stdio-common/printf-parsemb.c

   1 /* Helper functions for parsing printf format strings.
   2    Copyright (C) 1995-2024 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                     bool *failed)
  61 #else
  62 __parse_one_specmb (const UCHAR_T *format, size_t posn,
  63                     struct printf_spec *spec, size_t *max_ref_arg,
  64                     bool *failed)
  65 #endif
  66 {
  67   unsigned int n;
  68   size_t nargs = 0;
  69   bool is_fast;
  70
  71   /* Skip the '%'.  */
  72   ++format;
  73
  74   /* Clear information structure.  */
  75   spec->data_arg = -1;
  76   spec->info.alt = 0;
  77   spec->info.space = 0;
  78   spec->info.left = 0;
  79   spec->info.showsign = 0;
  80   spec->info.group = 0;
  81   spec->info.i18n = 0;
  82   spec->info.extra = 0;
  83   spec->info.pad = ' ';
  84   spec->info.wide = sizeof (UCHAR_T) > 1;
  85   spec->info.is_binary128 = 0;
  86
  87   *failed = false;
  88
  89   /* Test for positional argument.  */
  90   if (ISDIGIT (*format))
  91     {
  92       const UCHAR_T *begin = format;
  93
  94       n = read_int (&format);
  95
  96       if (n != 0 && *format == L_('$'))
  97         /* Is positional parameter.  */
  98         {
  99           ++format;             /* Skip the '$'.  */
 100           if (n != -1)
 101             {
 102               spec->data_arg = n - 1;
 103               *max_ref_arg = MAX (*max_ref_arg, n);
 104             }
 105         }
 106       else
 107         /* Oops; that was actually the width and/or 0 padding flag.
 108            Step back and read it again.  */
 109         format = begin;
 110     }
 111
 112   /* Check for spec modifiers.  */
 113   do
 114     {
 115       switch (*format)
 116         {
 117         case L_(' '):
 118           /* Output a space in place of a sign, when there is no sign.  */
 119           spec->info.space = 1;
 120           continue;
 121         case L_('+'):
 122           /* Always output + or - for numbers.  */
 123           spec->info.showsign = 1;
 124           continue;
 125         case L_('-'):
 126           /* Left-justify things.  */
 127           spec->info.left = 1;
 128           continue;
 129         case L_('#'):
 130           /* Use the "alternate form":
 131              Hex has 0x or 0X, FP always has a decimal point.  */
 132           spec->info.alt = 1;
 133           continue;
 134         case L_('0'):
 135           /* Pad with 0s.  */
 136           spec->info.pad = '0';
 137           continue;
 138         case L_('\''):
 139           /* Show grouping in numbers if the locale information
 140              indicates any.  */
 141           spec->info.group = 1;
 142           continue;
 143         case L_('I'):
 144           /* Use the internationalized form of the output.  Currently
 145              means to use the `outdigits' of the current locale.  */
 146           spec->info.i18n = 1;
 147           continue;
 148         default:
 149           break;
 150         }
 151       break;
 152     }
 153   while (*++format);
 154
 155   if (spec->info.left)
 156     spec->info.pad = ' ';
 157
 158   /* Get the field width.  */
 159   spec->width_arg = -1;
 160   spec->info.width = 0;
 161   if (*format == L_('*'))
 162     {
 163       /* The field width is given in an argument.
 164          A negative field width indicates left justification.  */
 165       const UCHAR_T *begin = ++format;
 166
 167       if (ISDIGIT (*format))
 168         {
 169           /* The width argument might be found in a positional parameter.  */
 170           n = read_int (&format);
 171
 172           if (n != 0 && *format == L_('$'))
 173             {
 174               if (n != -1)
 175                 {
 176                   spec->width_arg = n - 1;
 177                   *max_ref_arg = MAX (*max_ref_arg, n);
 178                 }
 179               ++format;         /* Skip '$'.  */
 180             }
 181         }
 182
 183       if (spec->width_arg < 0)
 184         {
 185           /* Not in a positional parameter.  Consume one argument.  */
 186           spec->width_arg = posn++;
 187           ++nargs;
 188           format = begin;       /* Step back and reread.  */
 189         }
 190     }
 191   else if (ISDIGIT (*format))
 192     {
 193       int n = read_int (&format);
 194
 195       /* Constant width specification.  */
 196       if (n != -1)
 197         spec->info.width = n;
 198     }
 199   /* Get the precision.  */
 200   spec->prec_arg = -1;
 201   /* -1 means none given; 0 means explicit 0.  */
 202   spec->info.prec = -1;
 203   if (*format == L_('.'))
 204     {
 205       ++format;
 206       if (*format == L_('*'))
 207         {
 208           /* The precision is given in an argument.  */
 209           const UCHAR_T *begin = ++format;
 210
 211           if (ISDIGIT (*format))
 212             {
 213               n = read_int (&format);
 214
 215               if (n != 0 && *format == L_('$'))
 216                 {
 217                   if (n != -1)
 218                     {
 219                       spec->prec_arg = n - 1;
 220                       *max_ref_arg = MAX (*max_ref_arg, n);
 221                     }
 222                   ++format;
 223                 }
 224             }
 225
 226           if (spec->prec_arg < 0)
 227             {
 228               /* Not in a positional parameter.  */
 229               spec->prec_arg = posn++;
 230               ++nargs;
 231               format = begin;
 232             }
 233         }
 234       else if (ISDIGIT (*format))
 235         {
 236           int n = read_int (&format);
 237
 238           if (n != -1)
 239             spec->info.prec = n;
 240         }
 241       else
 242         /* "%.?" is treated like "%.0?".  */
 243         spec->info.prec = 0;
 244     }
 245
 246   /* Check for type modifiers.  */
 247   spec->info.is_long_double = 0;
 248   spec->info.is_short = 0;
 249   spec->info.is_long = 0;
 250   spec->info.is_char = 0;
 251   spec->info.user = 0;
 252
 253   if (__builtin_expect (__printf_modifier_table == NULL, 1)
 254       || __printf_modifier_table[*format] == NULL
 255       || HANDLE_REGISTERED_MODIFIER (&format, &spec->info) != 0)
 256     switch (*format++)
 257       {
 258       case L_('h'):
 259         /* ints are short ints or chars.  */
 260         if (*format != L_('h'))
 261           spec->info.is_short = 1;
 262         else
 263           {
 264             ++format;
 265             spec->info.is_char = 1;
 266           }
 267         break;
 268       case L_('l'):
 269         /* ints are long ints.  */
 270         spec->info.is_long = 1;
 271         if (*format != L_('l'))
 272           break;
 273         ++format;
 274         /* FALLTHROUGH */
 275       case L_('L'):
 276         /* doubles are long doubles, and ints are long long ints.  */
 277       case L_('q'):
 278         /* 4.4 uses this for long long.  */
 279         spec->info.is_long_double = 1;
 280         break;
 281       case L_('z'):
 282       case L_('Z'):
 283         /* ints are size_ts.  */
 284         assert (sizeof (size_t) <= sizeof (unsigned long long int));
 285 #if LONG_MAX != LONG_LONG_MAX
 286         spec->info.is_long_double = (sizeof (size_t)
 287                                      > sizeof (unsigned long int));
 288 #endif
 289         spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 290         break;
 291       case L_('t'):
 292         assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 293 #if LONG_MAX != LONG_LONG_MAX
 294         spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 295 #endif
 296         spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 297         break;
 298       case L_('j'):
 299         assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 300 #if LONG_MAX != LONG_LONG_MAX
 301         spec->info.is_long_double = (sizeof (uintmax_t)
 302                                      > sizeof (unsigned long int));
 303 #endif
 304         spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 305         break;
 306       case L_('w'):
 307         is_fast = false;
 308         if (*format == L_('f'))
 309           {
 310             ++format;
 311             is_fast = true;
 312           }
 313         int bitwidth = 0;
 314         if (ISDIGIT (*format))
 315           bitwidth = read_int (&format);
 316         if (is_fast)
 317           switch (bitwidth)
 318             {
 319             case 8:
 320               bitwidth = INT_FAST8_WIDTH;
 321               break;
 322             case 16:
 323               bitwidth = INT_FAST16_WIDTH;
 324               break;
 325             case 32:
 326               bitwidth = INT_FAST32_WIDTH;
 327               break;
 328             case 64:
 329               bitwidth = INT_FAST64_WIDTH;
 330               break;
 331             }
 332         switch (bitwidth)
 333           {
 334           case 8:
 335             spec->info.is_char = 1;
 336             break;
 337           case 16:
 338             spec->info.is_short = 1;
 339             break;
 340           case 32:
 341             break;
 342           case 64:
 343             spec->info.is_long_double = 1;
 344             spec->info.is_long = 1;
 345             break;
 346           default:
 347             /* ISO C requires this error to be detected.  */
 348             __set_errno (EINVAL);
 349             *failed = true;
 350             break;
 351           }
 352         break;
 353       default:
 354         /* Not a recognized modifier.  Backup.  */
 355         --format;
 356         break;
 357       }
 358
 359   /* Get the format specification.  */
 360   spec->info.spec = (wchar_t) *format++;
 361   spec->size = -1;
 362   if (__builtin_expect (__printf_function_table == NULL, 1)
 363       || spec->info.spec > UCHAR_MAX
 364       || __printf_arginfo_table[spec->info.spec] == NULL
 365       /* We don't try to get the types for all arguments if the format
 366          uses more than one.  The normal case is covered though.  If
 367          the call returns -1 we continue with the normal specifiers.  */
 368       || (int) (spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 369                                    (&spec->info, 1, &spec->data_arg_type,
 370                                     &spec->size)) < 0)
 371     {
 372       /* Find the data argument types of a built-in spec.  */
 373       spec->ndata_args = 1;
 374
 375       switch (spec->info.spec)
 376         {
 377         case L'i':
 378         case L'd':
 379         case L'u':
 380         case L'o':
 381         case L'X':
 382         case L'x':
 383         case L'B':
 384         case L'b':
 385 #if LONG_MAX != LONG_LONG_MAX
 386           if (spec->info.is_long_double)
 387             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 388           else
 389 #endif
 390             if (spec->info.is_long)
 391               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 392             else if (spec->info.is_short)
 393               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 394             else if (spec->info.is_char)
 395               spec->data_arg_type = PA_CHAR;
 396             else
 397               spec->data_arg_type = PA_INT;
 398           break;
 399         case L'e':
 400         case L'E':
 401         case L'f':
 402         case L'F':
 403         case L'g':
 404         case L'G':
 405         case L'a':
 406         case L'A':
 407           if (spec->info.is_long_double)
 408             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 409           else
 410             spec->data_arg_type = PA_DOUBLE;
 411           break;
 412         case L'c':
 413           spec->data_arg_type = PA_CHAR;
 414           break;
 415         case L'C':
 416           spec->data_arg_type = PA_WCHAR;
 417           break;
 418         case L's':
 419           spec->data_arg_type = PA_STRING;
 420           break;
 421         case L'S':
 422           spec->data_arg_type = PA_WSTRING;
 423           break;
 424         case L'p':
 425           spec->data_arg_type = PA_POINTER;
 426           break;
 427         case L'n':
 428           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 429           break;
 430
 431         case L'm':
 432         default:
 433           /* An unknown spec will consume no args.  */
 434           spec->ndata_args = 0;
 435           break;
 436         }
 437     }
 438
 439   if (spec->data_arg == -1 && spec->ndata_args > 0)
 440     {
 441       /* There are args consumed, but no positional spec.  Use the
 442          next sequential arg position.  */
 443       spec->data_arg = posn;
 444       nargs += spec->ndata_args;
 445     }
 446
 447   if (spec->info.spec == L'\0')
 448     /* Format ended before this spec was complete.  */
 449     spec->end_of_fmt = spec->next_fmt = format - 1;
 450   else
 451     {
 452       /* Find the next format spec.  */
 453       spec->end_of_fmt = format;
 454 #ifdef COMPILE_WPRINTF
 455       spec->next_fmt = __find_specwc (format);
 456 #else
 457       spec->next_fmt = __find_specmb (format);
 458 #endif
 459     }
 460
 461   return nargs;
 462 }