stdio-common/printf-parsemb.c

   1 /* Helper functions for parsing printf format strings.
   2    Copyright (C) 1995-2000,2002-2004,2006,2009 Free Software Foundation, Inc.
   3    This file is part of th GNU C Library.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, see
  17    <http://www.gnu.org/licenses/>.  */
  18
  19 #include <ctype.h>
  20 #include <limits.h>
  21 #include <stdlib.h>
  22 #include <string.h>
  23 #include <sys/param.h>
  24 #include <wchar.h>
  25 #include <wctype.h>
  26
  27 #ifndef COMPILE_WPRINTF
  28 # define CHAR_T         char
  29 # define UCHAR_T        unsigned char
  30 # define INT_T          int
  31 # define L_(Str)        Str
  32 # define ISDIGIT(Ch)    isdigit (Ch)
  33 # define HANDLE_REGISTERED_MODIFIER __handle_registered_modifier_mb
  34 #else
  35 # define CHAR_T         wchar_t
  36 # define UCHAR_T        unsigned int
  37 # define INT_T          wint_t
  38 # define L_(Str)        L##Str
  39 # define ISDIGIT(Ch)    iswdigit (Ch)
  40 # define HANDLE_REGISTERED_MODIFIER __handle_registered_modifier_wc
  41 #endif
  42
  43 #include "printf-parse.h"
  44
  45 #define NDEBUG 1
  46 #include <assert.h>
  47
  48
  49
  50 /* FORMAT must point to a '%' at the beginning of a spec.  Fills in *SPEC
  51    with the parsed details.  POSN is the number of arguments already
  52    consumed.  At most MAXTYPES - POSN types are filled in TYPES.  Return
  53    the number of args consumed by this spec; *MAX_REF_ARG is updated so it
  54    remains the highest argument index used.  */
  55 size_t
  56 attribute_hidden
  57 #ifdef COMPILE_WPRINTF
  58 __parse_one_specwc (const UCHAR_T *format, size_t posn,
  59                     struct printf_spec *spec, size_t *max_ref_arg)
  60 #else
  61 __parse_one_specmb (const UCHAR_T *format, size_t posn,
  62                     struct printf_spec *spec, size_t *max_ref_arg)
  63 #endif
  64 {
  65   unsigned int n;
  66   size_t nargs = 0;
  67
  68   /* Skip the '%'.  */
  69   ++format;
  70
  71   /* Clear information structure.  */
  72   spec->data_arg = -1;
  73   spec->info.alt = 0;
  74   spec->info.space = 0;
  75   spec->info.left = 0;
  76   spec->info.showsign = 0;
  77   spec->info.group = 0;
  78   spec->info.i18n = 0;
  79   spec->info.extra = 0;
  80   spec->info.pad = ' ';
  81   spec->info.wide = sizeof (UCHAR_T) > 1;
  82
  83   /* Test for positional argument.  */
  84   if (ISDIGIT (*format))
  85     {
  86       const UCHAR_T *begin = format;
  87
  88       n = read_int (&format);
  89
  90       if (n > 0 && *format == L_('$'))
  91         /* Is positional parameter.  */
  92         {
  93           ++format;             /* Skip the '$'.  */
  94           spec->data_arg = n - 1;
  95           *max_ref_arg = MAX (*max_ref_arg, n);
  96         }
  97       else
  98         /* Oops; that was actually the width and/or 0 padding flag.
  99            Step back and read it again.  */
 100         format = begin;
 101     }
 102
 103   /* Check for spec modifiers.  */
 104   do
 105     {
 106       switch (*format)
 107         {
 108         case L_(' '):
 109           /* Output a space in place of a sign, when there is no sign.  */
 110           spec->info.space = 1;
 111           continue;
 112         case L_('+'):
 113           /* Always output + or - for numbers.  */
 114           spec->info.showsign = 1;
 115           continue;
 116         case L_('-'):
 117           /* Left-justify things.  */
 118           spec->info.left = 1;
 119           continue;
 120         case L_('#'):
 121           /* Use the "alternate form":
 122              Hex has 0x or 0X, FP always has a decimal point.  */
 123           spec->info.alt = 1;
 124           continue;
 125         case L_('0'):
 126           /* Pad with 0s.  */
 127           spec->info.pad = '0';
 128           continue;
 129         case L_('\''):
 130           /* Show grouping in numbers if the locale information
 131              indicates any.  */
 132           spec->info.group = 1;
 133           continue;
 134         case L_('I'):
 135           /* Use the internationalized form of the output.  Currently
 136              means to use the `outdigits' of the current locale.  */
 137           spec->info.i18n = 1;
 138           continue;
 139         default:
 140           break;
 141         }
 142       break;
 143     }
 144   while (*++format);
 145
 146   if (spec->info.left)
 147     spec->info.pad = ' ';
 148
 149   /* Get the field width.  */
 150   spec->width_arg = -1;
 151   spec->info.width = 0;
 152   if (*format == L_('*'))
 153     {
 154       /* The field width is given in an argument.
 155          A negative field width indicates left justification.  */
 156       const UCHAR_T *begin = ++format;
 157
 158       if (ISDIGIT (*format))
 159         {
 160           /* The width argument might be found in a positional parameter.  */
 161           n = read_int (&format);
 162
 163           if (n > 0 && *format == L_('$'))
 164             {
 165               spec->width_arg = n - 1;
 166               *max_ref_arg = MAX (*max_ref_arg, n);
 167               ++format;         /* Skip '$'.  */
 168             }
 169         }
 170
 171       if (spec->width_arg < 0)
 172         {
 173           /* Not in a positional parameter.  Consume one argument.  */
 174           spec->width_arg = posn++;
 175           ++nargs;
 176           format = begin;       /* Step back and reread.  */
 177         }
 178     }
 179   else if (ISDIGIT (*format))
 180     /* Constant width specification.  */
 181     spec->info.width = read_int (&format);
 182
 183   /* Get the precision.  */
 184   spec->prec_arg = -1;
 185   /* -1 means none given; 0 means explicit 0.  */
 186   spec->info.prec = -1;
 187   if (*format == L_('.'))
 188     {
 189       ++format;
 190       if (*format == L_('*'))
 191         {
 192           /* The precision is given in an argument.  */
 193           const UCHAR_T *begin = ++format;
 194
 195           if (ISDIGIT (*format))
 196             {
 197               n = read_int (&format);
 198
 199               if (n > 0 && *format == L_('$'))
 200                 {
 201                   spec->prec_arg = n - 1;
 202                   *max_ref_arg = MAX (*max_ref_arg, n);
 203                   ++format;
 204                 }
 205             }
 206
 207           if (spec->prec_arg < 0)
 208             {
 209               /* Not in a positional parameter.  */
 210               spec->prec_arg = posn++;
 211               ++nargs;
 212               format = begin;
 213             }
 214         }
 215       else if (ISDIGIT (*format))
 216         spec->info.prec = read_int (&format);
 217       else
 218         /* "%.?" is treated like "%.0?".  */
 219         spec->info.prec = 0;
 220     }
 221
 222   /* Check for type modifiers.  */
 223   spec->info.is_long_double = 0;
 224   spec->info.is_short = 0;
 225   spec->info.is_long = 0;
 226   spec->info.is_char = 0;
 227   spec->info.user = 0;
 228
 229   if (__builtin_expect (__printf_modifier_table == NULL, 1)
 230       || __printf_modifier_table[*format] == NULL
 231       || HANDLE_REGISTERED_MODIFIER (&format, &spec->info) != 0)
 232     switch (*format++)
 233       {
 234       case L_('h'):
 235         /* ints are short ints or chars.  */
 236         if (*format != L_('h'))
 237           spec->info.is_short = 1;
 238         else
 239           {
 240             ++format;
 241             spec->info.is_char = 1;
 242           }
 243         break;
 244       case L_('l'):
 245         /* ints are long ints.  */
 246         spec->info.is_long = 1;
 247         if (*format != L_('l'))
 248           break;
 249         ++format;
 250         /* FALLTHROUGH */
 251       case L_('L'):
 252         /* doubles are long doubles, and ints are long long ints.  */
 253       case L_('q'):
 254         /* 4.4 uses this for long long.  */
 255         spec->info.is_long_double = 1;
 256         break;
 257       case L_('z'):
 258       case L_('Z'):
 259         /* ints are size_ts.  */
 260         assert (sizeof (size_t) <= sizeof (unsigned long long int));
 261 #if LONG_MAX != LONG_LONG_MAX
 262         spec->info.is_long_double = (sizeof (size_t)
 263                                      > sizeof (unsigned long int));
 264 #endif
 265         spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 266         break;
 267       case L_('t'):
 268         assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 269 #if LONG_MAX != LONG_LONG_MAX
 270         spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 271 #endif
 272         spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 273         break;
 274       case L_('j'):
 275         assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 276 #if LONG_MAX != LONG_LONG_MAX
 277         spec->info.is_long_double = (sizeof (uintmax_t)
 278                                      > sizeof (unsigned long int));
 279 #endif
 280         spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 281         break;
 282       default:
 283         /* Not a recognized modifier.  Backup.  */
 284         --format;
 285         break;
 286       }
 287
 288   /* Get the format specification.  */
 289   spec->info.spec = (wchar_t) *format++;
 290   spec->size = -1;
 291   if (__builtin_expect (__printf_function_table == NULL, 1)
 292       || spec->info.spec > UCHAR_MAX
 293       || __printf_arginfo_table[spec->info.spec] == NULL
 294       /* We don't try to get the types for all arguments if the format
 295          uses more than one.  The normal case is covered though.  If
 296          the call returns -1 we continue with the normal specifiers.  */
 297       || (int) (spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 298                                    (&spec->info, 1, &spec->data_arg_type,
 299                                     &spec->size)) < 0)
 300     {
 301       /* Find the data argument types of a built-in spec.  */
 302       spec->ndata_args = 1;
 303
 304       switch (spec->info.spec)
 305         {
 306         case L'i':
 307         case L'd':
 308         case L'u':
 309         case L'o':
 310         case L'X':
 311         case L'x':
 312 #if LONG_MAX != LONG_LONG_MAX
 313           if (spec->info.is_long_double)
 314             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 315           else
 316 #endif
 317             if (spec->info.is_long)
 318               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 319             else if (spec->info.is_short)
 320               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 321             else if (spec->info.is_char)
 322               spec->data_arg_type = PA_CHAR;
 323             else
 324               spec->data_arg_type = PA_INT;
 325           break;
 326         case L'e':
 327         case L'E':
 328         case L'f':
 329         case L'F':
 330         case L'g':
 331         case L'G':
 332         case L'a':
 333         case L'A':
 334           if (spec->info.is_long_double)
 335             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 336           else
 337             spec->data_arg_type = PA_DOUBLE;
 338           break;
 339         case L'c':
 340           spec->data_arg_type = PA_CHAR;
 341           break;
 342         case L'C':
 343           spec->data_arg_type = PA_WCHAR;
 344           break;
 345         case L's':
 346           spec->data_arg_type = PA_STRING;
 347           break;
 348         case L'S':
 349           spec->data_arg_type = PA_WSTRING;
 350           break;
 351         case L'p':
 352           spec->data_arg_type = PA_POINTER;
 353           break;
 354         case L'n':
 355           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 356           break;
 357
 358         case L'm':
 359         default:
 360           /* An unknown spec will consume no args.  */
 361           spec->ndata_args = 0;
 362           break;
 363         }
 364     }
 365
 366   if (spec->data_arg == -1 && spec->ndata_args > 0)
 367     {
 368       /* There are args consumed, but no positional spec.  Use the
 369          next sequential arg position.  */
 370       spec->data_arg = posn;
 371       nargs += spec->ndata_args;
 372     }
 373
 374   if (spec->info.spec == L'\0')
 375     /* Format ended before this spec was complete.  */
 376     spec->end_of_fmt = spec->next_fmt = format - 1;
 377   else
 378     {
 379       /* Find the next format spec.  */
 380       spec->end_of_fmt = format;
 381 #ifdef COMPILE_WPRINTF
 382       spec->next_fmt = __find_specwc (format);
 383 #else
 384       spec->next_fmt = __find_specmb (format);
 385 #endif
 386     }
 387
 388   return nargs;
 389 }