stdio-common/printf-parse.h

   1 /* Internal header for parsing printf format strings.
   2    Copyright (C) 1995, 1996, 1997, 1998, 1999 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
 153   /* Test for positional argument.  */
 154   if (ISDIGIT (*format))
 155     {
 156       const UCHAR_T *begin = format;
 157
 158       n = read_int (&format);
 159
 160       if (n > 0 && *format == L_('$'))
 161         /* Is positional parameter.  */
 162         {
 163           ++format;             /* Skip the '$'.  */
 164           spec->data_arg = n - 1;
 165           *max_ref_arg = MAX (*max_ref_arg, n);
 166         }
 167       else
 168         /* Oops; that was actually the width and/or 0 padding flag.
 169            Step back and read it again.  */
 170         format = begin;
 171     }
 172
 173   /* Check for spec modifiers.  */
 174   while (*format == L_(' ') || *format == L_('+') || *format == L_('-') ||
 175          *format == L_('#') || *format == L_('0') || *format == L_('\''))
 176     switch (*format++)
 177       {
 178       case L_(' '):
 179         /* Output a space in place of a sign, when there is no sign.  */
 180         spec->info.space = 1;
 181         break;
 182       case L_('+'):
 183         /* Always output + or - for numbers.  */
 184         spec->info.showsign = 1;
 185         break;
 186       case L_('-'):
 187         /* Left-justify things.  */
 188         spec->info.left = 1;
 189         break;
 190       case L_('#'):
 191         /* Use the "alternate form":
 192            Hex has 0x or 0X, FP always has a decimal point.  */
 193         spec->info.alt = 1;
 194         break;
 195       case L_('0'):
 196         /* Pad with 0s.  */
 197         spec->info.pad = '0';
 198         break;
 199       case L_('\''):
 200         /* Show grouping in numbers if the locale information
 201            indicates any.  */
 202         spec->info.group = 1;
 203         break;
 204       }
 205   if (spec->info.left)
 206     spec->info.pad = ' ';
 207
 208   /* Get the field width.  */
 209   spec->width_arg = -1;
 210   spec->info.width = 0;
 211   if (*format == L_('*'))
 212     {
 213       /* The field width is given in an argument.
 214          A negative field width indicates left justification.  */
 215       const UCHAR_T *begin = ++format;
 216
 217       if (ISDIGIT (*format))
 218         {
 219           /* The width argument might be found in a positional parameter.  */
 220           n = read_int (&format);
 221
 222           if (n > 0 && *format == L_('$'))
 223             {
 224               spec->width_arg = n - 1;
 225               *max_ref_arg = MAX (*max_ref_arg, n);
 226               ++format;         /* Skip '$'.  */
 227             }
 228         }
 229
 230       if (spec->width_arg < 0)
 231         {
 232           /* Not in a positional parameter.  Consume one argument.  */
 233           spec->width_arg = posn++;
 234           ++nargs;
 235           format = begin;       /* Step back and reread.  */
 236         }
 237     }
 238   else if (ISDIGIT (*format))
 239     /* Constant width specification.  */
 240     spec->info.width = read_int (&format);
 241
 242   /* Get the precision.  */
 243   spec->prec_arg = -1;
 244   /* -1 means none given; 0 means explicit 0.  */
 245   spec->info.prec = -1;
 246   if (*format == L_('.'))
 247     {
 248       ++format;
 249       if (*format == L_('*'))
 250         {
 251           /* The precision is given in an argument.  */
 252           const UCHAR_T *begin = ++format;
 253
 254           if (ISDIGIT (*format))
 255             {
 256               n = read_int (&format);
 257
 258               if (n > 0 && *format == L_('$'))
 259                 {
 260                   spec->prec_arg = n - 1;
 261                   *max_ref_arg = MAX (*max_ref_arg, n);
 262                   ++format;
 263                 }
 264             }
 265
 266           if (spec->prec_arg < 0)
 267             {
 268               /* Not in a positional parameter.  */
 269               spec->prec_arg = posn++;
 270               ++nargs;
 271               format = begin;
 272             }
 273         }
 274       else if (ISDIGIT (*format))
 275         spec->info.prec = read_int (&format);
 276       else
 277         /* "%.?" is treated like "%.0?".  */
 278         spec->info.prec = 0;
 279     }
 280
 281   /* Check for type modifiers.  */
 282   spec->info.is_long_double = 0;
 283   spec->info.is_short = 0;
 284   spec->info.is_long = 0;
 285   spec->info.is_char = 0;
 286
 287   switch (*format++)
 288     {
 289     case L_('h'):
 290       /* ints are short ints or chars.  */
 291       if (*format != L_('h'))
 292         spec->info.is_short = 1;
 293       else
 294         {
 295           ++format;
 296           spec->info.is_char = 1;
 297         }
 298       break;
 299     case L_('l'):
 300       /* ints are long ints.  */
 301       spec->info.is_long = 1;
 302       if (*format != L_('l'))
 303         break;
 304       ++format;
 305       /* FALLTHROUGH */
 306     case L_('L'):
 307       /* doubles are long doubles, and ints are long long ints.  */
 308     case L_('q'):
 309       /* 4.4 uses this for long long.  */
 310       spec->info.is_long_double = 1;
 311       break;
 312     case L_('z'):
 313     case L_('Z'):
 314       /* ints are size_ts.  */
 315       assert (sizeof (size_t) <= sizeof (unsigned long long int));
 316 #if LONG_MAX != LONG_LONG_MAX
 317       spec->info.is_long_double = sizeof (size_t) > sizeof (unsigned long int);
 318 #endif
 319       spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 320       break;
 321     case L_('t'):
 322       assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 323 #if LONG_MAX != LONG_LONG_MAX
 324       spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 325 #endif
 326       spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 327       break;
 328     case L_('j'):
 329       assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 330 #if LONG_MAX != LONG_LONG_MAX
 331       spec->info.is_long_double = (sizeof (uintmax_t)
 332                                    > sizeof (unsigned long int));
 333 #endif
 334       spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 335       break;
 336     default:
 337       /* Not a recognized modifier.  Backup.  */
 338       --format;
 339       break;
 340     }
 341
 342   /* Get the format specification.  */
 343   spec->info.spec = (wchar_t) *format++;
 344   if (__printf_function_table != NULL
 345       && spec->info.spec <= UCHAR_MAX
 346       && __printf_arginfo_table[spec->info.spec] != NULL)
 347     /* We don't try to get the types for all arguments if the format
 348        uses more than one.  The normal case is covered though.  */
 349     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 350       (&spec->info, 1, &spec->data_arg_type);
 351   else
 352     {
 353       /* Find the data argument types of a built-in spec.  */
 354       spec->ndata_args = 1;
 355
 356       switch (spec->info.spec)
 357         {
 358         case L'i':
 359         case L'd':
 360         case L'u':
 361         case L'o':
 362         case L'X':
 363         case L'x':
 364 #if LONG_MAX != LONG_LONG_MAX
 365           if (spec->info.is_long_double)
 366             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 367           else
 368 #endif
 369             if (spec->info.is_long)
 370               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 371             else if (spec->info.is_short)
 372               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 373             else if (spec->info.is_char)
 374               spec->data_arg_type = PA_CHAR;
 375             else
 376               spec->data_arg_type = PA_INT;
 377           break;
 378         case L'e':
 379         case L'E':
 380         case L'f':
 381         case L'g':
 382         case L'G':
 383         case L'a':
 384         case L'A':
 385           if (spec->info.is_long_double)
 386             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 387           else
 388             spec->data_arg_type = PA_DOUBLE;
 389           break;
 390         case L'c':
 391           spec->data_arg_type = PA_CHAR;
 392           break;
 393         case L'C':
 394           spec->data_arg_type = PA_WCHAR;
 395           break;
 396         case L's':
 397           spec->data_arg_type = PA_STRING;
 398           break;
 399         case L'S':
 400           spec->data_arg_type = PA_WSTRING;
 401           break;
 402         case L'p':
 403           spec->data_arg_type = PA_POINTER;
 404           break;
 405         case L'n':
 406           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 407           break;
 408
 409         case L'm':
 410         default:
 411           /* An unknown spec will consume no args.  */
 412           spec->ndata_args = 0;
 413           break;
 414         }
 415     }
 416
 417   if (spec->data_arg == -1 && spec->ndata_args > 0)
 418     {
 419       /* There are args consumed, but no positional spec.  Use the
 420          next sequential arg position.  */
 421       spec->data_arg = posn;
 422       nargs += spec->ndata_args;
 423     }
 424
 425   if (spec->info.spec == L'\0')
 426     /* Format ended before this spec was complete.  */
 427     spec->end_of_fmt = spec->next_fmt = format - 1;
 428   else
 429     {
 430       /* Find the next format spec.  */
 431       spec->end_of_fmt = format;
 432 #ifdef COMPILE_WPRINTF
 433       spec->next_fmt = find_spec (format);
 434 #else
 435       spec->next_fmt = find_spec (format, ps);
 436 #endif
 437     }
 438
 439   return nargs;
 440 }