stdio-common/printf-parsemb.c

   1 /* Helper functions for parsing printf format strings.
   2    Copyright (C) 1995-2000,2002,2003,2004,2006 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 Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the 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    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, write to the Free
  17    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  18    02111-1307 USA.  */
  19
  20 #include <ctype.h>
  21 #include <limits.h>
  22 #include <stdlib.h>
  23 #include <string.h>
  24 #include <sys/param.h>
  25 #include <wchar.h>
  26 #include <wctype.h>
  27
  28 #ifndef COMPILE_WPRINTF
  29 # define CHAR_T         char
  30 # define UCHAR_T        unsigned char
  31 # define INT_T          int
  32 # define L_(Str)        Str
  33 # define ISDIGIT(Ch)    isdigit (Ch)
  34 #else
  35 # define CHAR_T         wchar_t
  36 # define UCHAR_T        unsigned int
  37 # define INT_T          wint_t
  38 # define L_(Str)        L##Str
  39 # define ISDIGIT(Ch)    iswdigit (Ch)
  40 #endif
  41
  42 #include "printf-parse.h"
  43
  44 #define NDEBUG 1
  45 #include <assert.h>
  46
  47
  48
  49 /* Find the next spec in FORMAT, or the end of the string.  Returns
  50    a pointer into FORMAT, to a '%' or a '\0'.  */
  51 const UCHAR_T *
  52 #ifdef COMPILE_WPRINTF
  53 __find_specwc (const UCHAR_T *format)
  54 #else
  55 __find_specmb (const UCHAR_T *format, mbstate_t *ps)
  56 #endif
  57 {
  58 #ifdef COMPILE_WPRINTF
  59   return (const UCHAR_T *) __wcschrnul ((const CHAR_T *) format, L'%');
  60 #else
  61   while (*format != L_('\0') && *format != L_('%'))
  62     {
  63       int len;
  64
  65       /* Remove any hints of a wrong encoding.  */
  66       ps->__count = 0;
  67       if (! isascii (*format)
  68           && (len = __mbrlen ((const CHAR_T *) format, MB_CUR_MAX, ps)) > 0)
  69         format += len;
  70       else
  71         ++format;
  72     }
  73   return format;
  74 #endif
  75 }
  76
  77
  78 /* FORMAT must point to a '%' at the beginning of a spec.  Fills in *SPEC
  79    with the parsed details.  POSN is the number of arguments already
  80    consumed.  At most MAXTYPES - POSN types are filled in TYPES.  Return
  81    the number of args consumed by this spec; *MAX_REF_ARG is updated so it
  82    remains the highest argument index used.  */
  83 size_t
  84 attribute_hidden
  85 #ifdef COMPILE_WPRINTF
  86 __parse_one_specwc (const UCHAR_T *format, size_t posn,
  87                     struct printf_spec *spec, size_t *max_ref_arg)
  88 #else
  89 __parse_one_specmb (const UCHAR_T *format, size_t posn,
  90                     struct printf_spec *spec, size_t *max_ref_arg,
  91                     mbstate_t *ps)
  92 #endif
  93 {
  94   unsigned int n;
  95   size_t nargs = 0;
  96
  97   /* Skip the '%'.  */
  98   ++format;
  99
 100   /* Clear information structure.  */
 101   spec->data_arg = -1;
 102   spec->info.alt = 0;
 103   spec->info.space = 0;
 104   spec->info.left = 0;
 105   spec->info.showsign = 0;
 106   spec->info.group = 0;
 107   spec->info.i18n = 0;
 108   spec->info.extra = 0;
 109   spec->info.pad = ' ';
 110   spec->info.wide = sizeof (UCHAR_T) > 1;
 111
 112   /* Test for positional argument.  */
 113   if (ISDIGIT (*format))
 114     {
 115       const UCHAR_T *begin = format;
 116
 117       n = read_int (&format);
 118
 119       if (n > 0 && *format == L_('$'))
 120         /* Is positional parameter.  */
 121         {
 122           ++format;             /* Skip the '$'.  */
 123           spec->data_arg = n - 1;
 124           *max_ref_arg = MAX (*max_ref_arg, n);
 125         }
 126       else
 127         /* Oops; that was actually the width and/or 0 padding flag.
 128            Step back and read it again.  */
 129         format = begin;
 130     }
 131
 132   /* Check for spec modifiers.  */
 133   do
 134     {
 135       switch (*format)
 136         {
 137         case L_(' '):
 138           /* Output a space in place of a sign, when there is no sign.  */
 139           spec->info.space = 1;
 140           continue;
 141         case L_('+'):
 142           /* Always output + or - for numbers.  */
 143           spec->info.showsign = 1;
 144           continue;
 145         case L_('-'):
 146           /* Left-justify things.  */
 147           spec->info.left = 1;
 148           continue;
 149         case L_('#'):
 150           /* Use the "alternate form":
 151              Hex has 0x or 0X, FP always has a decimal point.  */
 152           spec->info.alt = 1;
 153           continue;
 154         case L_('0'):
 155           /* Pad with 0s.  */
 156           spec->info.pad = '0';
 157           continue;
 158         case L_('\''):
 159           /* Show grouping in numbers if the locale information
 160              indicates any.  */
 161           spec->info.group = 1;
 162           continue;
 163         case L_('I'):
 164           /* Use the internationalized form of the output.  Currently
 165              means to use the `outdigits' of the current locale.  */
 166           spec->info.i18n = 1;
 167           continue;
 168         default:
 169           break;
 170         }
 171       break;
 172     }
 173   while (*++format);
 174
 175   if (spec->info.left)
 176     spec->info.pad = ' ';
 177
 178   /* Get the field width.  */
 179   spec->width_arg = -1;
 180   spec->info.width = 0;
 181   if (*format == L_('*'))
 182     {
 183       /* The field width is given in an argument.
 184          A negative field width indicates left justification.  */
 185       const UCHAR_T *begin = ++format;
 186
 187       if (ISDIGIT (*format))
 188         {
 189           /* The width argument might be found in a positional parameter.  */
 190           n = read_int (&format);
 191
 192           if (n > 0 && *format == L_('$'))
 193             {
 194               spec->width_arg = n - 1;
 195               *max_ref_arg = MAX (*max_ref_arg, n);
 196               ++format;         /* Skip '$'.  */
 197             }
 198         }
 199
 200       if (spec->width_arg < 0)
 201         {
 202           /* Not in a positional parameter.  Consume one argument.  */
 203           spec->width_arg = posn++;
 204           ++nargs;
 205           format = begin;       /* Step back and reread.  */
 206         }
 207     }
 208   else if (ISDIGIT (*format))
 209     /* Constant width specification.  */
 210     spec->info.width = read_int (&format);
 211
 212   /* Get the precision.  */
 213   spec->prec_arg = -1;
 214   /* -1 means none given; 0 means explicit 0.  */
 215   spec->info.prec = -1;
 216   if (*format == L_('.'))
 217     {
 218       ++format;
 219       if (*format == L_('*'))
 220         {
 221           /* The precision is given in an argument.  */
 222           const UCHAR_T *begin = ++format;
 223
 224           if (ISDIGIT (*format))
 225             {
 226               n = read_int (&format);
 227
 228               if (n > 0 && *format == L_('$'))
 229                 {
 230                   spec->prec_arg = n - 1;
 231                   *max_ref_arg = MAX (*max_ref_arg, n);
 232                   ++format;
 233                 }
 234             }
 235
 236           if (spec->prec_arg < 0)
 237             {
 238               /* Not in a positional parameter.  */
 239               spec->prec_arg = posn++;
 240               ++nargs;
 241               format = begin;
 242             }
 243         }
 244       else if (ISDIGIT (*format))
 245         spec->info.prec = read_int (&format);
 246       else
 247         /* "%.?" is treated like "%.0?".  */
 248         spec->info.prec = 0;
 249     }
 250
 251   /* Check for type modifiers.  */
 252   spec->info.is_long_double = 0;
 253   spec->info.is_short = 0;
 254   spec->info.is_long = 0;
 255   spec->info.is_char = 0;
 256
 257   switch (*format++)
 258     {
 259     case L_('h'):
 260       /* ints are short ints or chars.  */
 261       if (*format != L_('h'))
 262         spec->info.is_short = 1;
 263       else
 264         {
 265           ++format;
 266           spec->info.is_char = 1;
 267         }
 268       break;
 269     case L_('l'):
 270       /* ints are long ints.  */
 271       spec->info.is_long = 1;
 272       if (*format != L_('l'))
 273         break;
 274       ++format;
 275       /* FALLTHROUGH */
 276     case L_('L'):
 277       /* doubles are long doubles, and ints are long long ints.  */
 278     case L_('q'):
 279       /* 4.4 uses this for long long.  */
 280       spec->info.is_long_double = 1;
 281       break;
 282     case L_('z'):
 283     case L_('Z'):
 284       /* ints are size_ts.  */
 285       assert (sizeof (size_t) <= sizeof (unsigned long long int));
 286 #if LONG_MAX != LONG_LONG_MAX
 287       spec->info.is_long_double = sizeof (size_t) > sizeof (unsigned long int);
 288 #endif
 289       spec->info.is_long = sizeof (size_t) > sizeof (unsigned int);
 290       break;
 291     case L_('t'):
 292       assert (sizeof (ptrdiff_t) <= sizeof (long long int));
 293 #if LONG_MAX != LONG_LONG_MAX
 294       spec->info.is_long_double = (sizeof (ptrdiff_t) > sizeof (long int));
 295 #endif
 296       spec->info.is_long = sizeof (ptrdiff_t) > sizeof (int);
 297       break;
 298     case L_('j'):
 299       assert (sizeof (uintmax_t) <= sizeof (unsigned long long int));
 300 #if LONG_MAX != LONG_LONG_MAX
 301       spec->info.is_long_double = (sizeof (uintmax_t)
 302                                    > sizeof (unsigned long int));
 303 #endif
 304       spec->info.is_long = sizeof (uintmax_t) > sizeof (unsigned int);
 305       break;
 306     default:
 307       /* Not a recognized modifier.  Backup.  */
 308       --format;
 309       break;
 310     }
 311
 312   /* Get the format specification.  */
 313   spec->info.spec = (wchar_t) *format++;
 314   if (__builtin_expect (__printf_function_table != NULL, 0)
 315       && spec->info.spec <= UCHAR_MAX
 316       && __printf_arginfo_table[spec->info.spec] != NULL)
 317     /* We don't try to get the types for all arguments if the format
 318        uses more than one.  The normal case is covered though.  */
 319     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 320       (&spec->info, 1, &spec->data_arg_type);
 321   else
 322     {
 323       /* Find the data argument types of a built-in spec.  */
 324       spec->ndata_args = 1;
 325
 326       switch (spec->info.spec)
 327         {
 328         case L'i':
 329         case L'd':
 330         case L'u':
 331         case L'o':
 332         case L'X':
 333         case L'x':
 334 #if LONG_MAX != LONG_LONG_MAX
 335           if (spec->info.is_long_double)
 336             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 337           else
 338 #endif
 339             if (spec->info.is_long)
 340               spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 341             else if (spec->info.is_short)
 342               spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 343             else if (spec->info.is_char)
 344               spec->data_arg_type = PA_CHAR;
 345             else
 346               spec->data_arg_type = PA_INT;
 347           break;
 348         case L'e':
 349         case L'E':
 350         case L'f':
 351         case L'F':
 352         case L'g':
 353         case L'G':
 354         case L'a':
 355         case L'A':
 356           if (spec->info.is_long_double)
 357             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 358           else
 359             spec->data_arg_type = PA_DOUBLE;
 360           break;
 361         case L'c':
 362           spec->data_arg_type = PA_CHAR;
 363           break;
 364         case L'C':
 365           spec->data_arg_type = PA_WCHAR;
 366           break;
 367         case L's':
 368           spec->data_arg_type = PA_STRING;
 369           break;
 370         case L'S':
 371           spec->data_arg_type = PA_WSTRING;
 372           break;
 373         case L'p':
 374           spec->data_arg_type = PA_POINTER;
 375           break;
 376         case L'n':
 377           spec->data_arg_type = PA_INT|PA_FLAG_PTR;
 378           break;
 379
 380         case L'm':
 381         default:
 382           /* An unknown spec will consume no args.  */
 383           spec->ndata_args = 0;
 384           break;
 385         }
 386     }
 387
 388   if (spec->data_arg == -1 && spec->ndata_args > 0)
 389     {
 390       /* There are args consumed, but no positional spec.  Use the
 391          next sequential arg position.  */
 392       spec->data_arg = posn;
 393       nargs += spec->ndata_args;
 394     }
 395
 396   if (spec->info.spec == L'\0')
 397     /* Format ended before this spec was complete.  */
 398     spec->end_of_fmt = spec->next_fmt = format - 1;
 399   else
 400     {
 401       /* Find the next format spec.  */
 402       spec->end_of_fmt = format;
 403 #ifdef COMPILE_WPRINTF
 404       spec->next_fmt = __find_specwc (format);
 405 #else
 406       spec->next_fmt = __find_specmb (format, ps);
 407 #endif
 408     }
 409
 410   return nargs;
 411 }