stdio-common/printf-parsemb.c

   1 /* Helper functions for parsing printf format strings.
   2    Copyright (C) 1995-2013 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           if (n != -1)
  95             {
  96               spec->data_arg = n - 1;
  97               *max_ref_arg = MAX (*max_ref_arg, n);
  98             }
  99         }
 100       else
 101         /* Oops; that was actually the width and/or 0 padding flag.
 102            Step back and read it again.  */
 103         format = begin;
 104     }
 105
 106   /* Check for spec modifiers.  */
 107   do
 108     {
 109       switch (*format)
 110         {
 111         case L_(' '):
 112           /* Output a space in place of a sign, when there is no sign.  */
 113           spec->info.space = 1;
 114           continue;
 115         case L_('+'):
 116           /* Always output + or - for numbers.  */
 117           spec->info.showsign = 1;
 118           continue;
 119         case L_('-'):
 120           /* Left-justify things.  */
 121           spec->info.left = 1;
 122           continue;
 123         case L_('#'):
 124           /* Use the "alternate form":
 125              Hex has 0x or 0X, FP always has a decimal point.  */
 126           spec->info.alt = 1;
 127           continue;
 128         case L_('0'):
 129           /* Pad with 0s.  */
 130           spec->info.pad = '0';
 131           continue;
 132         case L_('\''):
 133           /* Show grouping in numbers if the locale information
 134              indicates any.  */
 135           spec->info.group = 1;
 136           continue;
 137         case L_('I'):
 138           /* Use the internationalized form of the output.  Currently
 139              means to use the `outdigits' of the current locale.  */
 140           spec->info.i18n = 1;
 141           continue;
 142         default:
 143           break;
 144         }
 145       break;
 146     }
 147   while (*++format);
 148
 149   if (spec->info.left)
 150     spec->info.pad = ' ';
 151
 152   /* Get the field width.  */
 153   spec->width_arg = -1;
 154   spec->info.width = 0;
 155   if (*format == L_('*'))
 156     {
 157       /* The field width is given in an argument.
 158          A negative field width indicates left justification.  */
 159       const UCHAR_T *begin = ++format;
 160
 161       if (ISDIGIT (*format))
 162         {
 163           /* The width argument might be found in a positional parameter.  */
 164           n = read_int (&format);
 165
 166           if (n != 0 && *format == L_('$'))
 167             {
 168               if (n != -1)
 169                 {
 170                   spec->width_arg = n - 1;
 171                   *max_ref_arg = MAX (*max_ref_arg, n);
 172                 }
 173               ++format;         /* Skip '$'.  */
 174             }
 175         }
 176
 177       if (spec->width_arg < 0)
 178         {
 179           /* Not in a positional parameter.  Consume one argument.  */
 180           spec->width_arg = posn++;
 181           ++nargs;
 182           format = begin;       /* Step back and reread.  */
 183         }
 184     }
 185   else if (ISDIGIT (*format))
 186     {
 187       int n = read_int (&format);
 188
 189       /* Constant width specification.  */
 190       if (n != -1)
 191         spec->info.width = n;
 192     }
 193   /* Get the precision.  */
 194   spec->prec_arg = -1;
 195   /* -1 means none given; 0 means explicit 0.  */
 196   spec->info.prec = -1;
 197   if (*format == L_('.'))
 198     {
 199       ++format;
 200       if (*format == L_('*'))
 201         {
 202           /* The precision is given in an argument.  */
 203           const UCHAR_T *begin = ++format;
 204
 205           if (ISDIGIT (*format))
 206             {
 207               n = read_int (&format);
 208
 209               if (n != 0 && *format == L_('$'))
 210                 {
 211                   if (n != -1)
 212                     {
 213                       spec->prec_arg = n - 1;
 214                       *max_ref_arg = MAX (*max_ref_arg, n);
 215                     }
 216                   ++format;
 217                 }
 218             }
 219
 220           if (spec->prec_arg < 0)
 221             {
 222               /* Not in a positional parameter.  */
 223               spec->prec_arg = posn++;
 224               ++nargs;
 225               format = begin;
 226             }
 227         }
 228       else if (ISDIGIT (*format))
 229         {
 230           int n = read_int (&format);
 231
 232           if (n != -1)
 233             spec->info.prec = n;
 234         }
 235       else
 236         /* "%.?" is treated like "%.0?".  */
 237         spec->info.prec = 0;
 238     }
 239
 240   /* Check for type modifiers.  */
 241   spec->info.is_long_double = 0;
 242   spec->info.is_short = 0;
 243   spec->info.is_long = 0;
 244   spec->info.is_char = 0;
 245   spec->info.user = 0;
 246
 247   if (__builtin_expect (__printf_modifier_table == NULL, 1)
 248       || __printf_modifier_table[*format] == NULL
 249       || HANDLE_REGISTERED_MODIFIER (&format, &spec->info) != 0)
 250     switch (*format++)
 251       {
 252       case L_('h'):
 253         /* ints are short ints or chars.  */
 254         if (*format != L_('h'))
 255           spec->info.is_short = 1;
 256         else
 257           {
 258             ++format;
 259             spec->info.is_char = 1;
 260           }
 261         break;
 262       case L_('l'):
 263         /* ints are long ints.  */
 264         spec->info.is_long = 1;
 265         if (*format != L_('l'))
 266           break;
 267         ++format;
 268         /* FALLTHROUGH */
 269       case L_('L'):
 270         /* doubles are long doubles, and ints are long long ints.  */
 271       case L_('q'):
 272         /* 4.4 uses this for long long.  */
 273         spec->info.is_long_double = 1;
 274         break;
 275       case L_('z'):
 276       case L_('Z'):
 277         /* ints are size_ts.  */
 278         assert (sizeof (size_t) <= sizeof (unsigned long long int));
 279 #if LONG_MAX != LONG_LONG_MAX
 280         spec->info.is_long_double = (sizeof (size_t)
 281                                      > sizeof (unsigned long int));
 282 #endif
 283         spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 284         break;
 285       case L_('t'):
 286         assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 287 #if LONG_MAX != LONG_LONG_MAX
 288         spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 289 #endif
 290         spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 291         break;
 292       case L_('j'):
 293         assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 294 #if LONG_MAX != LONG_LONG_MAX
 295         spec->info.is_long_double = (sizeof (uintmax_t)
 296                                      > sizeof (unsigned long int));
 297 #endif
 298         spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 299         break;
 300       default:
 301         /* Not a recognized modifier.  Backup.  */
 302         --format;
 303         break;
 304       }
 305
 306   /* Get the format specification.  */
 307   spec->info.spec = (wchar_t) *format++;
 308   spec->size = -1;
 309   if (__builtin_expect (__printf_function_table == NULL, 1)
 310       || spec->info.spec > UCHAR_MAX
 311       || __printf_arginfo_table[spec->info.spec] == NULL
 312       /* We don't try to get the types for all arguments if the format
 313          uses more than one.  The normal case is covered though.  If
 314          the call returns -1 we continue with the normal specifiers.  */
 315       || (int) (spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 316                                    (&spec->info, 1, &spec->data_arg_type,
 317                                     &spec->size)) < 0)
 318     {
 319       /* Find the data argument types of a built-in spec.  */
 320       spec->ndata_args = 1;
 321
 322       switch (spec->info.spec)
 323         {
 324         case L'i':
 325         case L'd':
 326         case L'u':
 327         case L'o':
 328         case L'X':
 329         case L'x':
 330 #if LONG_MAX != LONG_LONG_MAX
 331           if (spec->info.is_long_double)
 332             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 333           else
 334 #endif
 335             if (spec->info.is_long)
 336               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 337             else if (spec->info.is_short)
 338               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 339             else if (spec->info.is_char)
 340               spec->data_arg_type = PA_CHAR;
 341             else
 342               spec->data_arg_type = PA_INT;
 343           break;
 344         case L'e':
 345         case L'E':
 346         case L'f':
 347         case L'F':
 348         case L'g':
 349         case L'G':
 350         case L'a':
 351         case L'A':
 352           if (spec->info.is_long_double)
 353             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 354           else
 355             spec->data_arg_type = PA_DOUBLE;
 356           break;
 357         case L'c':
 358           spec->data_arg_type = PA_CHAR;
 359           break;
 360         case L'C':
 361           spec->data_arg_type = PA_WCHAR;
 362           break;
 363         case L's':
 364           spec->data_arg_type = PA_STRING;
 365           break;
 366         case L'S':
 367           spec->data_arg_type = PA_WSTRING;
 368           break;
 369         case L'p':
 370           spec->data_arg_type = PA_POINTER;
 371           break;
 372         case L'n':
 373           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 374           break;
 375
 376         case L'm':
 377         default:
 378           /* An unknown spec will consume no args.  */
 379           spec->ndata_args = 0;
 380           break;
 381         }
 382     }
 383
 384   if (spec->data_arg == -1 && spec->ndata_args > 0)
 385     {
 386       /* There are args consumed, but no positional spec.  Use the
 387          next sequential arg position.  */
 388       spec->data_arg = posn;
 389       nargs += spec->ndata_args;
 390     }
 391
 392   if (spec->info.spec == L'\0')
 393     /* Format ended before this spec was complete.  */
 394     spec->end_of_fmt = spec->next_fmt = format - 1;
 395   else
 396     {
 397       /* Find the next format spec.  */
 398       spec->end_of_fmt = format;
 399 #ifdef COMPILE_WPRINTF
 400       spec->next_fmt = __find_specwc (format);
 401 #else
 402       spec->next_fmt = __find_specmb (format);
 403 #endif
 404     }
 405
 406   return nargs;
 407 }