stdio-common/printf-parse.h

   1 /* Internal header for parsing printf format strings.
   2    Copyright (C) 1995, 1996, 1997, 1998 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 <printf.h>
  22 #include <stdint.h>
  23 #include <stddef.h>
  24 #include <string.h>
  25
  26 #define NDEBUG 1
  27 #include <assert.h>
  28
  29
  30 struct printf_spec
  31   {
  32     /* Information parsed from the format spec.  */
  33     struct printf_info info;
  34
  35     /* Pointers into the format string for the end of this format
  36        spec and the next (or to the end of the string if no more).  */
  37     const char *end_of_fmt, *next_fmt;
  38
  39     /* Position of arguments for precision and width, or -1 if `info' has
  40        the constant value.  */
  41     int prec_arg, width_arg;
  42
  43     int data_arg;               /* Position of data argument.  */
  44     int data_arg_type;          /* Type of first argument.  */
  45     /* Number of arguments consumed by this format specifier.  */
  46     size_t ndata_args;
  47   };
  48
  49
  50 /* The various kinds off arguments that can be passed to printf.  */
  51 union printf_arg
  52   {
  53     unsigned char pa_char;
  54     wchar_t pa_wchar;
  55     short int pa_short_int;
  56     int pa_int;
  57     long int pa_long_int;
  58     long long int pa_long_long_int;
  59     unsigned short int pa_u_short_int;
  60     unsigned int pa_u_int;
  61     unsigned long int pa_u_long_int;
  62     unsigned long long int pa_u_long_long_int;
  63     float pa_float;
  64     double pa_double;
  65     long double pa_long_double;
  66     const char *pa_string;
  67     const wchar_t *pa_wstring;
  68     void *pa_pointer;
  69   };
  70
  71
  72 /* Read a simple integer from a string and update the string pointer.
  73    It is assumed that the first character is a digit.  */
  74 static inline unsigned int
  75 read_int (const UCHAR_T * *pstr)
  76 {
  77   unsigned int retval = **pstr - L_('0');
  78
  79   while (ISDIGIT (*++(*pstr)))
  80     {
  81       retval *= 10;
  82       retval += **pstr - L_('0');
  83     }
  84
  85   return retval;
  86 }
  87
  88
  89
  90 /* Find the next spec in FORMAT, or the end of the string.  Returns
  91    a pointer into FORMAT, to a '%' or a '\0'.  */
  92 static inline const char *
  93 find_spec (const char *format, mbstate_t *ps)
  94 {
  95   while (*format != '\0' && *format != '%')
  96     {
  97       int len;
  98
  99       /* Remove any hints of a wrong encoding.  */
 100       ps->count = 0;
 101       if (isascii (*format) || (len = mbrlen (format, MB_CUR_MAX, ps)) <= 0)
 102         ++format;
 103       else
 104         format += len;
 105     }
 106   return format;
 107 }
 108
 109
 110 /* These are defined in reg-printf.c.  */
 111 extern printf_arginfo_function *__printf_arginfo_table[];
 112 extern printf_function **__printf_function_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   spec->info.is_char = 0;
 273
 274   switch (*format++)
 275     {
 276     case L_('h'):
 277       /* ints are short ints or chars.  */
 278       if (*format != L_('h'))
 279         spec->info.is_short = 1;
 280       else
 281         {
 282           ++format;
 283           spec->info.is_char = 1;
 284         }
 285       break;
 286     case L_('l'):
 287       /* ints are long ints.  */
 288       spec->info.is_long = 1;
 289       if (*format != L_('l'))
 290         break;
 291       ++format;
 292       /* FALLTHROUGH */
 293     case L_('L'):
 294       /* doubles are long doubles, and ints are long long ints.  */
 295     case L_('q'):
 296       /* 4.4 uses this for long long.  */
 297       spec->info.is_long_double = 1;
 298       break;
 299     case L_('z'):
 300     case L_('Z'):
 301       /* ints are size_ts.  */
 302       assert (sizeof (size_t) <= sizeof (unsigned long long int));
 303       spec->info.is_longlong = sizeof (size_t) > sizeof (unsigned long int);
 304       spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 305       break;
 306     case L_('t'):
 307       assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 308       spec->info.is_longlong = (sizeof (ptrdiff_t) > sizeof (long int));
 309       spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 310       break;
 311     case L_('j'):
 312       assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 313       spec->info.is_longlong = (sizeof (uintmax_t)
 314                                 > sizeof (unsigned long int));
 315       spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 316       break;
 317     default:
 318       /* Not a recognized modifier.  Backup.  */
 319       --format;
 320       break;
 321     }
 322
 323   /* Get the format specification.  */
 324   spec->info.spec = (wchar_t) *format++;
 325   if (__printf_function_table != NULL
 326       && spec->info.spec <= UCHAR_MAX
 327       && __printf_arginfo_table[spec->info.spec] != NULL)
 328     /* We don't try to get the types for all arguments if the format
 329        uses more than one.  The normal case is covered though.  */
 330     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 331       (&spec->info, 1, &spec->data_arg_type);
 332   else
 333     {
 334       /* Find the data argument types of a built-in spec.  */
 335       spec->ndata_args = 1;
 336
 337       switch (spec->info.spec)
 338         {
 339         case L'i':
 340         case L'd':
 341         case L'u':
 342         case L'o':
 343         case L'X':
 344         case L'x':
 345           if (spec->info.is_longlong)
 346             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 347           else if (spec->info.is_long)
 348             spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 349           else if (spec->info.is_short)
 350             spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 351           else if (spec->info.is_char)
 352             spec->data_arg_type = PA_CHAR;
 353           else
 354             spec->data_arg_type = PA_INT;
 355           break;
 356         case L'e':
 357         case L'E':
 358         case L'f':
 359         case L'g':
 360         case L'G':
 361           if (spec->info.is_long_double)
 362             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 363           else
 364             spec->data_arg_type = PA_DOUBLE;
 365           break;
 366         case L'c':
 367           spec->data_arg_type = PA_CHAR;
 368           break;
 369         case L'C':
 370           spec->data_arg_type = PA_WCHAR;
 371           break;
 372         case L's':
 373           spec->data_arg_type = PA_STRING;
 374           break;
 375         case L'S':
 376           spec->data_arg_type = PA_WSTRING;
 377           break;
 378         case L'p':
 379           spec->data_arg_type = PA_POINTER;
 380           break;
 381         case L'n':
 382           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 383           break;
 384
 385         case L'm':
 386         default:
 387           /* An unknown spec will consume no args.  */
 388           spec->ndata_args = 0;
 389           break;
 390         }
 391     }
 392
 393   if (spec->data_arg == -1 && spec->ndata_args > 0)
 394     {
 395       /* There are args consumed, but no positional spec.  Use the
 396          next sequential arg position.  */
 397       spec->data_arg = posn;
 398       nargs += spec->ndata_args;
 399     }
 400
 401   if (spec->info.spec == L'\0')
 402     /* Format ended before this spec was complete.  */
 403     spec->end_of_fmt = spec->next_fmt = format - 1;
 404   else
 405     {
 406       /* Find the next format spec.  */
 407       spec->end_of_fmt = format;
 408       spec->next_fmt = find_spec (format, ps);
 409     }
 410
 411   return nargs;
 412 }