigit( *str ) )               \
-    {                                  \
-       num *= 10 ;                     \
-       num += NUM( *str++ ) ;          \
-    }
+** The INS_CHAR macro inserts a character in the buffer and writes
+** the buffer back to disk if necessary
+** It uses the char pointers sp and bep:
+**      sp points to the next available character in the buffer
+**      bep points to the end-of-buffer+1
+** While using this macro, note that the nextb pointer is NOT updated.
+**
+** NOTE: Evaluation of the c argument should not have any side-effects
+*/
+#define INS_CHAR(c, sp, bep, cc) \
+       { \
+               if (sp >= bep) { \
+                       vbuff->curpos = sp; \
+                       if (flush_func(vbuff)) \
+                               return -1; \
+                       sp = vbuff->curpos; \
+                       bep = vbuff->endpos; \
+               } \
+               *sp++ = (c); \
+               cc++; \
+       }
 
-/*
- * This macro does zero padding so that the precision
- * requirement is satisfied. The padding is done by
- * adding '0's to the left of the string that is going
- * to be printed.
- */
-#define FIX_PRECISION( adjust, precision, s, s_len )   \
-    if ( adjust )                                      \
-       while ( s_len < precision )                     \
-       {                                               \
-           *--s = '0' ;                                \
-           s_len++ ;                                   \
+#define NUM( c ) ( c - '0' )
+
+#define STR_TO_DEC( str, num ) \
+       num = NUM( *str++ ) ; \
+       while ( ap_isdigit( *str ) ) \
+       { \
+               num *= 10 ; \
+               num += NUM( *str++ ) ; \
        }
 
 /*
- * Macro that does padding. The padding is done by printing
- * the character ch.
- */
-#define PAD( width, len, ch )  do              \
-       {                                       \
-           INS_CHAR( ch, sp, bep, cc ) ;       \
-           width-- ;                           \
-       }                                       \
+** This macro does zero padding so that the precision
+** requirement is satisfied. The padding is done by
+** adding '0's to the left of the string that is going
+** to be printed.
+*/
+#define FIX_PRECISION( adjust, precision, s, s_len ) \
+       if ( adjust ) \
+               while ( s_len < precision ) \
+               { \
+                       *--s = '0' ; \
+                       s_len++ ; \
+               }
+
+/*
+** Macro that does padding. The padding is done by printing
+** the character ch.
+*/
+#define PAD( width, len, ch ) do \
+       { \
+               INS_CHAR( ch, sp, bep, cc ) ; \
+               width-- ; \
+       } \
        while ( width > len )
 
 /*
- * Prefix the character ch to the string str
- * Increase length
- * Set the has_prefix flag
- */
-#define PREFIX( str, length, ch )       *--str = ch ; length++ ; has_prefix = YES
+** Prefix the character ch to the string str
+** Increase length
+** Set the has_prefix flag
+*/
+#define PREFIX( str, length, ch ) *--str = ch ; length++ ; has_prefix = YES
 
 
 /*
- * Convert num to its decimal format.
- * Return value:
- *   - a pointer to a string containing the number (no sign)
- *   - len contains the length of the string
- *   - is_negative is set to TRUE or FALSE depending on the sign
- *     of the number (always set to FALSE if is_unsigned is TRUE)
- *
- * The caller provides a buffer for the string: that is the buf_end argument
- * which is a pointer to the END of the buffer + 1 (i.e. if the buffer
- * is declared as buf[ 100 ], buf_end should be &buf[ 100 ])
- */
-static char *conv_10(register wide_int num, register bool_int is_unsigned,
-                    register bool_int *is_negative, char *buf_end,
-                    register int *len)
+** Convert num to its decimal format.
+** Return value:
+**   - a pointer to a string containing the number (no sign)
+**   - len contains the length of the string
+**   - is_negative is set to TRUE or FALSE depending on the sign
+**     of the number (always set to FALSE if is_unsigned is TRUE)
+**
+** The caller provides a buffer for the string: that is the buf_end argument
+** which is a pointer to the END of the buffer + 1 (i.e. if the buffer
+** is declared as buf[ 100 ], buf_end should be &buf[ 100 ])
+*/
+static char *
+conv_10(register wide_int num, register bool_int is_unsigned,
+        register bool_int *is_negative, char *buf_end,
+        register int *len)
 {
-    register char *p = buf_end;
-    register u_wide_int magnitude;
+       register char *p = buf_end;
+       register u_wide_int magnitude;
+
+       if (is_unsigned) {
+               magnitude = (u_wide_int) num;
+               *is_negative = FALSE;
+       } else {
+               *is_negative = (num < 0);
 
-    if (is_unsigned) {
-       magnitude = (u_wide_int) num;
-       *is_negative = FALSE;
-    }
-    else {
-       *is_negative = (num < 0);
+               /*
+               ** On a 2's complement machine, negating the most negative
+               ** integer results in a number that cannot be represented
+               ** as a signed integer.
+               ** Here is what we do to obtain the number's magnitude:
+               **   a. add 1 to the number
+               **   b. negate it (becomes positive)
+               **   c. convert it to unsigned
+               **   d. add 1
+               */
+               if (*is_negative) {
+                       wide_int t = num + 1;
+
+                       magnitude = ((u_wide_int) -t) + 1;
+               } else
+                       magnitude = (u_wide_int) num;
+       }
 
        /*
-        * On a 2's complement machine, negating the most negative integer 
-        * results in a number that cannot be represented as a signed integer.
-        * Here is what we do to obtain the number's magnitude:
-        *      a. add 1 to the number
-        *      b. negate it (becomes positive)
-        *      c. convert it to unsigned
-        *      d. add 1
-        */
-       if (*is_negative) {
-           wide_int t = num + 1;
-
-           magnitude = ((u_wide_int) -t) + 1;
+       ** We use a do-while loop so that we write at least 1 digit
+       */
+       do {
+               register u_wide_int new_magnitude = magnitude / 10;
+
+               *--p = (char) (magnitude - new_magnitude * 10 + '0');
+               magnitude = new_magnitude;
        }
-       else
-           magnitude = (u_wide_int) num;
-    }
-
-    /*
-     * We use a do-while loop so that we write at least 1 digit 
-     */
-    do {
-       register u_wide_int new_magnitude = magnitude / 10;
-
-       *--p = (char) (magnitude - new_magnitude * 10 + '0');
-       magnitude = new_magnitude;
-    }
-    while (magnitude);
-
-    *len = buf_end - p;
-    return (p);
+       while (magnitude);
+
+       *len = buf_end - p;
+       return (p);
 }
 
 
 
-static char *conv_in_addr(struct in_addr *ia, char *buf_end, int *len)
+static char *
+conv_in_addr(struct in_addr *ia, char *buf_end, int *len)
 {
-    unsigned addr = ntohl(ia->s_addr);
-    char *p = buf_end;
-    bool_int is_negative;
-    int sub_len;
-
-    p = conv_10((addr & 0x000000FF)      , TRUE, &is_negative, p, &sub_len);
-    *--p = '.';
-    p = conv_10((addr & 0x0000FF00) >>  8, TRUE, &is_negative, p, &sub_len);
-    *--p = '.';
-    p = conv_10((addr & 0x00FF0000) >> 16, TRUE, &is_negative, p, &sub_len);
-    *--p = '.';
-    p = conv_10((addr & 0xFF000000) >> 24, TRUE, &is_negative, p, &sub_len);
-
-    *len = buf_end - p;
-    return (p);
+       unsigned addr = ntohl(ia->s_addr);
+       char *p = buf_end;
+       bool_int is_negative;
+       int sub_len;
+
+       p = conv_10((addr & 0x000000FF)  , TRUE, &is_negative, p, &sub_len);
+       *--p = '.';
+       p = conv_10((addr & 0x0000FF00) >>  8, TRUE, &is_negative, p, &sub_len);
+       *--p = '.';
+       p = conv_10((addr & 0x00FF0000) >> 16, TRUE, &is_negative, p, &sub_len);
+       *--p = '.';
+       p = conv_10((addr & 0xFF000000) >> 24, TRUE, &is_negative, p, &sub_len);
+
+       *len = buf_end - p;
+       return (p);
 }
 
 
 
-static char *conv_sockaddr_in(struct sockaddr_in *si, char *buf_end, int *len)
+static char *
+conv_sockaddr_in(struct sockaddr_in *si, char *buf_end, int *len)
 {
-    char *p = buf_end;
-    bool_int is_negative;
-    int sub_len;
+       char *p = buf_end;
+       bool_int is_negative;
+       int sub_len;
 
-    p = conv_10(ntohs(si->sin_port), TRUE, &is_negative, p, &sub_len);
-    *--p = ':';
-    p = conv_in_addr(&si->sin_addr, p, &sub_len);
+       p = conv_10(ntohs(si->sin_port), TRUE, &is_negative, p, &sub_len);
+       *--p = ':';
+       p = conv_in_addr(&si->sin_addr, p, &sub_len);
 
-    *len = buf_end - p;
-    return (p);
+       *len = buf_end - p;
+       return (p);
 }
 
 
 
 /*
- * Convert a floating point number to a string formats 'f', 'e' or 'E'.
- * The result is placed in buf, and len denotes the length of the string
- * The sign is returned in the is_negative argument (and is not placed
- * in buf).
- */
-static char *conv_fp(register char format, register double num,
-    boolean_e add_dp, int precision, bool_int *is_negative,
-    char *buf, int *len)
+** Convert a floating point number to a string formats 'f', 'e' or 'E'.
+** The result is placed in buf, and len denotes the length of the string
+** The sign is returned in the is_negative argument (and is not placed
+** in buf).
+*/
+static char *
+conv_fp(register char format, register double num,
+       boolean_e add_dp, int precision, bool_int *is_negative,
+       char *buf, int *len)
 {
-    register char *s = buf;
-    register char *p;
-    int decimal_point;
-    char buf1[NDIG];
-
-    if (format == 'f')
-       p = ap_fcvt(num, precision, &decimal_point, is_negative, buf1);
-    else                       /* either e or E format */
-       p = ap_ecvt(num, precision + 1, &decimal_point, is_negative, buf1);
-
-    /*
-     * Check for Infinity and NaN
-     */
-    if (ap_isalpha(*p)) {
-       *len = strlen(strcpy(buf, p));
-       *is_negative = FALSE;
-       return (buf);
-    }
-
-    if (format == 'f') {
-       if (decimal_point <= 0) {
-           *s++ = '0';
-           if (precision > 0) {
-               *s++ = '.';
-               while (decimal_point++ < 0)
-                   *s++ = '0';
-           }
-           else if (add_dp)
-               *s++ = '.';
+       register char *s = buf;
+       register char *p;
+       int decimal_point;
+       char buf1[NDIG];
+
+       if (format == 'f')
+               p = ap_fcvt(num, precision, &decimal_point, is_negative, buf1);
+       else  /* either e or E format */
+               p = ap_ecvt(num, precision + 1, &decimal_point, is_negative, buf1);
+
+       /*
+       ** Check for Infinity and NaN
+       */
+       if (ap_isalpha(*p)) {
+               *len = strlen(strcpy(buf, p));
+               *is_negative = FALSE;
+               return (buf);
        }
-       else {
-           while (decimal_point-- > 0)
+
+       if (format == 'f') {
+               if (decimal_point <= 0) {
+                       *s++ = '0';
+                       if (precision > 0) {
+                               *s++ = '.';
+                               while (decimal_point++ < 0)
+                                       *s++ = '0';
+                       } else if (add_dp)
+                               *s++ = '.';
+               } else {
+                       while (decimal_point-- > 0)
+                               *s++ = *p++;
+                       if (precision > 0 || add_dp)
+                               *s++ = '.';
+               }
+       } else {
                *s++ = *p++;
-           if (precision > 0 || add_dp)
-               *s++ = '.';
+               if (precision > 0 || add_dp)
+                       *s++ = '.';
        }
-    }
-    else {
-       *s++ = *p++;
-       if (precision > 0 || add_dp)
-           *s++ = '.';
-    }
-
-    /*
-     * copy the rest of p, the NUL is NOT copied
-     */
-    while (*p)
-       *s++ = *p++;
-
-    if (format != 'f') {
-       char temp[EXPONENT_LENGTH];     /* for exponent conversion */
-       int t_len;
-       bool_int exponent_is_negative;
-
-       *s++ = format;          /* either e or E */
-       decimal_point--;
-       if (decimal_point != 0) {
-           p = conv_10((wide_int) decimal_point, FALSE, &exponent_is_negative,
-                       &temp[EXPONENT_LENGTH], &t_len);
-           *s++ = exponent_is_negative ? '-' : '+';
-
-           /*
-            * Make sure the exponent has at least 2 digits
-            */
-           if (t_len == 1)
-               *s++ = '0';
-           while (t_len--)
+
+       /*
+       ** copy the rest of p, the NUL is NOT copied
+       */
+       while (*p)
                *s++ = *p++;
+
+       if (format != 'f') {
+               char temp[EXPONENT_LENGTH];  /* for exponent conversion */
+               int t_len;
+               bool_int exponent_is_negative;
+
+               *s++ = format;  /* either e or E */
+               decimal_point--;
+               if (decimal_point != 0) {
+                       p = conv_10((wide_int) decimal_point, FALSE,
+                               &exponent_is_negative, &temp[EXPONENT_LENGTH],
+                               &t_len);
+                       *s++ = exponent_is_negative ? '-' : '+';
+
+                       /*
+                       ** Make sure the exponent has at least 2 digits
+                       */
+                       if (t_len == 1)
+                               *s++ = '0';
+                       while (t_len--)
+                               *s++ = *p++;
+               } else {
+                       *s++ = '+';
+                       *s++ = '0';
+                       *s++ = '0';
+               }
        }
-       else {
-           *s++ = '+';
-           *s++ = '0';
-           *s++ = '0';
-       }
-    }
 
-    *len = s - buf;
-    return (buf);
+       *len = s - buf;
+       return (buf);
 }
 
 
 /*
- * Convert num to a base X number where X is a power of 2. nbits determines X.
- * For example, if nbits is 3, we do base 8 conversion
- * Return value:
- *      a pointer to a string containing the number
- *
- * The caller provides a buffer for the string: that is the buf_end argument
- * which is a pointer to the END of the buffer + 1 (i.e. if the buffer
- * is declared as buf[ 100 ], buf_end should be &buf[ 100 ])
- */
-static char *conv_p2(register u_wide_int num, register int nbits,
-                    char format, char *buf_end, register int *len)
+** Convert num to a base X number where X is a power of 2. nbits determines X.
+** For example, if nbits is 3, we do base 8 conversion
+** Return value:
+**      a pointer to a string containing the number
+**
+** The caller provides a buffer for the string: that is the buf_end argument
+** which is a pointer to the END of the buffer + 1 (i.e. if the buffer
+** is declared as buf[ 100 ], buf_end should be &buf[ 100 ])
+*/
+static char *
+conv_p2(register u_wide_int num, register int nbits,
+       char format, char *buf_end, register int *len)
 {
-    register int mask = (1 << nbits) - 1;
-    register char *p = buf_end;
-    static const char low_digits[] = "0123456789abcdef";
-    static const char upper_digits[] = "0123456789ABCDEF";
-    register const char *digits = (format == 'X') ? upper_digits : low_digits;
-
-    do {
-       *--p = digits[num & mask];
-       num >>= nbits;
-    }
-    while (num);
-
-    *len = buf_end - p;
-    return (p);
+       register int mask = (1 << nbits) - 1;
+       register char *p = buf_end;
+       static const char low_digits[] = "0123456789abcdef";
+       static const char upper_digits[] = "0123456789ABCDEF";
+       register const char *digits = (format == 'X') ? upper_digits : low_digits;
+
+       do {
+               *--p = digits[num & mask];
+               num >>= nbits;
+       }
+       while (num);
+
+       *len = buf_end - p;
+       return (p);
 }
 
 
 /*
- * Do format conversion placing the output in buffer
- */
-int ap_vformatter(int (*flush_func)(ap_vformatter_buff *),
-    ap_vformatter_buff *vbuff, const char *fmt, va_list ap)
+** Do format conversion placing the output in buffer
+*/
+static int
+ap_vformatter(int (*flush_func)(ap_vformatter_buff *),
+       ap_vformatter_buff *vbuff, const char *fmt, va_list ap)
 {
-    register char *sp;
-    register char *bep;
-    register int cc = 0;
-    register int i;
-
-    register char *s = NULL;
-    char *q;
-    int s_len;
-
-    register int min_width = 0;
-    int precision = 0;
-    enum {
-       LEFT, RIGHT
-    } adjust;
-    char pad_char;
-    char prefix_char;
-
-    double fp_num;
-    wide_int i_num = (wide_int) 0;
-    u_wide_int ui_num;
-
-    char num_buf[NUM_BUF_SIZE];
-    char char_buf[2];          /* for printing %% and %<unknown> */
-
-    /*
-     * Flag variables
-     */
-    boolean_e is_long;
-    boolean_e alternate_form;
-    boolean_e print_sign;
-    boolean_e print_blank;
-    boolean_e adjust_precision;
-    boolean_e adjust_width;
-    bool_int is_negative;
-
-    sp = vbuff->curpos;
-    bep = vbuff->endpos;
-
-    while (*fmt) {
-       if (*fmt != '%') {
-           INS_CHAR(*fmt, sp, bep, cc);
-       }
-       else {
-           /*
-            * Default variable settings
-            */
-           adjust = RIGHT;
-           alternate_form = print_sign = print_blank = NO;
-           pad_char = ' ';
-           prefix_char = NUL;
-
-           fmt++;
-
-           /*
-            * Try to avoid checking for flags, width or precision
-            */
-           if (!ap_islower(*fmt)) {
-               /*
-                * Recognize flags: -, #, BLANK, +
-                */
-               for (;; fmt++) {
-                   if (*fmt == '-')
-                       adjust = LEFT;
-                   else if (*fmt == '+')
-                       print_sign = YES;
-                   else if (*fmt == '#')
-                       alternate_form = YES;
-                   else if (*fmt == ' ')
-                       print_blank = YES;
-                   else if (*fmt == '0')
-                       pad_char = '0';
-                   else
-                       break;
-               }
+       register char *sp;
+       register char *bep;
+       register int cc = 0;
+       register int i;
+
+       register char *s = NULL;
+       char *q;
+       int s_len;
+
+       register int min_width = 0;
+       int precision = 0;
+       enum {
+               LEFT, RIGHT
+       } adjust;
+       char pad_char;
+       char prefix_char;
+
+       double fp_num;
+       wide_int i_num = (wide_int) 0;
+       u_wide_int ui_num;
+
+       char num_buf[NUM_BUF_SIZE];
+       char char_buf[2];  /* for printing %% and %<unknown> */
 
-               /*
-                * Check if a width was specified
-                */
-               if (ap_isdigit(*fmt)) {
-                   STR_TO_DEC(fmt, min_width);
-                   adjust_width = YES;
-               }
-               else if (*fmt == '*') {
-                   min_width = va_arg(ap, int);
-                   fmt++;
-                   adjust_width = YES;
-                   if (min_width < 0) {
-                       adjust = LEFT;
-                       min_width = -min_width;
-                   }
-               }
-               else
-                   adjust_width = NO;
+       /*
+       ** Flag variables
+       */
+       boolean_e is_long;
+       boolean_e alternate_form;
+       boolean_e print_sign;
+       boolean_e print_blank;
+       boolean_e adjust_precision;
+       boolean_e adjust_width;
+       bool_int is_negative;
+
+       sp = vbuff->curpos;
+       bep = vbuff->endpos;
+
+       while (*fmt) {
+               if (*fmt != '%') {
+                       INS_CHAR(*fmt, sp, bep, cc);
+               } else {
+                       /*
+                       ** Default variable settings
+                       */
+                       adjust = RIGHT;
+                       alternate_form = print_sign = print_blank = NO;
+                       pad_char = ' ';
+                       prefix_char = NUL;
 
-               /*
-                * Check if a precision was specified
-                *
-                * XXX: an unreasonable amount of precision may be specified
-                * resulting in overflow of num_buf. Currently we
-                * ignore this possibility.
-                */
-               if (*fmt == '.') {
-                   adjust_precision = YES;
-                   fmt++;
-                   if (ap_isdigit(*fmt)) {
-                       STR_TO_DEC(fmt, precision);
-                   }
-                   else if (*fmt == '*') {
-                       precision = va_arg(ap, int);
                        fmt++;
-                       if (precision < 0)
-                           precision = 0;
-                   }
-                   else
-                       precision = 0;
-               }
-               else
-                   adjust_precision = NO;
-           }
-           else
-               adjust_precision = adjust_width = NO;
-
-           /*
-            * Modifier check
-            */
-           if (*fmt == 'l') {
-               is_long = YES;
-               fmt++;
-           }
-           else {
-               if (*fmt == 'h')  /* "short" backward compatibility */
-                   ++fmt;
-               is_long = NO;
-           }
-
-           /*
-            * Argument extraction and printing.
-            * First we determine the argument type.
-            * Then, we convert the argument to a string.
-            * On exit from the switch, s points to the string that
-            * must be printed, s_len has the length of the string
-            * The precision requirements, if any, are reflected in s_len.
-            *
-            * NOTE: pad_char may be set to '0' because of the 0 flag.
-            *   It is reset to ' ' by non-numeric formats
-            */
-           switch (*fmt) {
-           case 'u':
-               if (is_long)
-                   i_num = va_arg(ap, u_wide_int);
-               else
-                   i_num = (wide_int) va_arg(ap, unsigned int);
-               s = conv_10(i_num, 1, &is_negative,
-                           &num_buf[NUM_BUF_SIZE], &s_len);
-               FIX_PRECISION(adjust_precision, precision, s, s_len);
-               break;
-
-           case 'd':
-           case 'i':
-               if (is_long)
-                   i_num = va_arg(ap, wide_int);
-               else
-                   i_num = (wide_int) va_arg(ap, int);
-               s = conv_10(i_num, 0, &is_negative,
-                           &num_buf[NUM_BUF_SIZE], &s_len);
-               FIX_PRECISION(adjust_precision, precision, s, s_len);
-
-               if (is_negative)
-                   prefix_char = '-';
-               else if (print_sign)
-                   prefix_char = '+';
-               else if (print_blank)
-                   prefix_char = ' ';
-               break;
-
-
-           case 'o':
-               if (is_long)
-                   ui_num = va_arg(ap, u_wide_int);
-               else
-                   ui_num = (u_wide_int) va_arg(ap, unsigned int);
-               s = conv_p2(ui_num, 3, *fmt,
-                           &num_buf[NUM_BUF_SIZE], &s_len);
-               FIX_PRECISION(adjust_precision, precision, s, s_len);
-               if (alternate_form && *s != '0') {
-                   *--s = '0';
-                   s_len++;
-               }
-               break;
-
-
-           case 'x':
-           case 'X':
-               if (is_long)
-                   ui_num = (u_wide_int) va_arg(ap, u_wide_int);
-               else
-                   ui_num = (u_wide_int) va_arg(ap, unsigned int);
-               s = conv_p2(ui_num, 4, *fmt,
-                           &num_buf[NUM_BUF_SIZE], &s_len);
-               FIX_PRECISION(adjust_precision, precision, s, s_len);
-               if (alternate_form && i_num != 0) {
-                   *--s = *fmt;        /* 'x' or 'X' */
-                   *--s = '0';
-                   s_len += 2;
-               }
-               break;
 
-
-           case 's':
-               s = va_arg(ap, char *);
-               if (s != NULL) {
-                   s_len = strlen(s);
-                   if (adjust_precision && precision < s_len)
-                       s_len = precision;
-               }
-               else {
-                   s = S_NULL;
-                   s_len = S_NULL_LEN;
-               }
-               pad_char = ' ';
-               break;
-
-
-           case 'f':
-           case 'e':
-           case 'E':
-               fp_num = va_arg(ap, double);
-               /*
-                * * We use &num_buf[ 1 ], so that we have room for the sign
-                */
-               s = conv_fp(*fmt, fp_num, alternate_form,
-                       (adjust_precision == NO) ? FLOAT_DIGITS : precision,
-                           &is_negative, &num_buf[1], &s_len);
-               if (is_negative)
-                   prefix_char = '-';
-               else if (print_sign)
-                   prefix_char = '+';
-               else if (print_blank)
-                   prefix_char = ' ';
-               break;
-
-
-           case 'g':
-           case 'G':
-               if (adjust_precision == NO)
-                   precision = FLOAT_DIGITS;
-               else if (precision == 0)
-                   precision = 1;
-               /*
-                * * We use &num_buf[ 1 ], so that we have room for the sign
-                */
-               s = ap_gcvt(va_arg(ap, double), precision, &num_buf[1],
-                           alternate_form);
-               if (*s == '-')
-                   prefix_char = *s++;
-               else if (print_sign)
-                   prefix_char = '+';
-               else if (print_blank)
-                   prefix_char = ' ';
-
-               s_len = strlen(s);
-
-               if (alternate_form && (q = strchr(s, '.')) == NULL) {
-                   s[s_len++] = '.';
-                   s[s_len] = '\0'; /* delimit for following strchr() */
-               }
-               if (*fmt == 'G' && (q = strchr(s, 'e')) != NULL)
-                   *q = 'E';
-               break;
-
-
-           case 'c':
-               char_buf[0] = (char) (va_arg(ap, int));
-               s = &char_buf[0];
-               s_len = 1;
-               pad_char = ' ';
-               break;
-
-
-           case '%':
-               char_buf[0] = '%';
-               s = &char_buf[0];
-               s_len = 1;
-               pad_char = ' ';
-               break;
-
-
-           case 'n':
-               *(va_arg(ap, int *)) = cc;
-               break;
-
-               /*
-                * This is where we extend the printf format, with a second
-                * type specifier
-                */
-           case 'p':
-               switch(*++fmt) {
-                   /*
-                    * If the pointer size is equal to the size of an unsigned
-                    * integer we convert the pointer to a hex number, otherwise 
-                    * we print "%p" to indicate that we don't handle "%p".
-                    */
-               case 'p':
-                   ui_num = (u_wide_int) va_arg(ap, void *);
-
-                   if (sizeof(char *) <= sizeof(u_wide_int))
-                               s = conv_p2(ui_num, 4, 'x',
-                                           &num_buf[NUM_BUF_SIZE], &s_len);
-                   else {
-                       s = "%p";
-                       s_len = 2;
-                       prefix_char = NUL;
-                   }
-                   pad_char = ' ';
-                   break;
-
-                   /* print a struct sockaddr_in as a.b.c.d:port */
-               case 'I':
-                   {
-                       struct sockaddr_in *si;
-
-                       si = va_arg(ap, struct sockaddr_in *);
-                       if (si != NULL) {
-                           s = conv_sockaddr_in(si, &num_buf[NUM_BUF_SIZE], &s_len);
-                           if (adjust_precision && precision < s_len)
-                               s_len = precision;
-                       }
-                       else {
-                           s = S_NULL;
-                           s_len = S_NULL_LEN;
+                       /*
+                       ** Try to avoid checking for flags, width or precision
+                       */
+                       if (!ap_islower(*fmt)) {
+                               /*
+                               ** Recognize flags: -, #, BLANK, +
+                               */
+                               for (;; fmt++) {
+                                       if (*fmt == '-')
+                                               adjust = LEFT;
+                                       else if (*fmt == '+')
+                                               print_sign = YES;
+                                       else if (*fmt == '#')
+                                               alternate_form = YES;
+                                       else if (*fmt == ' ')
+                                               print_blank = YES;
+                                       else if (*fmt == '0')
+                                               pad_char = '0';
+                                       else
+                                               break;
+                               }
+
+                               /*
+                               ** Check if a width was specified
+                               */
+                               if (ap_isdigit(*fmt)) {
+                                       STR_TO_DEC(fmt, min_width);
+                                       adjust_width = YES;
+                               } else if (*fmt == '*') {
+                                       min_width = va_arg(ap, int);
+                                       fmt++;
+                                       adjust_width = YES;
+                                       if (min_width < 0) {
+                                               adjust = LEFT;
+                                               min_width = -min_width;
+                                       }
+                               } else
+                                       adjust_width = NO;
+
+                               /*
+                               ** Check if a precision was specified
+                               **
+                               ** XXX: an unreasonable amount of precision
+                               ** may be specified resulting in overflow of
+                               ** num_buf. Currently we ignore this
+                               ** possibility.
+                               */
+                               if (*fmt == '.') {
+                                       adjust_precision = YES;
+                                       fmt++;
+                                       if (ap_isdigit(*fmt)) {
+                                               STR_TO_DEC(fmt, precision);
+                                       } else if (*fmt == '*') {
+                                               precision = va_arg(ap, int);
+                                               fmt++;
+                                               if (precision < 0)
+                                                       precision = 0;
+                                       } else
+                                               precision = 0;
+                               } else
+                                       adjust_precision = NO;
+                       } else
+                               adjust_precision = adjust_width = NO;
+
+                       /*
+                       ** Modifier check
+                       */
+                       if (*fmt == 'l') {
+                               is_long = YES;
+                               fmt++;
+                       } else {
+                               if (*fmt == 'h')  /* "short" backward compatibility */
+                                       ++fmt;
+                               is_long = NO;
                        }
-                       pad_char = ' ';
-                   }
-                   break;
-
-                   /* print a struct in_addr as a.b.c.d */
-               case 'A':
-                   {
-                       struct in_addr *ia;
-
-                       ia = va_arg(ap, struct in_addr *);
-                       if (ia != NULL) {
-                           s = conv_in_addr(ia, &num_buf[NUM_BUF_SIZE], &s_len);
-                           if (adjust_precision && precision < s_len)
-                               s_len = precision;
+
+                       /*
+                       ** Argument extraction and printing.
+                       ** First we determine the argument type.
+                       ** Then, we convert the argument to a string.
+                       ** On exit from the switch, s points to the string that
+                       ** must be printed, s_len has the length of the string
+                       ** The precision requirements, if any, are reflected
+                       ** in s_len.
+                       **
+                       ** NOTE: pad_char may be set to '0' because of the
+                       ** 0 flag.
+                       **   It is reset to ' ' by non-numeric formats
+                       */
+                       switch (*fmt) {
+                       case 'u':
+                               if (is_long)
+                                       i_num = va_arg(ap, u_wide_int);
+                               else
+                                       i_num = (wide_int) va_arg(ap, unsigned int);
+                               s = conv_10(i_num, 1, &is_negative,
+                                                       &num_buf[NUM_BUF_SIZE], &s_len);
+                               FIX_PRECISION(adjust_precision, precision, s, s_len);
+                               break;
+
+                       case 'd':
+                       case 'i':
+                               if (is_long)
+                                       i_num = va_arg(ap, wide_int);
+                               else
+                                       i_num = (wide_int) va_arg(ap, int);
+                               s = conv_10(i_num, 0, &is_negative,
+                                                       &num_buf[NUM_BUF_SIZE], &s_len);
+                               FIX_PRECISION(adjust_precision, precision, s, s_len);
+
+                               if (is_negative)
+                                       prefix_char = '-';
+                               else if (print_sign)
+                                       prefix_char = '+';
+                               else if (print_blank)
+                                       prefix_char = ' ';
+                               break;
+
+
+                       case 'o':
+                               if (is_long)
+                                       ui_num = va_arg(ap, u_wide_int);
+                               else
+                                       ui_num = (u_wide_int) va_arg(ap, unsigned int);
+                               s = conv_p2(ui_num, 3, *fmt,
+                                                       &num_buf[NUM_BUF_SIZE], &s_len);
+                               FIX_PRECISION(adjust_precision, precision, s, s_len);
+                               if (alternate_form && *s != '0') {
+                                       *--s = '0';
+                                       s_len++;
+                               }
+                               break;
+
+
+                       case 'x':
+                       case 'X':
+                               if (is_long)
+                                       ui_num = (u_wide_int) va_arg(ap, u_wide_int);
+                               else
+                                       ui_num = (u_wide_int) va_arg(ap, unsigned int);
+                               s = conv_p2(ui_num, 4, *fmt,
+                                                       &num_buf[NUM_BUF_SIZE], &s_len);
+                               FIX_PRECISION(adjust_precision, precision, s, s_len);
+                               if (alternate_form && i_num != 0) {
+                                       *--s = *fmt;  /* 'x' or 'X' */
+                                       *--s = '0';
+                                       s_len += 2;
+                               }
+                               break;
+
+
+                       case 's':
+                               s = va_arg(ap, char *);
+                               if (s != NULL) {
+                                       s_len = strlen(s);
+                                       if (adjust_precision && precision < s_len)
+                                               s_len = precision;
+                               }
+                               else {
+                                       s = S_NULL;
+                                       s_len = S_NULL_LEN;
+                               }
+                               pad_char = ' ';
+                               break;
+
+
+                       case 'f':
+                       case 'e':
+                       case 'E':
+                               fp_num = va_arg(ap, double);
+                               /*
+                               ** We use &num_buf[ 1 ], so that we have
+                               ** room for the sign
+                               */
+                               s = conv_fp(*fmt, fp_num, alternate_form,
+                                               (adjust_precision == NO) ? FLOAT_DIGITS : precision,
+                                                       &is_negative, &num_buf[1], &s_len);
+                               if (is_negative)
+                                       prefix_char = '-';
+                               else if (print_sign)
+                                       prefix_char = '+';
+                               else if (print_blank)
+                                       prefix_char = ' ';
+                               break;
+
+
+                       case 'g':
+                       case 'G':
+                               if (adjust_precision == NO)
+                                       precision = FLOAT_DIGITS;
+                               else if (precision == 0)
+                                       precision = 1;
+                               /*
+                               ** We use &num_buf[ 1 ], so that we have
+                               ** room for the sign
+                               */
+                               s = ap_gcvt(va_arg(ap, double), precision, &num_buf[1],
+                                                       alternate_form);
+                               if (*s == '-')
+                                       prefix_char = *s++;
+                               else if (print_sign)
+                                       prefix_char = '+';
+                               else if (print_blank)
+                                       prefix_char = ' ';
+
+                               s_len = strlen(s);
+
+                               if (alternate_form && (q = strchr(s, '.')) == NULL) {
+                                       s[s_len++] = '.';
+                                       s[s_len] = '\0'; /* delimit for following strchr() */
+                               }
+                               if (*fmt == 'G' && (q = strchr(s, 'e')) != NULL)
+                                       *q = 'E';
+                               break;
+
+
+                       case 'c':
+                               char_buf[0] = (char) (va_arg(ap, int));
+                               s = &char_buf[0];
+                               s_len = 1;
+                               pad_char = ' ';
+                               break;
+
+
+                       case '%':
+                               char_buf[0] = '%';
+                               s = &char_buf[0];
+                               s_len = 1;
+                               pad_char = ' ';
+                               break;
+
+
+                       case 'n':
+                               *(va_arg(ap, int *)) = cc;
+                               break;
+
+                               /*
+                               ** This is where we extend the printf format,
+                               ** with a second type specifier
+                               */
+                       case 'p':
+                               switch(*++fmt) {
+                                       /*
+                                       ** If the pointer size is equal to
+                                       ** the size of an unsigned integer
+                                       ** we convert the pointer to a hex
+                                       ** number, otherwise we print "%p"
+                                       ** to indicate that we don't handle
+                                       ** "%p".
+                                       */
+                               case 'p':
+                                       ui_num = (u_wide_int) va_arg(ap, void *);
+
+                                       if (sizeof(char *) <= sizeof(u_wide_int))
+                                                               s = conv_p2(ui_num, 4, 'x',
+                                                                                       &num_buf[NUM_BUF_SIZE], &s_len);
+                                       else {
+                                               s = "%p";
+                                               s_len = 2;
+                                               prefix_char = NUL;
+                                       }
+                                       pad_char = ' ';
+                                       break;
+
+                                       /*
+                                       ** print a struct sockaddr_in as
+                                       ** a.b.c.d:port
+                                       */
+                               case 'I':
+                                       {
+                                               struct sockaddr_in *si;
+
+                                               si = va_arg(ap, struct sockaddr_in *);
+                                               if (si != NULL) {
+                                                       s = conv_sockaddr_in(si, &num_buf[NUM_BUF_SIZE], &s_len);
+                                                       if (adjust_precision && precision < s_len)
+                                                               s_len = precision;
+                                               } else {
+                                                       s = S_NULL;
+                                                       s_len = S_NULL_LEN;
+                                               }
+                                               pad_char = ' ';
+                                       }
+                                       break;
+
+                                       /* print a struct in_addr as a.b.c.d */
+                               case 'A':
+                                       {
+                                               struct in_addr *ia;
+
+                                               ia = va_arg(ap, struct in_addr *);
+                                               if (ia != NULL) {
+                                                       s = conv_in_addr(ia, &num_buf[NUM_BUF_SIZE], &s_len);
+                                                       if (adjust_precision && precision < s_len)
+                                                               s_len = precision;
+                                               } else {
+                                                       s = S_NULL;
+                                                       s_len = S_NULL_LEN;
+                                               }
+                                               pad_char = ' ';
+                                       }
+                                       break;
+
+                               case NUL:
+                                       /* if %p ends the string, oh well ignore it */
+                                       continue;
+
+                               default:
+                                       s = "bogus %p";
+                                       s_len = 8;
+                                       prefix_char = NUL;
+                                       break;
+                               }
+                               break;
+
+                       case NUL:
+                               /*
+                               ** The last character of the format string
+                               ** was %.  We ignore it.
+                               */
+                               continue;
+
+
+                               /*
+                               ** The default case is for unrecognized %'s.
+                               ** We print %<char> to help the user identify
+                               ** what option is not understood.  This is
+                               ** also useful in case the user wants to
+                               ** pass the output of format_converter
+                               ** to another function that understands
+                               ** some other %<char> (like syslog).
+                               ** Note that we can't point s inside fmt
+                               ** because the unknown <char> could be
+                               ** preceded by width etc.
+                               */
+                       default:
+                               char_buf[0] = '%';
+                               char_buf[1] = *fmt;
+                               s = char_buf;
+                               s_len = 2;
+                               pad_char = ' ';
+                               break;
                        }
-                       else {
-                           s = S_NULL;
-                           s_len = S_NULL_LEN;
+
+                       if (prefix_char != NUL && s != S_NULL && s != char_buf) {
+                               *--s = prefix_char;
+                               s_len++;
                        }
-                       pad_char = ' ';
-                   }
-                   break;
-
-               case NUL:
-                   /* if %p ends the string, oh well ignore it */
-                   continue;
-
-               default:
-                   s = "bogus %p";
-                   s_len = 8;
-                   prefix_char = NUL;
-                   break;
-               }
-               break;
 
-           case NUL:
-               /*
-                * The last character of the format string was %.
-                * We ignore it.
-                */
-               continue;
+                       if (adjust_width && adjust == RIGHT && min_width > s_len) {
+                               if (pad_char == '0' && prefix_char != NUL) {
+                                       INS_CHAR(*s, sp, bep, cc);
+                                       s++;
+                                       s_len--;
+                                       min_width--;
+                               }
+                               PAD(min_width, s_len, pad_char);
+                       }
 
+                       /*
+                       ** Print the string s.
+                       */
+                       for (i = s_len; i != 0; i--) {
+                               INS_CHAR(*s, sp, bep, cc);
+                               s++;
+                       }
 
-               /*
-                * The default case is for unrecognized %'s.
-                * We print %<char> to help the user identify what
-                * option is not understood.
-                * This is also useful in case the user wants to pass
-                * the output of format_converter to another function
-                * that understands some other %<char> (like syslog).
-                * Note that we can't point s inside fmt because the
-                * unknown <char> could be preceded by width etc.
-                */
-           default:
-               char_buf[0] = '%';
-               char_buf[1] = *fmt;
-               s = char_buf;
-               s_len = 2;
-               pad_char = ' ';
-               break;
-           }
-
-           if (prefix_char != NUL && s != S_NULL && s != char_buf) {
-               *--s = prefix_char;
-               s_len++;
-           }
-
-           if (adjust_width && adjust == RIGHT && min_width > s_len) {
-               if (pad_char == '0' && prefix_char != NUL) {
-                   INS_CHAR(*s, sp, bep, cc);
-                   s++;
-                   s_len--;
-                   min_width--;
+                       if (adjust_width && adjust == LEFT && min_width > s_len)
+                               PAD(min_width, s_len, pad_char);
                }
-               PAD(min_width, s_len, pad_char);
-           }
-
-           /*
-            * Print the string s. 
-            */
-           for (i = s_len; i != 0; i--) {
-               INS_CHAR(*s, sp, bep, cc);
-               s++;
-           }
-
-           if (adjust_width && adjust == LEFT && min_width > s_len)
-               PAD(min_width, s_len, pad_char);
+               fmt++;
        }
-       fmt++;
-    }
-    vbuff->curpos = sp;
-    return cc;
+       vbuff->curpos = sp;
+       return cc;
 }
 
 
-static int snprintf_flush(ap_vformatter_buff *vbuff)
+static int
+snprintf_flush(ap_vformatter_buff *vbuff)
 {
-    /* if the buffer fills we have to abort immediately, there is no way
-     * to "flush" a snprintf... there's nowhere to flush it to.
-     */
-    return -1;
+       /*
+       ** if the buffer fills we have to abort immediately, there is no way
+       ** to "flush" a snprintf... there's nowhere to flush it to.
+       */
+       return -1;
 }
 
 
-int snprintf(char *buf, size_t len, const char *format,...)
+int
+snprintf(char *buf, size_t len, const char *format,...)
 {
-    int cc;
-    va_list ap;
-    ap_vformatter_buff vbuff;
-
-    if (len == 0)
-       return 0;
-
-    /* save one byte for nul terminator */
-    vbuff.curpos = buf;
-    vbuff.endpos = buf + len - 1;
-    va_start(ap, format);
-    cc = ap_vformatter(snprintf_flush, &vbuff, format, ap);
-    va_end(ap);
-    *vbuff.curpos = '\0';
-    return (cc == -1) ? len : cc;
+       int cc;
+       va_list ap;
+       ap_vformatter_buff vbuff;
+
+       if (len == 0)
+               return 0;
+
+       /* save one byte for nul terminator */
+       vbuff.curpos = buf;
+       vbuff.endpos = buf + len - 1;
+       va_start(ap, format);
+       cc = ap_vformatter(snprintf_flush, &vbuff, format, ap);
+       va_end(ap);
+       *vbuff.curpos = '\0';
+       return (cc == -1) ? len : cc;
 }
 
 
-int vsnprintf(char *buf, size_t len, const char *format, va_list ap)
+int
+vsnprintf(char *buf, size_t len, const char *format, va_list ap)
 {
-    int cc;
-    ap_vformatter_buff vbuff;
-
-    if (len == 0)
-       return 0;
-
-    /* save one byte for nul terminator */
-    vbuff.curpos = buf;
-    vbuff.endpos = buf + len - 1;
-    cc = ap_vformatter(snprintf_flush, &vbuff, format, ap);
-    *vbuff.curpos = '\0';
-    return (cc == -1) ? len : cc;
+       int cc;
+       ap_vformatter_buff vbuff;
+
+       if (len == 0)
+               return 0;
+
+       /* save one byte for nul terminator */
+       vbuff.curpos = buf;
+       vbuff.endpos = buf + len - 1;
+       cc = ap_vformatter(snprintf_flush, &vbuff, format, ap);
+       *vbuff.curpos = '\0';
+       return (cc == -1) ? len : cc;
 }