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 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 /* This is defined in reg-printf.c.  */
 110 extern printf_arginfo_function **__printf_arginfo_table;
 111
 112
 113 /* FORMAT must point to a '%' at the beginning of a spec.  Fills in *SPEC
 114    with the parsed details.  POSN is the number of arguments already
 115    consumed.  At most MAXTYPES - POSN types are filled in TYPES.  Return
 116    the number of args consumed by this spec; *MAX_REF_ARG is updated so it
 117    remains the highest argument index used.  */
 118 static inline size_t
 119 parse_one_spec (const UCHAR_T *format, size_t posn, struct printf_spec *spec,
 120                 size_t *max_ref_arg, mbstate_t *ps)
 121 {
 122   unsigned int n;
 123   size_t nargs = 0;
 124
 125   /* Skip the '%'.  */
 126   ++format;
 127
 128   /* Clear information structure.  */
 129   spec->data_arg = -1;
 130   spec->info.alt = 0;
 131   spec->info.space = 0;
 132   spec->info.left = 0;
 133   spec->info.showsign = 0;
 134   spec->info.group = 0;
 135   spec->info.pad = ' ';
 136
 137   /* Test for positional argument.  */
 138   if (ISDIGIT (*format))
 139     {
 140       const UCHAR_T *begin = format;
 141
 142       n = read_int (&format);
 143
 144       if (n > 0 && *format == L_('$'))
 145         /* Is positional parameter.  */
 146         {
 147           ++format;             /* Skip the '$'.  */
 148           spec->data_arg = n - 1;
 149           *max_ref_arg = MAX (*max_ref_arg, n);
 150         }
 151       else
 152         /* Oops; that was actually the width and/or 0 padding flag.
 153            Step back and read it again.  */
 154         format = begin;
 155     }
 156
 157   /* Check for spec modifiers.  */
 158   while (*format == L_(' ') || *format == L_('+') || *format == L_('-') ||
 159          *format == L_('#') || *format == L_('0') || *format == L_('\''))
 160     switch (*format++)
 161       {
 162       case L_(' '):
 163         /* Output a space in place of a sign, when there is no sign.  */
 164         spec->info.space = 1;
 165         break;
 166       case L_('+'):
 167         /* Always output + or - for numbers.  */
 168         spec->info.showsign = 1;
 169         break;
 170       case L_('-'):
 171         /* Left-justify things.  */
 172         spec->info.left = 1;
 173         break;
 174       case L_('#'):
 175         /* Use the "alternate form":
 176            Hex has 0x or 0X, FP always has a decimal point.  */
 177         spec->info.alt = 1;
 178         break;
 179       case L_('0'):
 180         /* Pad with 0s.  */
 181         spec->info.pad = '0';
 182         break;
 183       case L_('\''):
 184         /* Show grouping in numbers if the locale information
 185            indicates any.  */
 186         spec->info.group = 1;
 187         break;
 188       }
 189   if (spec->info.left)
 190     spec->info.pad = ' ';
 191
 192   /* Get the field width.  */
 193   spec->width_arg = -1;
 194   spec->info.width = 0;
 195   if (*format == L_('*'))
 196     {
 197       /* The field width is given in an argument.
 198          A negative field width indicates left justification.  */
 199       const UCHAR_T *begin = ++format;
 200
 201       if (ISDIGIT (*format))
 202         {
 203           /* The width argument might be found in a positional parameter.  */
 204           n = read_int (&format);
 205
 206           if (n > 0 && *format == L_('$'))
 207             {
 208               spec->width_arg = n - 1;
 209               *max_ref_arg = MAX (*max_ref_arg, n);
 210               ++format;         /* Skip '$'.  */
 211             }
 212         }
 213
 214       if (spec->width_arg < 0)
 215         {
 216           /* Not in a positional parameter.  Consume one argument.  */
 217           spec->width_arg = posn++;
 218           ++nargs;
 219           format = begin;       /* Step back and reread.  */
 220         }
 221     }
 222   else if (ISDIGIT (*format))
 223     /* Constant width specification.  */
 224     spec->info.width = read_int (&format);
 225
 226   /* Get the precision.  */
 227   spec->prec_arg = -1;
 228   /* -1 means none given; 0 means explicit 0.  */
 229   spec->info.prec = -1;
 230   if (*format == L_('.'))
 231     {
 232       ++format;
 233       if (*format == L_('*'))
 234         {
 235           /* The precision is given in an argument.  */
 236           const UCHAR_T *begin = ++format;
 237
 238           if (ISDIGIT (*format))
 239             {
 240               n = read_int (&format);
 241
 242               if (n > 0 && *format == L_('$'))
 243                 {
 244                   spec->prec_arg = n - 1;
 245                   *max_ref_arg = MAX (*max_ref_arg, n);
 246                   ++format;
 247                 }
 248             }
 249
 250           if (spec->prec_arg < 0)
 251             {
 252               /* Not in a positional parameter.  */
 253               spec->prec_arg = posn++;
 254               ++nargs;
 255               format = begin;
 256             }
 257         }
 258       else if (ISDIGIT (*format))
 259         spec->info.prec = read_int (&format);
 260       else
 261         /* "%.?" is treated like "%.0?".  */
 262         spec->info.prec = 0;
 263     }
 264
 265   /* Check for type modifiers.  */
 266 #define is_longlong is_long_double
 267   spec->info.is_long_double = 0;
 268   spec->info.is_short = 0;
 269   spec->info.is_long = 0;
 270
 271   if (*format == L_('h') || *format == L_('l') || *format == L_('L') ||
 272       *format == L_('Z') || *format == L_('q'))
 273     switch (*format++)
 274       {
 275       case L_('h'):
 276         /* int's are short int's.  */
 277         spec->info.is_short = 1;
 278         break;
 279       case L_('l'):
 280         /* int's are long int's.  */
 281         spec->info.is_long = 1;
 282         if (*format != L_('l'))
 283           break;
 284         ++format;
 285         /* FALLTHROUGH */
 286       case L_('L'):
 287         /* double's are long double's, and int's are long long int's.  */
 288       case L_('q'):
 289         /* 4.4 uses this for long long.  */
 290         spec->info.is_long_double = 1;
 291         break;
 292       case L_('Z'):
 293         /* int's are size_t's.  */
 294         assert (sizeof(size_t) <= sizeof(unsigned long long int));
 295         spec->info.is_longlong = sizeof(size_t) > sizeof(unsigned long int);
 296         spec->info.is_long = sizeof(size_t) > sizeof(unsigned int);
 297         break;
 298       }
 299
 300   /* Get the format specification.  */
 301   spec->info.spec = (wchar_t) *format++;
 302   if (__printf_arginfo_table != NULL &&
 303       __printf_arginfo_table[spec->info.spec] != NULL)
 304     /* We don't try to get the types for all arguments if the format
 305        uses more than one.  The normal case is covered though.  */
 306     spec->ndata_args = (*__printf_arginfo_table[spec->info.spec])
 307       (&spec->info, 1, &spec->data_arg_type);
 308   else
 309     {
 310       /* Find the data argument types of a built-in spec.  */
 311       spec->ndata_args = 1;
 312
 313       switch (spec->info.spec)
 314         {
 315         case L'i':
 316         case L'd':
 317         case L'u':
 318         case L'o':
 319         case L'X':
 320         case L'x':
 321           if (spec->info.is_longlong)
 322             spec->data_arg_type = PA_INT|PA_FLAG_LONG_LONG;
 323           else if (spec->info.is_long)
 324             spec->data_arg_type = PA_INT|PA_FLAG_LONG;
 325           else if (spec->info.is_short)
 326             spec->data_arg_type = PA_INT|PA_FLAG_SHORT;
 327           else
 328             spec->data_arg_type = PA_INT;
 329           break;
 330         case L'e':
 331         case L'E':
 332         case L'f':
 333         case L'g':
 334         case L'G':
 335           if (spec->info.is_long_double)
 336             spec->data_arg_type = PA_DOUBLE|PA_FLAG_LONG_DOUBLE;
 337           else
 338             spec->data_arg_type = PA_DOUBLE;
 339           break;
 340         case L'c':
 341           spec->data_arg_type = PA_CHAR;
 342           break;
 343         case L'C':
 344           spec->data_arg_type = PA_WCHAR;
 345           break;
 346         case L's':
 347           spec->data_arg_type = PA_STRING;
 348           break;
 349         case L'S':
 350           spec->data_arg_type = PA_WSTRING;
 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 }