lib/printf-parse.c

   1 /* Formatted output to strings.
   2    Copyright (C) 1999-2000, 2002-2003, 2006-2011 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 Lesser 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 Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser 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 #if defined IN_LIBINTL || defined IN_LIBASPRINTF
  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 /* memcpy().  */
  67 #include <string.h>
  68
  69 /* errno.  */
  70 #include <errno.h>
  71
  72 /* Checked size_t computations.  */
  73 #include "xsize.h"
  74
  75 #if CHAR_T_ONLY_ASCII
  76 /* c_isascii().  */
  77 # include "c-ctype.h"
  78 #endif
  79
  80 #ifdef STATIC
  81 STATIC
  82 #endif
  83 int
  84 PRINTF_PARSE (const CHAR_T *format, DIRECTIVES *d, arguments *a)
  85 {
  86   const CHAR_T *cp = format;    /* pointer into format */
  87   size_t arg_posn = 0;          /* number of regular arguments consumed */
  88   size_t d_allocated;           /* allocated elements of d->dir */
  89   size_t a_allocated;           /* allocated elements of a->arg */
  90   size_t max_width_length = 0;
  91   size_t max_precision_length = 0;
  92
  93   d->count = 0;
  94   d_allocated = N_DIRECT_ALLOC_DIRECTIVES;
  95   d->dir = d->direct_alloc_dir;
  96
  97   a->count = 0;
  98   a_allocated = N_DIRECT_ALLOC_ARGUMENTS;
  99   a->arg = a->direct_alloc_arg;
 100
 101 #define REGISTER_ARG(_index_,_type_) \
 102   {                                                                     \
 103     size_t n = (_index_);                                               \
 104     if (n >= a_allocated)                                               \
 105       {                                                                 \
 106         size_t memory_size;                                             \
 107         argument *memory;                                               \
 108                                                                         \
 109         a_allocated = xtimes (a_allocated, 2);                          \
 110         if (a_allocated <= n)                                           \
 111           a_allocated = xsum (n, 1);                                    \
 112         memory_size = xtimes (a_allocated, sizeof (argument));          \
 113         if (size_overflow_p (memory_size))                              \
 114           /* Overflow, would lead to out of memory.  */                 \
 115           goto out_of_memory;                                           \
 116         memory = (argument *) (a->arg != a->direct_alloc_arg            \
 117                                ? realloc (a->arg, memory_size)          \
 118                                : malloc (memory_size));                 \
 119         if (memory == NULL)                                             \
 120           /* Out of memory.  */                                         \
 121           goto out_of_memory;                                           \
 122         if (a->arg == a->direct_alloc_arg)                              \
 123           memcpy (memory, a->arg, a->count * sizeof (argument));        \
 124         a->arg = memory;                                                \
 125       }                                                                 \
 126     while (a->count <= n)                                               \
 127       a->arg[a->count++].type = TYPE_NONE;                              \
 128     if (a->arg[n].type == TYPE_NONE)                                    \
 129       a->arg[n].type = (_type_);                                        \
 130     else if (a->arg[n].type != (_type_))                                \
 131       /* Ambiguous type for positional argument.  */                    \
 132       goto error;                                                       \
 133   }
 134
 135   while (*cp != '\0')
 136     {
 137       CHAR_T c = *cp++;
 138       if (c == '%')
 139         {
 140           size_t arg_index = ARG_NONE;
 141           DIRECTIVE *dp = &d->dir[d->count]; /* pointer to next directive */
 142
 143           /* Initialize the next directive.  */
 144           dp->dir_start = cp - 1;
 145           dp->flags = 0;
 146           dp->width_start = NULL;
 147           dp->width_end = NULL;
 148           dp->width_arg_index = ARG_NONE;
 149           dp->precision_start = NULL;
 150           dp->precision_end = NULL;
 151           dp->precision_arg_index = ARG_NONE;
 152           dp->arg_index = ARG_NONE;
 153
 154           /* Test for positional argument.  */
 155           if (*cp >= '0' && *cp <= '9')
 156             {
 157               const CHAR_T *np;
 158
 159               for (np = cp; *np >= '0' && *np <= '9'; np++)
 160                 ;
 161               if (*np == '$')
 162                 {
 163                   size_t n = 0;
 164
 165                   for (np = cp; *np >= '0' && *np <= '9'; np++)
 166                     n = xsum (xtimes (n, 10), *np - '0');
 167                   if (n == 0)
 168                     /* Positional argument 0.  */
 169                     goto error;
 170                   if (size_overflow_p (n))
 171                     /* n too large, would lead to out of memory later.  */
 172                     goto error;
 173                   arg_index = n - 1;
 174                   cp = np + 1;
 175                 }
 176             }
 177
 178           /* Read the flags.  */
 179           for (;;)
 180             {
 181               if (*cp == '\'')
 182                 {
 183                   dp->flags |= FLAG_GROUP;
 184                   cp++;
 185                 }
 186               else if (*cp == '-')
 187                 {
 188                   dp->flags |= FLAG_LEFT;
 189                   cp++;
 190                 }
 191               else if (*cp == '+')
 192                 {
 193                   dp->flags |= FLAG_SHOWSIGN;
 194                   cp++;
 195                 }
 196               else if (*cp == ' ')
 197                 {
 198                   dp->flags |= FLAG_SPACE;
 199                   cp++;
 200                 }
 201               else if (*cp == '#')
 202                 {
 203                   dp->flags |= FLAG_ALT;
 204                   cp++;
 205                 }
 206               else if (*cp == '0')
 207                 {
 208                   dp->flags |= FLAG_ZERO;
 209                   cp++;
 210                 }
 211 #if __GLIBC__ >= 2 && !defined __UCLIBC__
 212               else if (*cp == 'I')
 213                 {
 214                   dp->flags |= FLAG_LOCALIZED;
 215                   cp++;
 216                 }
 217 #endif
 218               else
 219                 break;
 220             }
 221
 222           /* Parse the field width.  */
 223           if (*cp == '*')
 224             {
 225               dp->width_start = cp;
 226               cp++;
 227               dp->width_end = cp;
 228               if (max_width_length < 1)
 229                 max_width_length = 1;
 230
 231               /* Test for positional argument.  */
 232               if (*cp >= '0' && *cp <= '9')
 233                 {
 234                   const CHAR_T *np;
 235
 236                   for (np = cp; *np >= '0' && *np <= '9'; np++)
 237                     ;
 238                   if (*np == '$')
 239                     {
 240                       size_t n = 0;
 241
 242                       for (np = cp; *np >= '0' && *np <= '9'; np++)
 243                         n = xsum (xtimes (n, 10), *np - '0');
 244                       if (n == 0)
 245                         /* Positional argument 0.  */
 246                         goto error;
 247                       if (size_overflow_p (n))
 248                         /* n too large, would lead to out of memory later.  */
 249                         goto error;
 250                       dp->width_arg_index = n - 1;
 251                       cp = np + 1;
 252                     }
 253                 }
 254               if (dp->width_arg_index == ARG_NONE)
 255                 {
 256                   dp->width_arg_index = arg_posn++;
 257                   if (dp->width_arg_index == ARG_NONE)
 258                     /* arg_posn wrapped around.  */
 259                     goto error;
 260                 }
 261               REGISTER_ARG (dp->width_arg_index, TYPE_INT);
 262             }
 263           else if (*cp >= '0' && *cp <= '9')
 264             {
 265               size_t width_length;
 266
 267               dp->width_start = cp;
 268               for (; *cp >= '0' && *cp <= '9'; cp++)
 269                 ;
 270               dp->width_end = cp;
 271               width_length = dp->width_end - dp->width_start;
 272               if (max_width_length < width_length)
 273                 max_width_length = width_length;
 274             }
 275
 276           /* Parse the precision.  */
 277           if (*cp == '.')
 278             {
 279               cp++;
 280               if (*cp == '*')
 281                 {
 282                   dp->precision_start = cp - 1;
 283                   cp++;
 284                   dp->precision_end = cp;
 285                   if (max_precision_length < 2)
 286                     max_precision_length = 2;
 287
 288                   /* Test for positional argument.  */
 289                   if (*cp >= '0' && *cp <= '9')
 290                     {
 291                       const CHAR_T *np;
 292
 293                       for (np = cp; *np >= '0' && *np <= '9'; np++)
 294                         ;
 295                       if (*np == '$')
 296                         {
 297                           size_t n = 0;
 298
 299                           for (np = cp; *np >= '0' && *np <= '9'; np++)
 300                             n = xsum (xtimes (n, 10), *np - '0');
 301                           if (n == 0)
 302                             /* Positional argument 0.  */
 303                             goto error;
 304                           if (size_overflow_p (n))
 305                             /* n too large, would lead to out of memory
 306                                later.  */
 307                             goto error;
 308                           dp->precision_arg_index = n - 1;
 309                           cp = np + 1;
 310                         }
 311                     }
 312                   if (dp->precision_arg_index == ARG_NONE)
 313                     {
 314                       dp->precision_arg_index = arg_posn++;
 315                       if (dp->precision_arg_index == ARG_NONE)
 316                         /* arg_posn wrapped around.  */
 317                         goto error;
 318                     }
 319                   REGISTER_ARG (dp->precision_arg_index, TYPE_INT);
 320                 }
 321               else
 322                 {
 323                   size_t precision_length;
 324
 325                   dp->precision_start = cp - 1;
 326                   for (; *cp >= '0' && *cp <= '9'; cp++)
 327                     ;
 328                   dp->precision_end = cp;
 329                   precision_length = dp->precision_end - dp->precision_start;
 330                   if (max_precision_length < precision_length)
 331                     max_precision_length = precision_length;
 332                 }
 333             }
 334
 335           {
 336             arg_type type;
 337
 338             /* Parse argument type/size specifiers.  */
 339             {
 340               int flags = 0;
 341
 342               for (;;)
 343                 {
 344                   if (*cp == 'h')
 345                     {
 346                       flags |= (1 << (flags & 1));
 347                       cp++;
 348                     }
 349                   else if (*cp == 'L')
 350                     {
 351                       flags |= 4;
 352                       cp++;
 353                     }
 354                   else if (*cp == 'l')
 355                     {
 356                       flags += 8;
 357                       cp++;
 358                     }
 359                   else if (*cp == 'j')
 360                     {
 361                       if (sizeof (intmax_t) > sizeof (long))
 362                         {
 363                           /* intmax_t = long long */
 364                           flags += 16;
 365                         }
 366                       else if (sizeof (intmax_t) > sizeof (int))
 367                         {
 368                           /* intmax_t = long */
 369                           flags += 8;
 370                         }
 371                       cp++;
 372                     }
 373                   else if (*cp == 'z' || *cp == 'Z')
 374                     {
 375                       /* 'z' is standardized in ISO C 99, but glibc uses 'Z'
 376                          because the warning facility in gcc-2.95.2 understands
 377                          only 'Z' (see gcc-2.95.2/gcc/c-common.c:1784).  */
 378                       if (sizeof (size_t) > sizeof (long))
 379                         {
 380                           /* size_t = long long */
 381                           flags += 16;
 382                         }
 383                       else if (sizeof (size_t) > sizeof (int))
 384                         {
 385                           /* size_t = long */
 386                           flags += 8;
 387                         }
 388                       cp++;
 389                     }
 390                   else if (*cp == 't')
 391                     {
 392                       if (sizeof (ptrdiff_t) > sizeof (long))
 393                         {
 394                           /* ptrdiff_t = long long */
 395                           flags += 16;
 396                         }
 397                       else if (sizeof (ptrdiff_t) > sizeof (int))
 398                         {
 399                           /* ptrdiff_t = long */
 400                           flags += 8;
 401                         }
 402                       cp++;
 403                     }
 404 #if defined __APPLE__ && defined __MACH__
 405                   /* On MacOS X 10.3, PRIdMAX is defined as "qd".
 406                      We cannot change it to "lld" because PRIdMAX must also
 407                      be understood by the system's printf routines.  */
 408                   else if (*cp == 'q')
 409                     {
 410                       if (64 / 8 > sizeof (long))
 411                         {
 412                           /* int64_t = long long */
 413                           flags += 16;
 414                         }
 415                       else
 416                         {
 417                           /* int64_t = long */
 418                           flags += 8;
 419                         }
 420                       cp++;
 421                     }
 422 #endif
 423 #if (defined _WIN32 || defined __WIN32__) && ! defined __CYGWIN__
 424                   /* On native Win32, PRIdMAX is defined as "I64d".
 425                      We cannot change it to "lld" because PRIdMAX must also
 426                      be understood by the system's printf routines.  */
 427                   else if (*cp == 'I' && cp[1] == '6' && cp[2] == '4')
 428                     {
 429                       if (64 / 8 > sizeof (long))
 430                         {
 431                           /* __int64 = long long */
 432                           flags += 16;
 433                         }
 434                       else
 435                         {
 436                           /* __int64 = long */
 437                           flags += 8;
 438                         }
 439                       cp += 3;
 440                     }
 441 #endif
 442                   else
 443                     break;
 444                 }
 445
 446               /* Read the conversion character.  */
 447               c = *cp++;
 448               switch (c)
 449                 {
 450                 case 'd': case 'i':
 451 #if HAVE_LONG_LONG_INT
 452                   /* If 'long long' exists and is larger than 'long':  */
 453                   if (flags >= 16 || (flags & 4))
 454                     type = TYPE_LONGLONGINT;
 455                   else
 456 #endif
 457                   /* If 'long long' exists and is the same as 'long', we parse
 458                      "lld" into TYPE_LONGINT.  */
 459                   if (flags >= 8)
 460                     type = TYPE_LONGINT;
 461                   else if (flags & 2)
 462                     type = TYPE_SCHAR;
 463                   else if (flags & 1)
 464                     type = TYPE_SHORT;
 465                   else
 466                     type = TYPE_INT;
 467                   break;
 468                 case 'o': case 'u': case 'x': case 'X':
 469 #if HAVE_LONG_LONG_INT
 470                   /* If 'long long' exists and is larger than 'long':  */
 471                   if (flags >= 16 || (flags & 4))
 472                     type = TYPE_ULONGLONGINT;
 473                   else
 474 #endif
 475                   /* If 'unsigned long long' exists and is the same as
 476                      'unsigned long', we parse "llu" into TYPE_ULONGINT.  */
 477                   if (flags >= 8)
 478                     type = TYPE_ULONGINT;
 479                   else if (flags & 2)
 480                     type = TYPE_UCHAR;
 481                   else if (flags & 1)
 482                     type = TYPE_USHORT;
 483                   else
 484                     type = TYPE_UINT;
 485                   break;
 486                 case 'f': case 'F': case 'e': case 'E': case 'g': case 'G':
 487                 case 'a': case 'A':
 488                   if (flags >= 16 || (flags & 4))
 489                     type = TYPE_LONGDOUBLE;
 490                   else
 491                     type = TYPE_DOUBLE;
 492                   break;
 493                 case 'c':
 494                   if (flags >= 8)
 495 #if HAVE_WINT_T
 496                     type = TYPE_WIDE_CHAR;
 497 #else
 498                     goto error;
 499 #endif
 500                   else
 501                     type = TYPE_CHAR;
 502                   break;
 503 #if HAVE_WINT_T
 504                 case 'C':
 505                   type = TYPE_WIDE_CHAR;
 506                   c = 'c';
 507                   break;
 508 #endif
 509                 case 's':
 510                   if (flags >= 8)
 511 #if HAVE_WCHAR_T
 512                     type = TYPE_WIDE_STRING;
 513 #else
 514                     goto error;
 515 #endif
 516                   else
 517                     type = TYPE_STRING;
 518                   break;
 519 #if HAVE_WCHAR_T
 520                 case 'S':
 521                   type = TYPE_WIDE_STRING;
 522                   c = 's';
 523                   break;
 524 #endif
 525                 case 'p':
 526                   type = TYPE_POINTER;
 527                   break;
 528                 case 'n':
 529 #if HAVE_LONG_LONG_INT
 530                   /* If 'long long' exists and is larger than 'long':  */
 531                   if (flags >= 16 || (flags & 4))
 532                     type = TYPE_COUNT_LONGLONGINT_POINTER;
 533                   else
 534 #endif
 535                   /* If 'long long' exists and is the same as 'long', we parse
 536                      "lln" into TYPE_COUNT_LONGINT_POINTER.  */
 537                   if (flags >= 8)
 538                     type = TYPE_COUNT_LONGINT_POINTER;
 539                   else if (flags & 2)
 540                     type = TYPE_COUNT_SCHAR_POINTER;
 541                   else if (flags & 1)
 542                     type = TYPE_COUNT_SHORT_POINTER;
 543                   else
 544                     type = TYPE_COUNT_INT_POINTER;
 545                   break;
 546 #if ENABLE_UNISTDIO
 547                 /* The unistdio extensions.  */
 548                 case 'U':
 549                   if (flags >= 16)
 550                     type = TYPE_U32_STRING;
 551                   else if (flags >= 8)
 552                     type = TYPE_U16_STRING;
 553                   else
 554                     type = TYPE_U8_STRING;
 555                   break;
 556 #endif
 557                 case '%':
 558                   type = TYPE_NONE;
 559                   break;
 560                 default:
 561                   /* Unknown conversion character.  */
 562                   goto error;
 563                 }
 564             }
 565
 566             if (type != TYPE_NONE)
 567               {
 568                 dp->arg_index = arg_index;
 569                 if (dp->arg_index == ARG_NONE)
 570                   {
 571                     dp->arg_index = arg_posn++;
 572                     if (dp->arg_index == ARG_NONE)
 573                       /* arg_posn wrapped around.  */
 574                       goto error;
 575                   }
 576                 REGISTER_ARG (dp->arg_index, type);
 577               }
 578             dp->conversion = c;
 579             dp->dir_end = cp;
 580           }
 581
 582           d->count++;
 583           if (d->count >= d_allocated)
 584             {
 585               size_t memory_size;
 586               DIRECTIVE *memory;
 587
 588               d_allocated = xtimes (d_allocated, 2);
 589               memory_size = xtimes (d_allocated, sizeof (DIRECTIVE));
 590               if (size_overflow_p (memory_size))
 591                 /* Overflow, would lead to out of memory.  */
 592                 goto out_of_memory;
 593               memory = (DIRECTIVE *) (d->dir != d->direct_alloc_dir
 594                                       ? realloc (d->dir, memory_size)
 595                                       : malloc (memory_size));
 596               if (memory == NULL)
 597                 /* Out of memory.  */
 598                 goto out_of_memory;
 599               if (d->dir == d->direct_alloc_dir)
 600                 memcpy (memory, d->dir, d->count * sizeof (DIRECTIVE));
 601               d->dir = memory;
 602             }
 603         }
 604 #if CHAR_T_ONLY_ASCII
 605       else if (!c_isascii (c))
 606         {
 607           /* Non-ASCII character.  Not supported.  */
 608           goto error;
 609         }
 610 #endif
 611     }
 612   d->dir[d->count].dir_start = cp;
 613
 614   d->max_width_length = max_width_length;
 615   d->max_precision_length = max_precision_length;
 616   return 0;
 617
 618 error:
 619   if (a->arg != a->direct_alloc_arg)
 620     free (a->arg);
 621   if (d->dir != d->direct_alloc_dir)
 622     free (d->dir);
 623   errno = EINVAL;
 624   return -1;
 625
 626 out_of_memory:
 627   if (a->arg != a->direct_alloc_arg)
 628     free (a->arg);
 629   if (d->dir != d->direct_alloc_dir)
 630     free (d->dir);
 631   errno = ENOMEM;
 632   return -1;
 633 }
 634
 635 #undef PRINTF_PARSE
 636 #undef DIRECTIVES
 637 #undef DIRECTIVE
 638 #undef CHAR_T_ONLY_ASCII
 639 #undef CHAR_T