lib/printf-parse.c

   1 /* Formatted output to strings.
   2    Copyright (C) 1999-2000, 2002-2003, 2006-2007 Free Software Foundation, Inc.
   3
   4    This program is free software; you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 2, or (at your option)
   7    any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License along
  15    with this program; if not, write to the Free Software Foundation,
  16    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  17
  18 /* This file can be parametrized with the following macros:
  19      CHAR_T             The element type of the format string.
  20      CHAR_T_ONLY_ASCII  Set to 1 to enable verification that all characters
  21                         in the format string are ASCII.
  22      DIRECTIVE          Structure denoting a format directive.
  23                         Depends on CHAR_T.
  24      DIRECTIVES         Structure denoting the set of format directives of a
  25                         format string.  Depends on CHAR_T.
  26      PRINTF_PARSE       Function that parses a format string.
  27                         Depends on CHAR_T.
  28      STATIC             Set to 'static' to declare the function static.
  29      ENABLE_UNISTDIO    Set to 1 to enable the unistdio extensions.  */
  30
  31 #ifndef PRINTF_PARSE
  32 # include <config.h>
  33 #endif
  34
  35 /* Specification.  */
  36 #ifndef PRINTF_PARSE
  37 # include "printf-parse.h"
  38 #endif
  39
  40 /* Default parameters.  */
  41 #ifndef PRINTF_PARSE
  42 # define PRINTF_PARSE printf_parse
  43 # define CHAR_T char
  44 # define DIRECTIVE char_directive
  45 # define DIRECTIVES char_directives
  46 #endif
  47
  48 /* Get size_t, NULL.  */
  49 #include <stddef.h>
  50
  51 /* Get intmax_t.  */
  52 #ifdef IN_LIBINTL
  53 # if HAVE_STDINT_H_WITH_UINTMAX
  54 #  include <stdint.h>
  55 # endif
  56 # if HAVE_INTTYPES_H_WITH_UINTMAX
  57 #  include <inttypes.h>
  58 # endif
  59 #else
  60 # include <stdint.h>
  61 #endif
  62
  63 /* malloc(), realloc(), free().  */
  64 #include <stdlib.h>
  65
  66 /* Checked size_t computations.  */
  67 #include "xsize.h"
  68
  69 #if CHAR_T_ONLY_ASCII
  70 /* c_isascii().  */
  71 # include "c-ctype.h"
  72 #endif
  73
  74 #ifdef STATIC
  75 STATIC
  76 #endif
  77 int
  78 PRINTF_PARSE (const CHAR_T *format, DIRECTIVES *d, arguments *a)
  79 {
  80   const CHAR_T *cp = format;            /* pointer into format */
  81   size_t arg_posn = 0;          /* number of regular arguments consumed */
  82   size_t d_allocated;                   /* allocated elements of d->dir */
  83   size_t a_allocated;                   /* allocated elements of a->arg */
  84   size_t max_width_length = 0;
  85   size_t max_precision_length = 0;
  86
  87   d->count = 0;
  88   d_allocated = 1;
  89   d->dir = (DIRECTIVE *) malloc (d_allocated * sizeof (DIRECTIVE));
  90   if (d->dir == NULL)
  91     /* Out of memory.  */
  92     return -1;
  93
  94   a->count = 0;
  95   a_allocated = 0;
  96   a->arg = NULL;
  97
  98 #define REGISTER_ARG(_index_,_type_) \
  99   {                                                                     \
 100     size_t n = (_index_);                                               \
 101     if (n >= a_allocated)                                               \
 102       {                                                                 \
 103         size_t memory_size;                                             \
 104         argument *memory;                                               \
 105                                                                         \
 106         a_allocated = xtimes (a_allocated, 2);                          \
 107         if (a_allocated <= n)                                           \
 108           a_allocated = xsum (n, 1);                                    \
 109         memory_size = xtimes (a_allocated, sizeof (argument));          \
 110         if (size_overflow_p (memory_size))                              \
 111           /* Overflow, would lead to out of memory.  */                 \
 112           goto error;                                                   \
 113         memory = (argument *) (a->arg                                   \
 114                                ? realloc (a->arg, memory_size)          \
 115                                : malloc (memory_size));                 \
 116         if (memory == NULL)                                             \
 117           /* Out of memory.  */                                         \
 118           goto error;                                                   \
 119         a->arg = memory;                                                \
 120       }                                                                 \
 121     while (a->count <= n)                                               \
 122       a->arg[a->count++].type = TYPE_NONE;                              \
 123     if (a->arg[n].type == TYPE_NONE)                                    \
 124       a->arg[n].type = (_type_);                                        \
 125     else if (a->arg[n].type != (_type_))                                \
 126       /* Ambiguous type for positional argument.  */                    \
 127       goto error;                                                       \
 128   }
 129
 130   while (*cp != '\0')
 131     {
 132       CHAR_T c = *cp++;
 133       if (c == '%')
 134         {
 135           size_t arg_index = ARG_NONE;
 136           DIRECTIVE *dp = &d->dir[d->count]; /* pointer to next directive */
 137
 138           /* Initialize the next directive.  */
 139           dp->dir_start = cp - 1;
 140           dp->flags = 0;
 141           dp->width_start = NULL;
 142           dp->width_end = NULL;
 143           dp->width_arg_index = ARG_NONE;
 144           dp->precision_start = NULL;
 145           dp->precision_end = NULL;
 146           dp->precision_arg_index = ARG_NONE;
 147           dp->arg_index = ARG_NONE;
 148
 149           /* Test for positional argument.  */
 150           if (*cp >= '0' && *cp <= '9')
 151             {
 152               const CHAR_T *np;
 153
 154               for (np = cp; *np >= '0' && *np <= '9'; np++)
 155                 ;
 156               if (*np == '$')
 157                 {
 158                   size_t n = 0;
 159
 160                   for (np = cp; *np >= '0' && *np <= '9'; np++)
 161                     n = xsum (xtimes (n, 10), *np - '0');
 162                   if (n == 0)
 163                     /* Positional argument 0.  */
 164                     goto error;
 165                   if (size_overflow_p (n))
 166                     /* n too large, would lead to out of memory later.  */
 167                     goto error;
 168                   arg_index = n - 1;
 169                   cp = np + 1;
 170                 }
 171             }
 172
 173           /* Read the flags.  */
 174           for (;;)
 175             {
 176               if (*cp == '\'')
 177                 {
 178                   dp->flags |= FLAG_GROUP;
 179                   cp++;
 180                 }
 181               else if (*cp == '-')
 182                 {
 183                   dp->flags |= FLAG_LEFT;
 184                   cp++;
 185                 }
 186               else if (*cp == '+')
 187                 {
 188                   dp->flags |= FLAG_SHOWSIGN;
 189                   cp++;
 190                 }
 191               else if (*cp == ' ')
 192                 {
 193                   dp->flags |= FLAG_SPACE;
 194                   cp++;
 195                 }
 196               else if (*cp == '#')
 197                 {
 198                   dp->flags |= FLAG_ALT;
 199                   cp++;
 200                 }
 201               else if (*cp == '0')
 202                 {
 203                   dp->flags |= FLAG_ZERO;
 204                   cp++;
 205                 }
 206               else
 207                 break;
 208             }
 209
 210           /* Parse the field width.  */
 211           if (*cp == '*')
 212             {
 213               dp->width_start = cp;
 214               cp++;
 215               dp->width_end = cp;
 216               if (max_width_length < 1)
 217                 max_width_length = 1;
 218
 219               /* Test for positional argument.  */
 220               if (*cp >= '0' && *cp <= '9')
 221                 {
 222                   const CHAR_T *np;
 223
 224                   for (np = cp; *np >= '0' && *np <= '9'; np++)
 225                     ;
 226                   if (*np == '$')
 227                     {
 228                       size_t n = 0;
 229
 230                       for (np = cp; *np >= '0' && *np <= '9'; np++)
 231                         n = xsum (xtimes (n, 10), *np - '0');
 232                       if (n == 0)
 233                         /* Positional argument 0.  */
 234                         goto error;
 235                       if (size_overflow_p (n))
 236                         /* n too large, would lead to out of memory later.  */
 237                         goto error;
 238                       dp->width_arg_index = n - 1;
 239                       cp = np + 1;
 240                     }
 241                 }
 242               if (dp->width_arg_index == ARG_NONE)
 243                 {
 244                   dp->width_arg_index = arg_posn++;
 245                   if (dp->width_arg_index == ARG_NONE)
 246                     /* arg_posn wrapped around.  */
 247                     goto error;
 248                 }
 249               REGISTER_ARG (dp->width_arg_index, TYPE_INT);
 250             }
 251           else if (*cp >= '0' && *cp <= '9')
 252             {
 253               size_t width_length;
 254
 255               dp->width_start = cp;
 256               for (; *cp >= '0' && *cp <= '9'; cp++)
 257                 ;
 258               dp->width_end = cp;
 259               width_length = dp->width_end - dp->width_start;
 260               if (max_width_length < width_length)
 261                 max_width_length = width_length;
 262             }
 263
 264           /* Parse the precision.  */
 265           if (*cp == '.')
 266             {
 267               cp++;
 268               if (*cp == '*')
 269                 {
 270                   dp->precision_start = cp - 1;
 271                   cp++;
 272                   dp->precision_end = cp;
 273                   if (max_precision_length < 2)
 274                     max_precision_length = 2;
 275
 276                   /* Test for positional argument.  */
 277                   if (*cp >= '0' && *cp <= '9')
 278                     {
 279                       const CHAR_T *np;
 280
 281                       for (np = cp; *np >= '0' && *np <= '9'; np++)
 282                         ;
 283                       if (*np == '$')
 284                         {
 285                           size_t n = 0;
 286
 287                           for (np = cp; *np >= '0' && *np <= '9'; np++)
 288                             n = xsum (xtimes (n, 10), *np - '0');
 289                           if (n == 0)
 290                             /* Positional argument 0.  */
 291                             goto error;
 292                           if (size_overflow_p (n))
 293                             /* n too large, would lead to out of memory
 294                                later.  */
 295                             goto error;
 296                           dp->precision_arg_index = n - 1;
 297                           cp = np + 1;
 298                         }
 299                     }
 300                   if (dp->precision_arg_index == ARG_NONE)
 301                     {
 302                       dp->precision_arg_index = arg_posn++;
 303                       if (dp->precision_arg_index == ARG_NONE)
 304                         /* arg_posn wrapped around.  */
 305                         goto error;
 306                     }
 307                   REGISTER_ARG (dp->precision_arg_index, TYPE_INT);
 308                 }
 309               else
 310                 {
 311                   size_t precision_length;
 312
 313                   dp->precision_start = cp - 1;
 314                   for (; *cp >= '0' && *cp <= '9'; cp++)
 315                     ;
 316                   dp->precision_end = cp;
 317                   precision_length = dp->precision_end - dp->precision_start;
 318                   if (max_precision_length < precision_length)
 319                     max_precision_length = precision_length;
 320                 }
 321             }
 322
 323           {
 324             arg_type type;
 325
 326             /* Parse argument type/size specifiers.  */
 327             {
 328               int flags = 0;
 329
 330               for (;;)
 331                 {
 332                   if (*cp == 'h')
 333                     {
 334                       flags |= (1 << (flags & 1));
 335                       cp++;
 336                     }
 337                   else if (*cp == 'L')
 338                     {
 339                       flags |= 4;
 340                       cp++;
 341                     }
 342                   else if (*cp == 'l')
 343                     {
 344                       flags += 8;
 345                       cp++;
 346                     }
 347                   else if (*cp == 'j')
 348                     {
 349                       if (sizeof (intmax_t) > sizeof (long))
 350                         {
 351                           /* intmax_t = long long */
 352                           flags += 16;
 353                         }
 354                       else if (sizeof (intmax_t) > sizeof (int))
 355                         {
 356                           /* intmax_t = long */
 357                           flags += 8;
 358                         }
 359                       cp++;
 360                     }
 361                   else if (*cp == 'z' || *cp == 'Z')
 362                     {
 363                       /* 'z' is standardized in ISO C 99, but glibc uses 'Z'
 364                          because the warning facility in gcc-2.95.2 understands
 365                          only 'Z' (see gcc-2.95.2/gcc/c-common.c:1784).  */
 366                       if (sizeof (size_t) > sizeof (long))
 367                         {
 368                           /* size_t = long long */
 369                           flags += 16;
 370                         }
 371                       else if (sizeof (size_t) > sizeof (int))
 372                         {
 373                           /* size_t = long */
 374                           flags += 8;
 375                         }
 376                       cp++;
 377                     }
 378                   else if (*cp == 't')
 379                     {
 380                       if (sizeof (ptrdiff_t) > sizeof (long))
 381                         {
 382                           /* ptrdiff_t = long long */
 383                           flags += 16;
 384                         }
 385                       else if (sizeof (ptrdiff_t) > sizeof (int))
 386                         {
 387                           /* ptrdiff_t = long */
 388                           flags += 8;
 389                         }
 390                       cp++;
 391                     }
 392                   else
 393                     break;
 394                 }
 395
 396               /* Read the conversion character.  */
 397               c = *cp++;
 398               switch (c)
 399                 {
 400                 case 'd': case 'i':
 401 #if HAVE_LONG_LONG_INT
 402                   /* If 'long long' exists and is larger than 'long':  */
 403                   if (flags >= 16 || (flags & 4))
 404                     type = TYPE_LONGLONGINT;
 405                   else
 406 #endif
 407                   /* If 'long long' exists and is the same as 'long', we parse
 408                      "lld" into TYPE_LONGINT.  */
 409                   if (flags >= 8)
 410                     type = TYPE_LONGINT;
 411                   else if (flags & 2)
 412                     type = TYPE_SCHAR;
 413                   else if (flags & 1)
 414                     type = TYPE_SHORT;
 415                   else
 416                     type = TYPE_INT;
 417                   break;
 418                 case 'o': case 'u': case 'x': case 'X':
 419 #if HAVE_LONG_LONG_INT
 420                   /* If 'long long' exists and is larger than 'long':  */
 421                   if (flags >= 16 || (flags & 4))
 422                     type = TYPE_ULONGLONGINT;
 423                   else
 424 #endif
 425                   /* If 'unsigned long long' exists and is the same as
 426                      'unsigned long', we parse "llu" into TYPE_ULONGINT.  */
 427                   if (flags >= 8)
 428                     type = TYPE_ULONGINT;
 429                   else if (flags & 2)
 430                     type = TYPE_UCHAR;
 431                   else if (flags & 1)
 432                     type = TYPE_USHORT;
 433                   else
 434                     type = TYPE_UINT;
 435                   break;
 436                 case 'f': case 'F': case 'e': case 'E': case 'g': case 'G':
 437                 case 'a': case 'A':
 438                   if (flags >= 16 || (flags & 4))
 439                     type = TYPE_LONGDOUBLE;
 440                   else
 441                     type = TYPE_DOUBLE;
 442                   break;
 443                 case 'c':
 444                   if (flags >= 8)
 445 #if HAVE_WINT_T
 446                     type = TYPE_WIDE_CHAR;
 447 #else
 448                     goto error;
 449 #endif
 450                   else
 451                     type = TYPE_CHAR;
 452                   break;
 453 #if HAVE_WINT_T
 454                 case 'C':
 455                   type = TYPE_WIDE_CHAR;
 456                   c = 'c';
 457                   break;
 458 #endif
 459                 case 's':
 460                   if (flags >= 8)
 461 #if HAVE_WCHAR_T
 462                     type = TYPE_WIDE_STRING;
 463 #else
 464                     goto error;
 465 #endif
 466                   else
 467                     type = TYPE_STRING;
 468                   break;
 469 #if HAVE_WCHAR_T
 470                 case 'S':
 471                   type = TYPE_WIDE_STRING;
 472                   c = 's';
 473                   break;
 474 #endif
 475                 case 'p':
 476                   type = TYPE_POINTER;
 477                   break;
 478                 case 'n':
 479 #if HAVE_LONG_LONG_INT
 480                   /* If 'long long' exists and is larger than 'long':  */
 481                   if (flags >= 16 || (flags & 4))
 482                     type = TYPE_COUNT_LONGLONGINT_POINTER;
 483                   else
 484 #endif
 485                   /* If 'long long' exists and is the same as 'long', we parse
 486                      "lln" into TYPE_COUNT_LONGINT_POINTER.  */
 487                   if (flags >= 8)
 488                     type = TYPE_COUNT_LONGINT_POINTER;
 489                   else if (flags & 2)
 490                     type = TYPE_COUNT_SCHAR_POINTER;
 491                   else if (flags & 1)
 492                     type = TYPE_COUNT_SHORT_POINTER;
 493                   else
 494                     type = TYPE_COUNT_INT_POINTER;
 495                   break;
 496 #if ENABLE_UNISTDIO
 497                 /* The unistdio extensions.  */
 498                 case 'U':
 499                   if (flags >= 16)
 500                     type = TYPE_U32_STRING;
 501                   else if (flags >= 8)
 502                     type = TYPE_U16_STRING;
 503                   else
 504                     type = TYPE_U8_STRING;
 505                   break;
 506 #endif
 507                 case '%':
 508                   type = TYPE_NONE;
 509                   break;
 510                 default:
 511                   /* Unknown conversion character.  */
 512                   goto error;
 513                 }
 514             }
 515
 516             if (type != TYPE_NONE)
 517               {
 518                 dp->arg_index = arg_index;
 519                 if (dp->arg_index == ARG_NONE)
 520                   {
 521                     dp->arg_index = arg_posn++;
 522                     if (dp->arg_index == ARG_NONE)
 523                       /* arg_posn wrapped around.  */
 524                       goto error;
 525                   }
 526                 REGISTER_ARG (dp->arg_index, type);
 527               }
 528             dp->conversion = c;
 529             dp->dir_end = cp;
 530           }
 531
 532           d->count++;
 533           if (d->count >= d_allocated)
 534             {
 535               size_t memory_size;
 536               DIRECTIVE *memory;
 537
 538               d_allocated = xtimes (d_allocated, 2);
 539               memory_size = xtimes (d_allocated, sizeof (DIRECTIVE));
 540               if (size_overflow_p (memory_size))
 541                 /* Overflow, would lead to out of memory.  */
 542                 goto error;
 543               memory = (DIRECTIVE *) realloc (d->dir, memory_size);
 544               if (memory == NULL)
 545                 /* Out of memory.  */
 546                 goto error;
 547               d->dir = memory;
 548             }
 549         }
 550 #if CHAR_T_ONLY_ASCII
 551       else if (!c_isascii (c))
 552         {
 553           /* Non-ASCII character.  Not supported.  */
 554           goto error;
 555         }
 556 #endif
 557     }
 558   d->dir[d->count].dir_start = cp;
 559
 560   d->max_width_length = max_width_length;
 561   d->max_precision_length = max_precision_length;
 562   return 0;
 563
 564 error:
 565   if (a->arg)
 566     free (a->arg);
 567   if (d->dir)
 568     free (d->dir);
 569   return -1;
 570 }
 571
 572 #undef PRINTF_PARSE
 573 #undef DIRECTIVES
 574 #undef DIRECTIVE
 575 #undef CHAR_T_ONLY_ASCII
 576 #undef CHAR_T