stdio-common/printf-parsemb.c

   1 /* Helper functions for parsing printf format strings.
   2    Copyright (C) 1995-2023 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    <https://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   spec->info.is_binary128 = 0;
  83
  84   /* Test for positional argument.  */
  85   if (ISDIGIT (*format))
  86     {
  87       const UCHAR_T *begin = format;
  88
  89       n = read_int (&format);
  90
  91       if (n != 0 && *format == L_('$'))
  92         /* Is positional parameter.  */
  93         {
  94           ++format;             /* Skip the '$'.  */
  95           if (n != -1)
  96             {
  97               spec->data_arg = n - 1;
  98               *max_ref_arg = MAX (*max_ref_arg, n);
  99             }
 100         }
 101       else
 102         /* Oops; that was actually the width and/or 0 padding flag.
 103            Step back and read it again.  */
 104         format = begin;
 105     }
 106
 107   /* Check for spec modifiers.  */
 108   do
 109     {
 110       switch (*format)
 111         {
 112         case L_(' '):
 113           /* Output a space in place of a sign, when there is no sign.  */
 114           spec->info.space = 1;
 115           continue;
 116         case L_('+'):
 117           /* Always output + or - for numbers.  */
 118           spec->info.showsign = 1;
 119           continue;
 120         case L_('-'):
 121           /* Left-justify things.  */
 122           spec->info.left = 1;
 123           continue;
 124         case L_('#'):
 125           /* Use the "alternate form":
 126              Hex has 0x or 0X, FP always has a decimal point.  */
 127           spec->info.alt = 1;
 128           continue;
 129         case L_('0'):
 130           /* Pad with 0s.  */
 131           spec->info.pad = '0';
 132           continue;
 133         case L_('\''):
 134           /* Show grouping in numbers if the locale information
 135              indicates any.  */
 136           spec->info.group = 1;
 137           continue;
 138         case L_('I'):
 139           /* Use the internationalized form of the output.  Currently
 140              means to use the `outdigits' of the current locale.  */
 141           spec->info.i18n = 1;
 142           continue;
 143         default:
 144           break;
 145         }
 146       break;
 147     }
 148   while (*++format);
 149
 150   if (spec->info.left)
 151     spec->info.pad = ' ';
 152
 153   /* Get the field width.  */
 154   spec->width_arg = -1;
 155   spec->info.width = 0;
 156   if (*format == L_('*'))
 157     {
 158       /* The field width is given in an argument.
 159          A negative field width indicates left justification.  */
 160       const UCHAR_T *begin = ++format;
 161
 162       if (ISDIGIT (*format))
 163         {
 164           /* The width argument might be found in a positional parameter.  */
 165           n = read_int (&format);
 166
 167           if (n != 0 && *format == L_('$'))
 168             {
 169               if (n != -1)
 170                 {
 171                   spec->width_arg = n - 1;
 172                   *max_ref_arg = MAX (*max_ref_arg, n);
 173                 }
 174               ++format;         /* Skip '$'.  */
 175             }
 176         }
 177
 178       if (spec->width_arg < 0)
 179         {
 180           /* Not in a positional parameter.  Consume one argument.  */
 181           spec->width_arg = posn++;
 182           ++nargs;
 183           format = begin;       /* Step back and reread.  */
 184         }
 185     }
 186   else if (ISDIGIT (*format))
 187     {
 188       int n = read_int (&format);
 189
 190       /* Constant width specification.  */
 191       if (n != -1)
 192         spec->info.width = n;
 193     }
 194   /* Get the precision.  */
 195   spec->prec_arg = -1;
 196   /* -1 means none given; 0 means explicit 0.  */
 197   spec->info.prec = -1;
 198   if (*format == L_('.'))
 199     {
 200       ++format;
 201       if (*format == L_('*'))
 202         {
 203           /* The precision is given in an argument.  */
 204           const UCHAR_T *begin = ++format;
 205
 206           if (ISDIGIT (*format))
 207             {
 208               n = read_int (&format);
 209
 210               if (n != 0 && *format == L_('$'))
 211                 {
 212                   if (n != -1)
 213                     {
 214                       spec->prec_arg = n - 1;
 215                       *max_ref_arg = MAX (*max_ref_arg, n);
 216                     }
 217                   ++format;
 218                 }
 219             }
 220
 221           if (spec->prec_arg < 0)
 222             {
 223               /* Not in a positional parameter.  */
 224               spec->prec_arg = posn++;
 225               ++nargs;
 226               format = begin;
 227             }
 228         }
 229       else if (ISDIGIT (*format))
 230         {
 231           int n = read_int (&format);
 232
 233           if (n != -1)
 234             spec->info.prec = n;
 235         }
 236       else
 237         /* "%.?" is treated like "%.0?".  */
 238         spec->info.prec = 0;
 239     }
 240
 241   /* Check for type modifiers.  */
 242   spec->info.is_long_double = 0;
 243   spec->info.is_short = 0;
 244   spec->info.is_long = 0;
 245   spec->info.is_char = 0;
 246   spec->info.user = 0;
 247
 248   if (__builtin_expect (__printf_modifier_table == NULL, 1)
 249       || __printf_modifier_table[*format] == NULL
 250       || HANDLE_REGISTERED_MODIFIER (&format, &spec->info) != 0)
 251     switch (*format++)
 252       {
 253       case L_('h'):
 254         /* ints are short ints or chars.  */
 255         if (*format != L_('h'))
 256           spec->info.is_short = 1;
 257         else
 258           {
 259             ++format;
 260             spec->info.is_char = 1;
 261           }
 262         break;
 263       case L_('l'):
 264         /* ints are long ints.  */
 265         spec->info.is_long = 1;
 266         if (*format != L_('l'))
 267           break;
 268         ++format;
 269         /* FALLTHROUGH */
 270       case L_('L'):
 271         /* doubles are long doubles, and ints are long long ints.  */
 272       case L_('q'):
 273         /* 4.4 uses this for long long.  */
 274         spec->info.is_long_double = 1;
 275         break;
 276       case L_('z'):
 277       case L_('Z'):
 278         /* ints are size_ts.  */
 279         assert (sizeof (size_t) <= sizeof (unsigned long long int));
 280 #if LONG_MAX != LONG_LONG_MAX
 281         spec->info.is_long_double = (sizeof (size_t)
 282                                      > sizeof (unsigned long int));
 283 #endif
 284         spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 285         break;
 286       case L_('t'):
 287         assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 288 #if LONG_MAX != LONG_LONG_MAX
 289         spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 290 #endif
 291         spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 292         break;
 293       case L_('j'):
 294         assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 295 #if LONG_MAX != LONG_LONG_MAX
 296         spec->info.is_long_double = (sizeof (uintmax_t)
 297                                      > sizeof (unsigned long int));
 298 #endif
 299         spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 300         break;
 301       default:
 302         /* Not a recognized modifier.  Backup.  */
 303         --format;
 304         break;
 305       }
 306
 307   /* Get the format specification.  */
 308   spec->info.spec = (wchar_t) *format++;
 309   spec->size = -1;
 310   if (__builtin_expect (__printf_function_table == NULL, 1)
 311       || spec->info.spec > UCHAR_MAX
 312       || __printf_arginfo_table[spec->info.spec] == NULL
 313       /* We don't try to get the types for all arguments if the format
 314          uses more than one.  The normal case is covered though.  If
 315          the call returns -1 we continue with the normal specifiers.  */
 316       || (int) (spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 317                                    (&spec->info, 1, &spec->data_arg_type,
 318                                     &spec->size)) < 0)
 319     {
 320       /* Find the data argument types of a built-in spec.  */
 321       spec->ndata_args = 1;
 322
 323       switch (spec->info.spec)
 324         {
 325         case L'i':
 326         case L'd':
 327         case L'u':
 328         case L'o':
 329         case L'X':
 330         case L'x':
 331         case L'B':
 332         case L'b':
 333 #if LONG_MAX != LONG_LONG_MAX
 334           if (spec->info.is_long_double)
 335             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 336           else
 337 #endif
 338             if (spec->info.is_long)
 339               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 340             else if (spec->info.is_short)
 341               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 342             else if (spec->info.is_char)
 343               spec->data_arg_type = PA_CHAR;
 344             else
 345               spec->data_arg_type = PA_INT;
 346           break;
 347         case L'e':
 348         case L'E':
 349         case L'f':
 350         case L'F':
 351         case L'g':
 352         case L'G':
 353         case L'a':
 354         case L'A':
 355           if (spec->info.is_long_double)
 356             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 357           else
 358             spec->data_arg_type = PA_DOUBLE;
 359           break;
 360         case L'c':
 361           spec->data_arg_type = PA_CHAR;
 362           break;
 363         case L'C':
 364           spec->data_arg_type = PA_WCHAR;
 365           break;
 366         case L's':
 367           spec->data_arg_type = PA_STRING;
 368           break;
 369         case L'S':
 370           spec->data_arg_type = PA_WSTRING;
 371           break;
 372         case L'p':
 373           spec->data_arg_type = PA_POINTER;
 374           break;
 375         case L'n':
 376           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 377           break;
 378
 379         case L'm':
 380         default:
 381           /* An unknown spec will consume no args.  */
 382           spec->ndata_args = 0;
 383           break;
 384         }
 385     }
 386
 387   if (spec->data_arg == -1 && spec->ndata_args > 0)
 388     {
 389       /* There are args consumed, but no positional spec.  Use the
 390          next sequential arg position.  */
 391       spec->data_arg = posn;
 392       nargs += spec->ndata_args;
 393     }
 394
 395   if (spec->info.spec == L'\0')
 396     /* Format ended before this spec was complete.  */
 397     spec->end_of_fmt = spec->next_fmt = format - 1;
 398   else
 399     {
 400       /* Find the next format spec.  */
 401       spec->end_of_fmt = format;
 402 #ifdef COMPILE_WPRINTF
 403       spec->next_fmt = __find_specwc (format);
 404 #else
 405       spec->next_fmt = __find_specmb (format);
 406 #endif
 407     }
 408
 409   return nargs;
 410 }