stdio-common/printf-parse.h

   1 /* Internal header for parsing printf format strings.
   2 Copyright (C) 1995, 1996 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
  17 not, write to the Free Software Foundation, Inc., 675 Mass Ave,
  18 Cambridge, MA 02139, USA.  */
  19
  20 #include <ctype.h>
  21 #include <printf.h>
  22 #include <string.h>
  23 #include <stddef.h>
  24
  25 #define NDEBUG 1
  26 #include <assert.h>
  27
  28 #define MAX(a,b)        ({typeof(a) _a = (a); typeof(b) _b = (b);             \
  29                           _a > _b ? _a : _b; })
  30 #define MIN(a,b)        ({typeof(a) _a = (a); typeof(b) _b = (b);             \
  31                           _a < _b ? _a : _b; })
  32
  33 struct printf_spec
  34   {
  35     /* Information parsed from the format spec.  */
  36     struct printf_info info;
  37
  38     /* Pointers into the format string for the end of this format
  39        spec and the next (or to the end of the string if no more).  */
  40     const char *end_of_fmt, *next_fmt;
  41
  42     /* Position of arguments for precision and width, or -1 if `info' has
  43        the constant value.  */
  44     int prec_arg, width_arg;
  45
  46     int data_arg;               /* Position of data argument.  */
  47     int data_arg_type;          /* Type of first argument.  */
  48     /* Number of arguments consumed by this format specifier.  */
  49     size_t ndata_args;
  50   };
  51
  52
  53 /* The various kinds off arguments that can be passed to printf.  */
  54 union printf_arg
  55   {
  56     unsigned char pa_char;
  57     short int pa_short_int;
  58     int pa_int;
  59     long int pa_long_int;
  60     long long int pa_long_long_int;
  61     unsigned short int pa_u_short_int;
  62     unsigned int pa_u_int;
  63     unsigned long int pa_u_long_int;
  64     unsigned long long int pa_u_long_long_int;
  65     float pa_float;
  66     double pa_double;
  67     long double pa_long_double;
  68     const char *pa_string;
  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 char *
  94 find_spec (const char *format, mbstate_t *ps)
  95 {
  96   while (*format != '\0' && *format != '%')
  97     {
  98       int len;
  99
 100       /* Remove any hints of a wrong encoding.  */
 101       ps->count = 0;
 102       if (isascii (*format) || (len = mbrlen (format, MB_CUR_MAX, ps)) <= 0)
 103         ++format;
 104       else
 105         format += len;
 106     }
 107   return format;
 108 }
 109
 110
 111 /* This is defined in reg-printf.c.  */
 112 extern printf_arginfo_function **__printf_arginfo_table;
 113
 114
 115 /* FORMAT must point to a '%' at the beginning of a spec.  Fills in *SPEC
 116    with the parsed details.  POSN is the number of arguments already
 117    consumed.  At most MAXTYPES - POSN types are filled in TYPES.  Return
 118    the number of args consumed by this spec; *MAX_REF_ARG is updated so it
 119    remains the highest argument index used.  */
 120 static inline size_t
 121 parse_one_spec (const UCHAR_T *format, size_t posn, struct printf_spec *spec,
 122                 size_t *max_ref_arg, mbstate_t *ps)
 123 {
 124   unsigned int n;
 125   size_t nargs = 0;
 126
 127   /* Skip the '%'.  */
 128   ++format;
 129
 130   /* Clear information structure.  */
 131   spec->data_arg = -1;
 132   spec->info.alt = 0;
 133   spec->info.space = 0;
 134   spec->info.left = 0;
 135   spec->info.showsign = 0;
 136   spec->info.group = 0;
 137   spec->info.pad = ' ';
 138
 139   /* Test for positional argument.  */
 140   if (ISDIGIT (*format))
 141     {
 142       const UCHAR_T *begin = format;
 143
 144       n = read_int (&format);
 145
 146       if (n > 0 && *format == L_('$'))
 147         /* Is positional parameter.  */
 148         {
 149           ++format;             /* Skip the '$'.  */
 150           spec->data_arg = n - 1;
 151           *max_ref_arg = MAX (*max_ref_arg, n);
 152         }
 153       else
 154         /* Oops; that was actually the width and/or 0 padding flag.
 155            Step back and read it again.  */
 156         format = begin;
 157     }
 158
 159   /* Check for spec modifiers.  */
 160   while (*format == L_(' ') || *format == L_('+') || *format == L_('-') ||
 161          *format == L_('#') || *format == L_('0') || *format == L_('\''))
 162     switch (*format++)
 163       {
 164       case L_(' '):
 165         /* Output a space in place of a sign, when there is no sign.  */
 166         spec->info.space = 1;
 167         break;
 168       case L_('+'):
 169         /* Always output + or - for numbers.  */
 170         spec->info.showsign = 1;
 171         break;
 172       case L_('-'):
 173         /* Left-justify things.  */
 174         spec->info.left = 1;
 175         break;
 176       case L_('#'):
 177         /* Use the "alternate form":
 178            Hex has 0x or 0X, FP always has a decimal point.  */
 179         spec->info.alt = 1;
 180         break;
 181       case L_('0'):
 182         /* Pad with 0s.  */
 183         spec->info.pad = '0';
 184         break;
 185       case L_('\''):
 186         /* Show grouping in numbers if the locale information
 187            indicates any.  */
 188         spec->info.group = 1;
 189         break;
 190       }
 191   if (spec->info.left)
 192     spec->info.pad = ' ';
 193
 194   /* Get the field width.  */
 195   spec->width_arg = -1;
 196   spec->info.width = 0;
 197   if (*format == L_('*'))
 198     {
 199       /* The field width is given in an argument.
 200          A negative field width indicates left justification.  */
 201       const UCHAR_T *begin = ++format;
 202
 203       if (ISDIGIT (*format))
 204         {
 205           /* The width argument might be found in a positional parameter.  */
 206           n = read_int (&format);
 207
 208           if (n > 0 && *format == L_('$'))
 209             {
 210               spec->width_arg = n - 1;
 211               *max_ref_arg = MAX (*max_ref_arg, n);
 212               ++format;         /* Skip '$'.  */
 213             }
 214         }
 215
 216       if (spec->width_arg < 0)
 217         {
 218           /* Not in a positional parameter.  Consume one argument.  */
 219           spec->width_arg = posn++;
 220           ++nargs;
 221           format = begin;       /* Step back and reread.  */
 222         }
 223     }
 224   else if (ISDIGIT (*format))
 225     /* Constant width specification.  */
 226     spec->info.width = read_int (&format);
 227
 228   /* Get the precision.  */
 229   spec->prec_arg = -1;
 230   /* -1 means none given; 0 means explicit 0.  */
 231   spec->info.prec = -1;
 232   if (*format == L_('.'))
 233     {
 234       ++format;
 235       if (*format == L_('*'))
 236         {
 237           /* The precision is given in an argument.  */
 238           const UCHAR_T *begin = ++format;
 239
 240           if (ISDIGIT (*format))
 241             {
 242               n = read_int (&format);
 243
 244               if (n > 0 && *format == L_('$'))
 245                 {
 246                   spec->prec_arg = n - 1;
 247                   *max_ref_arg = MAX (*max_ref_arg, n);
 248                   ++format;
 249                 }
 250             }
 251
 252           if (spec->prec_arg < 0)
 253             {
 254               /* Not in a positional parameter.  */
 255               spec->prec_arg = posn++;
 256               ++nargs;
 257               format = begin;
 258             }
 259         }
 260       else if (ISDIGIT (*format))
 261         spec->info.prec = read_int (&format);
 262       else
 263         /* "%.?" is treated like "%.0?".  */
 264         spec->info.prec = 0;
 265     }
 266
 267   /* Check for type modifiers.  */
 268 #define is_longlong is_long_double
 269   spec->info.is_long_double = 0;
 270   spec->info.is_short = 0;
 271   spec->info.is_long = 0;
 272
 273   if (*format == L_('h') || *format == L_('l') || *format == L_('L') ||
 274       *format == L_('Z') || *format == L_('q'))
 275     switch (*format++)
 276       {
 277       case L_('h'):
 278         /* int's are short int's.  */
 279         spec->info.is_short = 1;
 280         break;
 281       case L_('l'):
 282         /* int's are long int's.  */
 283         spec->info.is_long = 1;
 284         if (*format != L_('l'))
 285           break;
 286         ++format;
 287         /* FALLTHROUGH */
 288       case L_('L'):
 289         /* double's are long double's, and int's are long long int's.  */
 290       case L_('q'):
 291         /* 4.4 uses this for long long.  */
 292         spec->info.is_long_double = 1;
 293         break;
 294       case L_('Z'):
 295         /* int's are size_t's.  */
 296         assert (sizeof(size_t) <= sizeof(unsigned long long int));
 297         spec->info.is_longlong = sizeof(size_t) > sizeof(unsigned long int);
 298         spec->info.is_long = sizeof(size_t) > sizeof(unsigned int);
 299         break;
 300       }
 301
 302   /* Get the format specification.  */
 303 #ifdef THIS_IS_INCOMPATIBLE_WITH_LINUX_LIBC
 304   spec->info.spec = (wchar_t) *format++;
 305 #else
 306   spec->info.spec = (char) *format++;
 307 #endif
 308   if (__printf_arginfo_table != NULL &&
 309       __printf_arginfo_table[spec->info.spec] != NULL)
 310     /* We don't try to get the types for all arguments if the format
 311        uses more than one.  The normal case is covered though.  */
 312     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 313       (&spec->info, 1, &spec->data_arg_type);
 314   else
 315     {
 316       /* Find the data argument types of a built-in spec.  */
 317       spec->ndata_args = 1;
 318
 319       switch (spec->info.spec)
 320         {
 321         case L'i':
 322         case L'd':
 323         case L'u':
 324         case L'o':
 325         case L'X':
 326         case L'x':
 327           if (spec->info.is_longlong)
 328             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 329           else if (spec->info.is_long)
 330             spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 331           else if (spec->info.is_short)
 332             spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 333           else
 334             spec->data_arg_type = PA_INT;
 335           break;
 336         case L'e':
 337         case L'E':
 338         case L'f':
 339         case L'g':
 340         case L'G':
 341           if (spec->info.is_long_double)
 342             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 343           else
 344             spec->data_arg_type = PA_DOUBLE;
 345           break;
 346         case L'c':
 347           spec->data_arg_type = PA_CHAR;
 348           break;
 349         case L's':
 350           spec->data_arg_type = PA_STRING;
 351           break;
 352         case L'p':
 353           spec->data_arg_type = PA_POINTER;
 354           break;
 355         case L'n':
 356           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 357           break;
 358
 359         case L'm':
 360         default:
 361           /* An unknown spec will consume no args.  */
 362           spec->ndata_args = 0;
 363           break;
 364         }
 365
 366       if (spec->data_arg == -1 && spec->ndata_args > 0)
 367         {
 368           /* There are args consumed, but no positional spec.
 369              Use the next sequential arg position.  */
 370           spec->data_arg = posn;
 371           posn += spec->ndata_args;
 372           nargs += spec->ndata_args;
 373         }
 374     }
 375
 376   if (spec->info.spec == L'\0')
 377     /* Format ended before this spec was complete.  */
 378     spec->end_of_fmt = spec->next_fmt = format - 1;
 379   else
 380     {
 381       /* Find the next format spec.  */
 382       spec->end_of_fmt = format;
 383       spec->next_fmt = find_spec (format, ps);
 384     }
 385
 386   return nargs;
 387 }