stdio-common/printf-parse.h

   1 /* Internal header for parsing printf format strings.
   2    Copyright (C) 1995, 1996, 1997 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 <string.h>
  23 #include <stddef.h>
  24
  25 #define NDEBUG 1
  26 #include <assert.h>
  27
  28
  29 struct printf_spec
  30   {
  31     /* Information parsed from the format spec.  */
  32     struct printf_info info;
  33
  34     /* Pointers into the format string for the end of this format
  35        spec and the next (or to the end of the string if no more).  */
  36     const char *end_of_fmt, *next_fmt;
  37
  38     /* Position of arguments for precision and width, or -1 if `info' has
  39        the constant value.  */
  40     int prec_arg, width_arg;
  41
  42     int data_arg;               /* Position of data argument.  */
  43     int data_arg_type;          /* Type of first argument.  */
  44     /* Number of arguments consumed by this format specifier.  */
  45     size_t ndata_args;
  46   };
  47
  48
  49 /* The various kinds off arguments that can be passed to printf.  */
  50 union printf_arg
  51   {
  52     unsigned char pa_char;
  53     wchar_t pa_wchar;
  54     short int pa_short_int;
  55     int pa_int;
  56     long int pa_long_int;
  57     long long int pa_long_long_int;
  58     unsigned short int pa_u_short_int;
  59     unsigned int pa_u_int;
  60     unsigned long int pa_u_long_int;
  61     unsigned long long int pa_u_long_long_int;
  62     float pa_float;
  63     double pa_double;
  64     long double pa_long_double;
  65     const char *pa_string;
  66     const wchar_t *pa_wstring;
  67     void *pa_pointer;
  68   };
  69
  70
  71 /* Read a simple integer from a string and update the string pointer.
  72    It is assumed that the first character is a digit.  */
  73 static inline unsigned int
  74 read_int (const UCHAR_T * *pstr)
  75 {
  76   unsigned int retval = **pstr - L_('0');
  77
  78   while (ISDIGIT (*++(*pstr)))
  79     {
  80       retval *= 10;
  81       retval += **pstr - L_('0');
  82     }
  83
  84   return retval;
  85 }
  86
  87
  88
  89 /* Find the next spec in FORMAT, or the end of the string.  Returns
  90    a pointer into FORMAT, to a '%' or a '\0'.  */
  91 static inline const char *
  92 find_spec (const char *format, mbstate_t *ps)
  93 {
  94   while (*format != '\0' && *format != '%')
  95     {
  96       int len;
  97
  98       /* Remove any hints of a wrong encoding.  */
  99       ps->count = 0;
 100       if (isascii (*format) || (len = mbrlen (format, MB_CUR_MAX, ps)) <= 0)
 101         ++format;
 102       else
 103         format += len;
 104     }
 105   return format;
 106 }
 107
 108
 109 /* These are defined in reg-printf.c.  */
 110 extern printf_arginfo_function *__printf_arginfo_table[];
 111 extern printf_function **__printf_function_table;
 112
 113
 114 /* FORMAT must point to a '%' at the beginning of a spec.  Fills in *SPEC
 115    with the parsed details.  POSN is the number of arguments already
 116    consumed.  At most MAXTYPES - POSN types are filled in TYPES.  Return
 117    the number of args consumed by this spec; *MAX_REF_ARG is updated so it
 118    remains the highest argument index used.  */
 119 static inline size_t
 120 parse_one_spec (const UCHAR_T *format, size_t posn, struct printf_spec *spec,
 121                 size_t *max_ref_arg, mbstate_t *ps)
 122 {
 123   unsigned int n;
 124   size_t nargs = 0;
 125
 126   /* Skip the '%'.  */
 127   ++format;
 128
 129   /* Clear information structure.  */
 130   spec->data_arg = -1;
 131   spec->info.alt = 0;
 132   spec->info.space = 0;
 133   spec->info.left = 0;
 134   spec->info.showsign = 0;
 135   spec->info.group = 0;
 136   spec->info.pad = ' ';
 137
 138   /* Test for positional argument.  */
 139   if (ISDIGIT (*format))
 140     {
 141       const UCHAR_T *begin = format;
 142
 143       n = read_int (&format);
 144
 145       if (n > 0 && *format == L_('$'))
 146         /* Is positional parameter.  */
 147         {
 148           ++format;             /* Skip the '$'.  */
 149           spec->data_arg = n - 1;
 150           *max_ref_arg = MAX (*max_ref_arg, n);
 151         }
 152       else
 153         /* Oops; that was actually the width and/or 0 padding flag.
 154            Step back and read it again.  */
 155         format = begin;
 156     }
 157
 158   /* Check for spec modifiers.  */
 159   while (*format == L_(' ') || *format == L_('+') || *format == L_('-') ||
 160          *format == L_('#') || *format == L_('0') || *format == L_('\''))
 161     switch (*format++)
 162       {
 163       case L_(' '):
 164         /* Output a space in place of a sign, when there is no sign.  */
 165         spec->info.space = 1;
 166         break;
 167       case L_('+'):
 168         /* Always output + or - for numbers.  */
 169         spec->info.showsign = 1;
 170         break;
 171       case L_('-'):
 172         /* Left-justify things.  */
 173         spec->info.left = 1;
 174         break;
 175       case L_('#'):
 176         /* Use the "alternate form":
 177            Hex has 0x or 0X, FP always has a decimal point.  */
 178         spec->info.alt = 1;
 179         break;
 180       case L_('0'):
 181         /* Pad with 0s.  */
 182         spec->info.pad = '0';
 183         break;
 184       case L_('\''):
 185         /* Show grouping in numbers if the locale information
 186            indicates any.  */
 187         spec->info.group = 1;
 188         break;
 189       }
 190   if (spec->info.left)
 191     spec->info.pad = ' ';
 192
 193   /* Get the field width.  */
 194   spec->width_arg = -1;
 195   spec->info.width = 0;
 196   if (*format == L_('*'))
 197     {
 198       /* The field width is given in an argument.
 199          A negative field width indicates left justification.  */
 200       const UCHAR_T *begin = ++format;
 201
 202       if (ISDIGIT (*format))
 203         {
 204           /* The width argument might be found in a positional parameter.  */
 205           n = read_int (&format);
 206
 207           if (n > 0 && *format == L_('$'))
 208             {
 209               spec->width_arg = n - 1;
 210               *max_ref_arg = MAX (*max_ref_arg, n);
 211               ++format;         /* Skip '$'.  */
 212             }
 213         }
 214
 215       if (spec->width_arg < 0)
 216         {
 217           /* Not in a positional parameter.  Consume one argument.  */
 218           spec->width_arg = posn++;
 219           ++nargs;
 220           format = begin;       /* Step back and reread.  */
 221         }
 222     }
 223   else if (ISDIGIT (*format))
 224     /* Constant width specification.  */
 225     spec->info.width = read_int (&format);
 226
 227   /* Get the precision.  */
 228   spec->prec_arg = -1;
 229   /* -1 means none given; 0 means explicit 0.  */
 230   spec->info.prec = -1;
 231   if (*format == L_('.'))
 232     {
 233       ++format;
 234       if (*format == L_('*'))
 235         {
 236           /* The precision is given in an argument.  */
 237           const UCHAR_T *begin = ++format;
 238
 239           if (ISDIGIT (*format))
 240             {
 241               n = read_int (&format);
 242
 243               if (n > 0 && *format == L_('$'))
 244                 {
 245                   spec->prec_arg = n - 1;
 246                   *max_ref_arg = MAX (*max_ref_arg, n);
 247                   ++format;
 248                 }
 249             }
 250
 251           if (spec->prec_arg < 0)
 252             {
 253               /* Not in a positional parameter.  */
 254               spec->prec_arg = posn++;
 255               ++nargs;
 256               format = begin;
 257             }
 258         }
 259       else if (ISDIGIT (*format))
 260         spec->info.prec = read_int (&format);
 261       else
 262         /* "%.?" is treated like "%.0?".  */
 263         spec->info.prec = 0;
 264     }
 265
 266   /* Check for type modifiers.  */
 267 #define is_longlong is_long_double
 268   spec->info.is_long_double = 0;
 269   spec->info.is_short = 0;
 270   spec->info.is_long = 0;
 271
 272   if (*format == L_('h') || *format == L_('l') || *format == L_('L') ||
 273       *format == L_('Z') || *format == L_('q'))
 274     switch (*format++)
 275       {
 276       case L_('h'):
 277         /* int's are short int's.  */
 278         if (spec->info.is_short == 0)
 279           spec->info.is_short = 1;
 280         else
 281           spec->info.is_short = 2;
 282         break;
 283       case L_('l'):
 284         /* int's are long int's.  */
 285         spec->info.is_long = 1;
 286         if (*format != L_('l'))
 287           break;
 288         ++format;
 289         /* FALLTHROUGH */
 290       case L_('L'):
 291         /* double's are long double's, and int's are long long int's.  */
 292       case L_('q'):
 293         /* 4.4 uses this for long long.  */
 294         spec->info.is_long_double = 1;
 295         break;
 296       case L_('Z'):
 297         /* int's are size_t's.  */
 298         assert (sizeof(size_t) <= sizeof(unsigned long long int));
 299         spec->info.is_longlong = sizeof(size_t) > sizeof(unsigned long int);
 300         spec->info.is_long = sizeof(size_t) > sizeof(unsigned int);
 301         break;
 302       }
 303
 304   /* Get the format specification.  */
 305   spec->info.spec = (wchar_t) *format++;
 306   if (__printf_function_table != NULL
 307       && spec->info.spec <= UCHAR_MAX
 308       && __printf_arginfo_table[spec->info.spec] != NULL)
 309     /* We don't try to get the types for all arguments if the format
 310        uses more than one.  The normal case is covered though.  */
 311     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 312       (&spec->info, 1, &spec->data_arg_type);
 313   else
 314     {
 315       /* Find the data argument types of a built-in spec.  */
 316       spec->ndata_args = 1;
 317
 318       switch (spec->info.spec)
 319         {
 320         case L'i':
 321         case L'd':
 322         case L'u':
 323         case L'o':
 324         case L'X':
 325         case L'x':
 326           if (spec->info.is_longlong)
 327             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 328           else if (spec->info.is_long)
 329             spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 330           else if (spec->info.is_short)
 331             spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 332           else
 333             spec->data_arg_type = PA_INT;
 334           break;
 335         case L'e':
 336         case L'E':
 337         case L'f':
 338         case L'g':
 339         case L'G':
 340           if (spec->info.is_long_double)
 341             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 342           else
 343             spec->data_arg_type = PA_DOUBLE;
 344           break;
 345         case L'c':
 346           spec->data_arg_type = PA_CHAR;
 347           break;
 348         case L'C':
 349           spec->data_arg_type = PA_WCHAR;
 350           break;
 351         case L's':
 352           spec->data_arg_type = PA_STRING;
 353           break;
 354         case L'S':
 355           spec->data_arg_type = PA_WSTRING;
 356           break;
 357         case L'p':
 358           spec->data_arg_type = PA_POINTER;
 359           break;
 360         case L'n':
 361           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 362           break;
 363
 364         case L'm':
 365         default:
 366           /* An unknown spec will consume no args.  */
 367           spec->ndata_args = 0;
 368           break;
 369         }
 370     }
 371
 372   if (spec->data_arg == -1 && spec->ndata_args > 0)
 373     {
 374       /* There are args consumed, but no positional spec.  Use the
 375          next sequential arg position.  */
 376       spec->data_arg = posn;
 377       nargs += spec->ndata_args;
 378     }
 379
 380   if (spec->info.spec == L'\0')
 381     /* Format ended before this spec was complete.  */
 382     spec->end_of_fmt = spec->next_fmt = format - 1;
 383   else
 384     {
 385       /* Find the next format spec.  */
 386       spec->end_of_fmt = format;
 387       spec->next_fmt = find_spec (format, ps);
 388     }
 389
 390   return nargs;
 391 }