lgl/printf-parse.c

   1 /* Formatted output to strings.
   2    Copyright (C) 1999-2000, 2002-2003, 2006-2008 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 2.1, 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 /* errno.  */
  67 #include <errno.h>
  68
  69 /* Checked size_t computations.  */
  70 #include "xsize.h"
  71
  72 #if CHAR_T_ONLY_ASCII
  73 /* c_isascii().  */
  74 # include "c-ctype.h"
  75 #endif
  76
  77 #ifdef STATIC
  78 STATIC
  79 #endif
  80 int
  81 PRINTF_PARSE (const CHAR_T *format, DIRECTIVES *d, arguments *a)
  82 {
  83   const CHAR_T *cp = format;            /* pointer into format */
  84   size_t arg_posn = 0;          /* number of regular arguments consumed */
  85   size_t d_allocated;                   /* allocated elements of d->dir */
  86   size_t a_allocated;                   /* allocated elements of a->arg */
  87   size_t max_width_length = 0;
  88   size_t max_precision_length = 0;
  89
  90   d->count = 0;
  91   d_allocated = 1;
  92   d->dir = (DIRECTIVE *) malloc (d_allocated * sizeof (DIRECTIVE));
  93   if (d->dir == NULL)
  94     /* Out of memory.  */
  95     goto out_of_memory_1;
  96
  97   a->count = 0;
  98   a_allocated = 0;
  99   a->arg = NULL;
 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                                   \
 117                                ? realloc (a->arg, memory_size)          \
 118                                : malloc (memory_size));                 \
 119         if (memory == NULL)                                             \
 120           /* Out of memory.  */                                         \
 121           goto out_of_memory;                                           \
 122         a->arg = memory;                                                \
 123       }                                                                 \
 124     while (a->count <= n)                                               \
 125       a->arg[a->count++].type = TYPE_NONE;                              \
 126     if (a->arg[n].type == TYPE_NONE)                                    \
 127       a->arg[n].type = (_type_);                                        \
 128     else if (a->arg[n].type != (_type_))                                \
 129       /* Ambiguous type for positional argument.  */                    \
 130       goto error;                                                       \
 131   }
 132
 133   while (*cp != '\0')
 134     {
 135       CHAR_T c = *cp++;
 136       if (c == '%')
 137         {
 138           size_t arg_index = ARG_NONE;
 139           DIRECTIVE *dp = &d->dir[d->count]; /* pointer to next directive */
 140
 141           /* Initialize the next directive.  */
 142           dp->dir_start = cp - 1;
 143           dp->flags = 0;
 144           dp->width_start = NULL;
 145           dp->width_end = NULL;
 146           dp->width_arg_index = ARG_NONE;
 147           dp->precision_start = NULL;
 148           dp->precision_end = NULL;
 149           dp->precision_arg_index = ARG_NONE;
 150           dp->arg_index = ARG_NONE;
 151
 152           /* Test for positional argument.  */
 153           if (*cp >= '0' && *cp <= '9')
 154             {
 155               const CHAR_T *np;
 156
 157               for (np = cp; *np >= '0' && *np <= '9'; np++)
 158                 ;
 159               if (*np == '$')
 160                 {
 161                   size_t n = 0;
 162
 163                   for (np = cp; *np >= '0' && *np <= '9'; np++)
 164                     n = xsum (xtimes (n, 10), *np - '0');
 165                   if (n == 0)
 166                     /* Positional argument 0.  */
 167                     goto error;
 168                   if (size_overflow_p (n))
 169                     /* n too large, would lead to out of memory later.  */
 170                     goto error;
 171                   arg_index = n - 1;
 172                   cp = np + 1;
 173                 }
 174             }
 175
 176           /* Read the flags.  */
 177           for (;;)
 178             {
 179               if (*cp == '\'')
 180                 {
 181                   dp->flags |= FLAG_GROUP;
 182                   cp++;
 183                 }
 184               else if (*cp == '-')
 185                 {
 186                   dp->flags |= FLAG_LEFT;
 187                   cp++;
 188                 }
 189               else if (*cp == '+')
 190                 {
 191                   dp->flags |= FLAG_SHOWSIGN;
 192                   cp++;
 193                 }
 194               else if (*cp == ' ')
 195                 {
 196                   dp->flags |= FLAG_SPACE;
 197                   cp++;
 198                 }
 199               else if (*cp == '#')
 200                 {
 201                   dp->flags |= FLAG_ALT;
 202                   cp++;
 203                 }
 204               else if (*cp == '0')
 205                 {
 206                   dp->flags |= FLAG_ZERO;
 207                   cp++;
 208                 }
 209               else
 210                 break;
 211             }
 212
 213           /* Parse the field width.  */
 214           if (*cp == '*')
 215             {
 216               dp->width_start = cp;
 217               cp++;
 218               dp->width_end = cp;
 219               if (max_width_length < 1)
 220                 max_width_length = 1;
 221
 222               /* Test for positional argument.  */
 223               if (*cp >= '0' && *cp <= '9')
 224                 {
 225                   const CHAR_T *np;
 226
 227                   for (np = cp; *np >= '0' && *np <= '9'; np++)
 228                     ;
 229                   if (*np == '$')
 230                     {
 231                       size_t n = 0;
 232
 233                       for (np = cp; *np >= '0' && *np <= '9'; np++)
 234                         n = xsum (xtimes (n, 10), *np - '0');
 235                       if (n == 0)
 236                         /* Positional argument 0.  */
 237                         goto error;
 238                       if (size_overflow_p (n))
 239                         /* n too large, would lead to out of memory later.  */
 240                         goto error;
 241                       dp->width_arg_index = n - 1;
 242                       cp = np + 1;
 243                     }
 244                 }
 245               if (dp->width_arg_index == ARG_NONE)
 246                 {
 247                   dp->width_arg_index = arg_posn++;
 248                   if (dp->width_arg_index == ARG_NONE)
 249                     /* arg_posn wrapped around.  */
 250                     goto error;
 251                 }
 252               REGISTER_ARG (dp->width_arg_index, TYPE_INT);
 253             }
 254           else if (*cp >= '0' && *cp <= '9')
 255             {
 256               size_t width_length;
 257
 258               dp->width_start = cp;
 259               for (; *cp >= '0' && *cp <= '9'; cp++)
 260                 ;
 261               dp->width_end = cp;
 262               width_length = dp->width_end - dp->width_start;
 263               if (max_width_length < width_length)
 264                 max_width_length = width_length;
 265             }
 266
 267           /* Parse the precision.  */
 268           if (*cp == '.')
 269             {
 270               cp++;
 271               if (*cp == '*')
 272                 {
 273                   dp->precision_start = cp - 1;
 274                   cp++;
 275                   dp->precision_end = cp;
 276                   if (max_precision_length < 2)
 277                     max_precision_length = 2;
 278
 279                   /* Test for positional argument.  */
 280                   if (*cp >= '0' && *cp <= '9')
 281                     {
 282                       const CHAR_T *np;
 283
 284                       for (np = cp; *np >= '0' && *np <= '9'; np++)
 285                         ;
 286                       if (*np == '$')
 287                         {
 288                           size_t n = 0;
 289
 290                           for (np = cp; *np >= '0' && *np <= '9'; np++)
 291                             n = xsum (xtimes (n, 10), *np - '0');
 292                           if (n == 0)
 293                             /* Positional argument 0.  */
 294                             goto error;
 295                           if (size_overflow_p (n))
 296                             /* n too large, would lead to out of memory
 297                                later.  */
 298                             goto error;
 299                           dp->precision_arg_index = n - 1;
 300                           cp = np + 1;
 301                         }
 302                     }
 303                   if (dp->precision_arg_index == ARG_NONE)
 304                     {
 305                       dp->precision_arg_index = arg_posn++;
 306                       if (dp->precision_arg_index == ARG_NONE)
 307                         /* arg_posn wrapped around.  */
 308                         goto error;
 309                     }
 310                   REGISTER_ARG (dp->precision_arg_index, TYPE_INT);
 311                 }
 312               else
 313                 {
 314                   size_t precision_length;
 315
 316                   dp->precision_start = cp - 1;
 317                   for (; *cp >= '0' && *cp <= '9'; cp++)
 318                     ;
 319                   dp->precision_end = cp;
 320                   precision_length = dp->precision_end - dp->precision_start;
 321                   if (max_precision_length < precision_length)
 322                     max_precision_length = precision_length;
 323                 }
 324             }
 325
 326           {
 327             arg_type type;
 328
 329             /* Parse argument type/size specifiers.  */
 330             {
 331               int flags = 0;
 332
 333               for (;;)
 334                 {
 335                   if (*cp == 'h')
 336                     {
 337                       flags |= (1 << (flags & 1));
 338                       cp++;
 339                     }
 340                   else if (*cp == 'L')
 341                     {
 342                       flags |= 4;
 343                       cp++;
 344                     }
 345                   else if (*cp == 'l')
 346                     {
 347                       flags += 8;
 348                       cp++;
 349                     }
 350                   else if (*cp == 'j')
 351                     {
 352                       if (sizeof (intmax_t) > sizeof (long))
 353                         {
 354                           /* intmax_t = long long */
 355                           flags += 16;
 356                         }
 357                       else if (sizeof (intmax_t) > sizeof (int))
 358                         {
 359                           /* intmax_t = long */
 360                           flags += 8;
 361                         }
 362                       cp++;
 363                     }
 364                   else if (*cp == 'z' || *cp == 'Z')
 365                     {
 366                       /* 'z' is standardized in ISO C 99, but glibc uses 'Z'
 367                          because the warning facility in gcc-2.95.2 understands
 368                          only 'Z' (see gcc-2.95.2/gcc/c-common.c:1784).  */
 369                       if (sizeof (size_t) > sizeof (long))
 370                         {
 371                           /* size_t = long long */
 372                           flags += 16;
 373                         }
 374                       else if (sizeof (size_t) > sizeof (int))
 375                         {
 376                           /* size_t = long */
 377                           flags += 8;
 378                         }
 379                       cp++;
 380                     }
 381                   else if (*cp == 't')
 382                     {
 383                       if (sizeof (ptrdiff_t) > sizeof (long))
 384                         {
 385                           /* ptrdiff_t = long long */
 386                           flags += 16;
 387                         }
 388                       else if (sizeof (ptrdiff_t) > sizeof (int))
 389                         {
 390                           /* ptrdiff_t = long */
 391                           flags += 8;
 392                         }
 393                       cp++;
 394                     }
 395 #if defined __APPLE__ && defined __MACH__
 396                   /* On MacOS X 10.3, PRIdMAX is defined as "qd".
 397                      We cannot change it to "lld" because PRIdMAX must also
 398                      be understood by the system's printf routines.  */
 399                   else if (*cp == 'q')
 400                     {
 401                       if (64 / 8 > sizeof (long))
 402                         {
 403                           /* int64_t = long long */
 404                           flags += 16;
 405                         }
 406                       else
 407                         {
 408                           /* int64_t = long */
 409                           flags += 8;
 410                         }
 411                       cp++;
 412                     }
 413 #endif
 414 #if (defined _WIN32 || defined __WIN32__) && ! defined __CYGWIN__
 415                   /* On native Win32, PRIdMAX is defined as "I64d".
 416                      We cannot change it to "lld" because PRIdMAX must also
 417                      be understood by the system's printf routines.  */
 418                   else if (*cp == 'I' && cp[1] == '6' && cp[2] == '4')
 419                     {
 420                       if (64 / 8 > sizeof (long))
 421                         {
 422                           /* __int64 = long long */
 423                           flags += 16;
 424                         }
 425                       else
 426                         {
 427                           /* __int64 = long */
 428                           flags += 8;
 429                         }
 430                       cp += 3;
 431                     }
 432 #endif
 433                   else
 434                     break;
 435                 }
 436
 437               /* Read the conversion character.  */
 438               c = *cp++;
 439               switch (c)
 440                 {
 441                 case 'd': case 'i':
 442 #if HAVE_LONG_LONG_INT
 443                   /* If 'long long' exists and is larger than 'long':  */
 444                   if (flags >= 16 || (flags & 4))
 445                     type = TYPE_LONGLONGINT;
 446                   else
 447 #endif
 448                   /* If 'long long' exists and is the same as 'long', we parse
 449                      "lld" into TYPE_LONGINT.  */
 450                   if (flags >= 8)
 451                     type = TYPE_LONGINT;
 452                   else if (flags & 2)
 453                     type = TYPE_SCHAR;
 454                   else if (flags & 1)
 455                     type = TYPE_SHORT;
 456                   else
 457                     type = TYPE_INT;
 458                   break;
 459                 case 'o': case 'u': case 'x': case 'X':
 460 #if HAVE_LONG_LONG_INT
 461                   /* If 'long long' exists and is larger than 'long':  */
 462                   if (flags >= 16 || (flags & 4))
 463                     type = TYPE_ULONGLONGINT;
 464                   else
 465 #endif
 466                   /* If 'unsigned long long' exists and is the same as
 467                      'unsigned long', we parse "llu" into TYPE_ULONGINT.  */
 468                   if (flags >= 8)
 469                     type = TYPE_ULONGINT;
 470                   else if (flags & 2)
 471                     type = TYPE_UCHAR;
 472                   else if (flags & 1)
 473                     type = TYPE_USHORT;
 474                   else
 475                     type = TYPE_UINT;
 476                   break;
 477                 case 'f': case 'F': case 'e': case 'E': case 'g': case 'G':
 478                 case 'a': case 'A':
 479                   if (flags >= 16 || (flags & 4))
 480                     type = TYPE_LONGDOUBLE;
 481                   else
 482                     type = TYPE_DOUBLE;
 483                   break;
 484                 case 'c':
 485                   if (flags >= 8)
 486 #if HAVE_WINT_T
 487                     type = TYPE_WIDE_CHAR;
 488 #else
 489                     goto error;
 490 #endif
 491                   else
 492                     type = TYPE_CHAR;
 493                   break;
 494 #if HAVE_WINT_T
 495                 case 'C':
 496                   type = TYPE_WIDE_CHAR;
 497                   c = 'c';
 498                   break;
 499 #endif
 500                 case 's':
 501                   if (flags >= 8)
 502 #if HAVE_WCHAR_T
 503                     type = TYPE_WIDE_STRING;
 504 #else
 505                     goto error;
 506 #endif
 507                   else
 508                     type = TYPE_STRING;
 509                   break;
 510 #if HAVE_WCHAR_T
 511                 case 'S':
 512                   type = TYPE_WIDE_STRING;
 513                   c = 's';
 514                   break;
 515 #endif
 516                 case 'p':
 517                   type = TYPE_POINTER;
 518                   break;
 519                 case 'n':
 520 #if HAVE_LONG_LONG_INT
 521                   /* If 'long long' exists and is larger than 'long':  */
 522                   if (flags >= 16 || (flags & 4))
 523                     type = TYPE_COUNT_LONGLONGINT_POINTER;
 524                   else
 525 #endif
 526                   /* If 'long long' exists and is the same as 'long', we parse
 527                      "lln" into TYPE_COUNT_LONGINT_POINTER.  */
 528                   if (flags >= 8)
 529                     type = TYPE_COUNT_LONGINT_POINTER;
 530                   else if (flags & 2)
 531                     type = TYPE_COUNT_SCHAR_POINTER;
 532                   else if (flags & 1)
 533                     type = TYPE_COUNT_SHORT_POINTER;
 534                   else
 535                     type = TYPE_COUNT_INT_POINTER;
 536                   break;
 537 #if ENABLE_UNISTDIO
 538                 /* The unistdio extensions.  */
 539                 case 'U':
 540                   if (flags >= 16)
 541                     type = TYPE_U32_STRING;
 542                   else if (flags >= 8)
 543                     type = TYPE_U16_STRING;
 544                   else
 545                     type = TYPE_U8_STRING;
 546                   break;
 547 #endif
 548                 case '%':
 549                   type = TYPE_NONE;
 550                   break;
 551                 default:
 552                   /* Unknown conversion character.  */
 553                   goto error;
 554                 }
 555             }
 556
 557             if (type != TYPE_NONE)
 558               {
 559                 dp->arg_index = arg_index;
 560                 if (dp->arg_index == ARG_NONE)
 561                   {
 562                     dp->arg_index = arg_posn++;
 563                     if (dp->arg_index == ARG_NONE)
 564                       /* arg_posn wrapped around.  */
 565                       goto error;
 566                   }
 567                 REGISTER_ARG (dp->arg_index, type);
 568               }
 569             dp->conversion = c;
 570             dp->dir_end = cp;
 571           }
 572
 573           d->count++;
 574           if (d->count >= d_allocated)
 575             {
 576               size_t memory_size;
 577               DIRECTIVE *memory;
 578
 579               d_allocated = xtimes (d_allocated, 2);
 580               memory_size = xtimes (d_allocated, sizeof (DIRECTIVE));
 581               if (size_overflow_p (memory_size))
 582                 /* Overflow, would lead to out of memory.  */
 583                 goto out_of_memory;
 584               memory = (DIRECTIVE *) realloc (d->dir, memory_size);
 585               if (memory == NULL)
 586                 /* Out of memory.  */
 587                 goto out_of_memory;
 588               d->dir = memory;
 589             }
 590         }
 591 #if CHAR_T_ONLY_ASCII
 592       else if (!c_isascii (c))
 593         {
 594           /* Non-ASCII character.  Not supported.  */
 595           goto error;
 596         }
 597 #endif
 598     }
 599   d->dir[d->count].dir_start = cp;
 600
 601   d->max_width_length = max_width_length;
 602   d->max_precision_length = max_precision_length;
 603   return 0;
 604
 605 error:
 606   if (a->arg)
 607     free (a->arg);
 608   if (d->dir)
 609     free (d->dir);
 610   errno = EINVAL;
 611   return -1;
 612
 613 out_of_memory:
 614   if (a->arg)
 615     free (a->arg);
 616   if (d->dir)
 617     free (d->dir);
 618 out_of_memory_1:
 619   errno = ENOMEM;
 620   return -1;
 621 }
 622
 623 #undef PRINTF_PARSE
 624 #undef DIRECTIVES
 625 #undef DIRECTIVE
 626 #undef CHAR_T_ONLY_ASCII
 627 #undef CHAR_T