stdio-common/printf-parse.h

   1 /* Internal header for parsing printf format strings.
   2    Copyright (C) 1995-1999, 2000 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 Library General Public License as
   7    published by the Free Software Foundation; either version 2 of the
   8    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    Library General Public License for more details.
  14
  15    You should have received a copy of the GNU Library General Public
  16    License along with the GNU C Library; see the file COPYING.LIB.  If not,
  17    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  18    Boston, MA 02111-1307, USA.  */
  19
  20 #include <ctype.h>
  21 #include <limits.h>
  22 #include <printf.h>
  23 #include <stdint.h>
  24 #include <stddef.h>
  25 #include <string.h>
  26
  27 #define NDEBUG 1
  28 #include <assert.h>
  29
  30
  31 struct printf_spec
  32   {
  33     /* Information parsed from the format spec.  */
  34     struct printf_info info;
  35
  36     /* Pointers into the format string for the end of this format
  37        spec and the next (or to the end of the string if no more).  */
  38     const UCHAR_T *end_of_fmt, *next_fmt;
  39
  40     /* Position of arguments for precision and width, or -1 if `info' has
  41        the constant value.  */
  42     int prec_arg, width_arg;
  43
  44     int data_arg;               /* Position of data argument.  */
  45     int data_arg_type;          /* Type of first argument.  */
  46     /* Number of arguments consumed by this format specifier.  */
  47     size_t ndata_args;
  48   };
  49
  50
  51 /* The various kinds off arguments that can be passed to printf.  */
  52 union printf_arg
  53   {
  54     unsigned char pa_char;
  55     wchar_t pa_wchar;
  56     short int pa_short_int;
  57     int pa_int;
  58     long int pa_long_int;
  59     long long int pa_long_long_int;
  60     unsigned short int pa_u_short_int;
  61     unsigned int pa_u_int;
  62     unsigned long int pa_u_long_int;
  63     unsigned long long int pa_u_long_long_int;
  64     float pa_float;
  65     double pa_double;
  66     long double pa_long_double;
  67     const char *pa_string;
  68     const wchar_t *pa_wstring;
  69     void *pa_pointer;
  70   };
  71
  72
  73 /* Read a simple integer from a string and update the string pointer.
  74    It is assumed that the first character is a digit.  */
  75 static inline unsigned int
  76 read_int (const UCHAR_T * *pstr)
  77 {
  78   unsigned int retval = **pstr - L_('0');
  79
  80   while (ISDIGIT (*++(*pstr)))
  81     {
  82       retval *= 10;
  83       retval += **pstr - L_('0');
  84     }
  85
  86   return retval;
  87 }
  88
  89
  90
  91 /* Find the next spec in FORMAT, or the end of the string.  Returns
  92    a pointer into FORMAT, to a '%' or a '\0'.  */
  93 static inline const UCHAR_T *
  94 #ifdef COMPILE_WPRINTF
  95 find_spec (const UCHAR_T *format)
  96 #else
  97 find_spec (const UCHAR_T *format, mbstate_t *ps)
  98 #endif
  99 {
 100 #ifdef COMPILE_WPRINTF
 101   return (const UCHAR_T *) __wcschrnul ((const CHAR_T *) format, L'%');
 102 #else
 103   while (*format != L_('\0') && *format != L_('%'))
 104     {
 105       int len;
 106
 107       /* Remove any hints of a wrong encoding.  */
 108       ps->__count = 0;
 109       if (! isascii (*format) && (len = mbrlen (format, MB_CUR_MAX, ps)) > 0)
 110         format += len;
 111       else
 112         ++format;
 113     }
 114   return format;
 115 #endif
 116 }
 117
 118
 119 /* These are defined in reg-printf.c.  */
 120 extern printf_arginfo_function *__printf_arginfo_table[];
 121 extern printf_function **__printf_function_table;
 122
 123
 124 /* FORMAT must point to a '%' at the beginning of a spec.  Fills in *SPEC
 125    with the parsed details.  POSN is the number of arguments already
 126    consumed.  At most MAXTYPES - POSN types are filled in TYPES.  Return
 127    the number of args consumed by this spec; *MAX_REF_ARG is updated so it
 128    remains the highest argument index used.  */
 129 static inline size_t
 130 #ifdef COMPILE_WPRINTF
 131 parse_one_spec (const UCHAR_T *format, size_t posn, struct printf_spec *spec,
 132                 size_t *max_ref_arg)
 133 #else
 134 parse_one_spec (const UCHAR_T *format, size_t posn, struct printf_spec *spec,
 135                 size_t *max_ref_arg, mbstate_t *ps)
 136 #endif
 137 {
 138   unsigned int n;
 139   size_t nargs = 0;
 140
 141   /* Skip the '%'.  */
 142   ++format;
 143
 144   /* Clear information structure.  */
 145   spec->data_arg = -1;
 146   spec->info.alt = 0;
 147   spec->info.space = 0;
 148   spec->info.left = 0;
 149   spec->info.showsign = 0;
 150   spec->info.group = 0;
 151   spec->info.pad = ' ';
 152   spec->info.wide = sizeof (UCHAR_T) > 1;
 153
 154   /* Test for positional argument.  */
 155   if (ISDIGIT (*format))
 156     {
 157       const UCHAR_T *begin = format;
 158
 159       n = read_int (&format);
 160
 161       if (n > 0 && *format == L_('$'))
 162         /* Is positional parameter.  */
 163         {
 164           ++format;             /* Skip the '$'.  */
 165           spec->data_arg = n - 1;
 166           *max_ref_arg = MAX (*max_ref_arg, n);
 167         }
 168       else
 169         /* Oops; that was actually the width and/or 0 padding flag.
 170            Step back and read it again.  */
 171         format = begin;
 172     }
 173
 174   /* Check for spec modifiers.  */
 175   while (*format == L_(' ') || *format == L_('+') || *format == L_('-') ||
 176          *format == L_('#') || *format == L_('0') || *format == L_('\''))
 177     switch (*format++)
 178       {
 179       case L_(' '):
 180         /* Output a space in place of a sign, when there is no sign.  */
 181         spec->info.space = 1;
 182         break;
 183       case L_('+'):
 184         /* Always output + or - for numbers.  */
 185         spec->info.showsign = 1;
 186         break;
 187       case L_('-'):
 188         /* Left-justify things.  */
 189         spec->info.left = 1;
 190         break;
 191       case L_('#'):
 192         /* Use the "alternate form":
 193            Hex has 0x or 0X, FP always has a decimal point.  */
 194         spec->info.alt = 1;
 195         break;
 196       case L_('0'):
 197         /* Pad with 0s.  */
 198         spec->info.pad = '0';
 199         break;
 200       case L_('\''):
 201         /* Show grouping in numbers if the locale information
 202            indicates any.  */
 203         spec->info.group = 1;
 204         break;
 205       case L_('I'):
 206         /* Use the internationalized form of the output.  Currently
 207            means to use the `outdigits' of the current locale.  */
 208         spec->info.i18n = 1;
 209         break;
 210       }
 211   if (spec->info.left)
 212     spec->info.pad = ' ';
 213
 214   /* Get the field width.  */
 215   spec->width_arg = -1;
 216   spec->info.width = 0;
 217   if (*format == L_('*'))
 218     {
 219       /* The field width is given in an argument.
 220          A negative field width indicates left justification.  */
 221       const UCHAR_T *begin = ++format;
 222
 223       if (ISDIGIT (*format))
 224         {
 225           /* The width argument might be found in a positional parameter.  */
 226           n = read_int (&format);
 227
 228           if (n > 0 && *format == L_('$'))
 229             {
 230               spec->width_arg = n - 1;
 231               *max_ref_arg = MAX (*max_ref_arg, n);
 232               ++format;         /* Skip '$'.  */
 233             }
 234         }
 235
 236       if (spec->width_arg < 0)
 237         {
 238           /* Not in a positional parameter.  Consume one argument.  */
 239           spec->width_arg = posn++;
 240           ++nargs;
 241           format = begin;       /* Step back and reread.  */
 242         }
 243     }
 244   else if (ISDIGIT (*format))
 245     /* Constant width specification.  */
 246     spec->info.width = read_int (&format);
 247
 248   /* Get the precision.  */
 249   spec->prec_arg = -1;
 250   /* -1 means none given; 0 means explicit 0.  */
 251   spec->info.prec = -1;
 252   if (*format == L_('.'))
 253     {
 254       ++format;
 255       if (*format == L_('*'))
 256         {
 257           /* The precision is given in an argument.  */
 258           const UCHAR_T *begin = ++format;
 259
 260           if (ISDIGIT (*format))
 261             {
 262               n = read_int (&format);
 263
 264               if (n > 0 && *format == L_('$'))
 265                 {
 266                   spec->prec_arg = n - 1;
 267                   *max_ref_arg = MAX (*max_ref_arg, n);
 268                   ++format;
 269                 }
 270             }
 271
 272           if (spec->prec_arg < 0)
 273             {
 274               /* Not in a positional parameter.  */
 275               spec->prec_arg = posn++;
 276               ++nargs;
 277               format = begin;
 278             }
 279         }
 280       else if (ISDIGIT (*format))
 281         spec->info.prec = read_int (&format);
 282       else
 283         /* "%.?" is treated like "%.0?".  */
 284         spec->info.prec = 0;
 285     }
 286
 287   /* Check for type modifiers.  */
 288   spec->info.is_long_double = 0;
 289   spec->info.is_short = 0;
 290   spec->info.is_long = 0;
 291   spec->info.is_char = 0;
 292
 293   switch (*format++)
 294     {
 295     case L_('h'):
 296       /* ints are short ints or chars.  */
 297       if (*format != L_('h'))
 298         spec->info.is_short = 1;
 299       else
 300         {
 301           ++format;
 302           spec->info.is_char = 1;
 303         }
 304       break;
 305     case L_('l'):
 306       /* ints are long ints.  */
 307       spec->info.is_long = 1;
 308       if (*format != L_('l'))
 309         break;
 310       ++format;
 311       /* FALLTHROUGH */
 312     case L_('L'):
 313       /* doubles are long doubles, and ints are long long ints.  */
 314     case L_('q'):
 315       /* 4.4 uses this for long long.  */
 316       spec->info.is_long_double = 1;
 317       break;
 318     case L_('z'):
 319     case L_('Z'):
 320       /* ints are size_ts.  */
 321       assert (sizeof (size_t) <= sizeof (unsigned long long int));
 322 #if LONG_MAX != LONG_LONG_MAX
 323       spec->info.is_long_double = sizeof (size_t) > sizeof (unsigned long int);
 324 #endif
 325       spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 326       break;
 327     case L_('t'):
 328       assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 329 #if LONG_MAX != LONG_LONG_MAX
 330       spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 331 #endif
 332       spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 333       break;
 334     case L_('j'):
 335       assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 336 #if LONG_MAX != LONG_LONG_MAX
 337       spec->info.is_long_double = (sizeof (uintmax_t)
 338                                    > sizeof (unsigned long int));
 339 #endif
 340       spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 341       break;
 342     default:
 343       /* Not a recognized modifier.  Backup.  */
 344       --format;
 345       break;
 346     }
 347
 348   /* Get the format specification.  */
 349   spec->info.spec = (wchar_t) *format++;
 350   if (__printf_function_table != NULL
 351       && spec->info.spec <= UCHAR_MAX
 352       && __printf_arginfo_table[spec->info.spec] != NULL)
 353     /* We don't try to get the types for all arguments if the format
 354        uses more than one.  The normal case is covered though.  */
 355     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 356       (&spec->info, 1, &spec->data_arg_type);
 357   else
 358     {
 359       /* Find the data argument types of a built-in spec.  */
 360       spec->ndata_args = 1;
 361
 362       switch (spec->info.spec)
 363         {
 364         case L'i':
 365         case L'd':
 366         case L'u':
 367         case L'o':
 368         case L'X':
 369         case L'x':
 370 #if LONG_MAX != LONG_LONG_MAX
 371           if (spec->info.is_long_double)
 372             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 373           else
 374 #endif
 375             if (spec->info.is_long)
 376               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 377             else if (spec->info.is_short)
 378               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 379             else if (spec->info.is_char)
 380               spec->data_arg_type = PA_CHAR;
 381             else
 382               spec->data_arg_type = PA_INT;
 383           break;
 384         case L'e':
 385         case L'E':
 386         case L'f':
 387         case L'g':
 388         case L'G':
 389         case L'a':
 390         case L'A':
 391           if (spec->info.is_long_double)
 392             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 393           else
 394             spec->data_arg_type = PA_DOUBLE;
 395           break;
 396         case L'c':
 397           spec->data_arg_type = PA_CHAR;
 398           break;
 399         case L'C':
 400           spec->data_arg_type = PA_WCHAR;
 401           break;
 402         case L's':
 403           spec->data_arg_type = PA_STRING;
 404           break;
 405         case L'S':
 406           spec->data_arg_type = PA_WSTRING;
 407           break;
 408         case L'p':
 409           spec->data_arg_type = PA_POINTER;
 410           break;
 411         case L'n':
 412           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 413           break;
 414
 415         case L'm':
 416         default:
 417           /* An unknown spec will consume no args.  */
 418           spec->ndata_args = 0;
 419           break;
 420         }
 421     }
 422
 423   if (spec->data_arg == -1 && spec->ndata_args > 0)
 424     {
 425       /* There are args consumed, but no positional spec.  Use the
 426          next sequential arg position.  */
 427       spec->data_arg = posn;
 428       nargs += spec->ndata_args;
 429     }
 430
 431   if (spec->info.spec == L'\0')
 432     /* Format ended before this spec was complete.  */
 433     spec->end_of_fmt = spec->next_fmt = format - 1;
 434   else
 435     {
 436       /* Find the next format spec.  */
 437       spec->end_of_fmt = format;
 438 #ifdef COMPILE_WPRINTF
 439       spec->next_fmt = find_spec (format);
 440 #else
 441       spec->next_fmt = find_spec (format, ps);
 442 #endif
 443     }
 444
 445   return nargs;
 446 }