gl/string_.h

   1 /* A GNU-like <string.h>.
   2
   3    Copyright (C) 1995-1996, 2001-2007 Free Software Foundation, Inc.
   4
   5    This program is free software; you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published by
   7    the Free Software Foundation; either version 2, or (at your option)
   8    any later version.
   9
  10    This program 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
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program; if not, write to the Free Software Foundation,
  17    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  18
  19 #ifdef __DECC
  20 # include_next <string.h>
  21 #endif
  22
  23 #ifndef _GL_STRING_H
  24 #define _GL_STRING_H
  25
  26 /* This #pragma avoids a warning with "gcc -Wmissing-prototypes" on some
  27    mingw systems.  */
  28 #ifdef __GNUC__
  29 # pragma GCC system_header
  30 #endif
  31
  32 #ifndef __DECC
  33 # include @ABSOLUTE_STRING_H@
  34 #endif
  35
  36
  37 /* The definition of GL_LINK_WARNING is copied here.  */
  38
  39
  40 #ifdef __cplusplus
  41 extern "C" {
  42 #endif
  43
  44 /* Return the first occurrence of NEEDLE in HAYSTACK.  */
  45 #if @GNULIB_MEMMEM@
  46 # if ! @HAVE_DECL_MEMMEM@
  47 extern void *memmem (void const *__haystack, size_t __haystack_len,
  48                      void const *__needle, size_t __needle_len);
  49 # endif
  50 #elif defined GNULIB_POSIXCHECK
  51 # undef memmem
  52 # define memmem(a,al,b,bl) \
  53     (GL_LINK_WARNING ("memmem is unportable - " \
  54                       "use gnulib module memmem for portability"), \
  55      memmem (a, al, b, bl))
  56 #endif
  57
  58 /* Copy N bytes of SRC to DEST, return pointer to bytes after the
  59    last written byte.  */
  60 #if @GNULIB_MEMPCPY@
  61 # if ! @HAVE_MEMPCPY@
  62 extern void *mempcpy (void *restrict __dest, void const *restrict __src,
  63                       size_t __n);
  64 # endif
  65 #elif defined GNULIB_POSIXCHECK
  66 # undef mempcpy
  67 # define mempcpy(a,b,n) \
  68     (GL_LINK_WARNING ("mempcpy is unportable - " \
  69                       "use gnulib module mempcpy for portability"), \
  70      mempcpy (a, b, n))
  71 #endif
  72
  73 /* Search backwards through a block for a byte (specified as an int).  */
  74 #if @GNULIB_MEMRCHR@
  75 # if ! @HAVE_DECL_MEMRCHR@
  76 extern void *memrchr (void const *, int, size_t);
  77 # endif
  78 #elif defined GNULIB_POSIXCHECK
  79 # undef memrchr
  80 # define memrchr(a,b,c) \
  81     (GL_LINK_WARNING ("memrchr is unportable - " \
  82                       "use gnulib module memrchr for portability"), \
  83      memrchr (a, b, c))
  84 #endif
  85
  86 /* Copy SRC to DST, returning the address of the terminating '\0' in DST.  */
  87 #if @GNULIB_STPCPY@
  88 # if ! @HAVE_STPCPY@
  89 extern char *stpcpy (char *restrict __dst, char const *restrict __src);
  90 # endif
  91 #elif defined GNULIB_POSIXCHECK
  92 # undef stpcpy
  93 # define stpcpy(a,b) \
  94     (GL_LINK_WARNING ("stpcpy is unportable - " \
  95                       "use gnulib module stpcpy for portability"), \
  96      stpcpy (a, b))
  97 #endif
  98
  99 /* Copy no more than N bytes of SRC to DST, returning a pointer past the
 100    last non-NUL byte written into DST.  */
 101 #if @GNULIB_STPNCPY@
 102 # if ! @HAVE_STPNCPY@
 103 #  define stpncpy gnu_stpncpy
 104 extern char *stpncpy (char *restrict __dst, char const *restrict __src,
 105                       size_t __n);
 106 # endif
 107 #elif defined GNULIB_POSIXCHECK
 108 # undef stpncpy
 109 # define stpncpy(a,b,n) \
 110     (GL_LINK_WARNING ("stpncpy is unportable - " \
 111                       "use gnulib module stpncpy for portability"), \
 112      stpncpy (a, b, n))
 113 #endif
 114
 115 /* Compare strings S1 and S2, ignoring case, returning less than, equal to or
 116    greater than zero if S1 is lexicographically less than, equal to or greater
 117    than S2.
 118    Note: This function does not work in multibyte locales.  */
 119 #if ! @HAVE_STRCASECMP@
 120 extern int strcasecmp (char const *s1, char const *s2);
 121 #endif
 122 #if defined GNULIB_POSIXCHECK
 123 /* strcasecmp() does not work with multibyte strings:
 124    POSIX says that it operates on "strings", and "string" in POSIX is defined
 125    as a sequence of bytes, not of characters.   */
 126 # undef strcasecmp
 127 # define strcasecmp(a,b) \
 128     (GL_LINK_WARNING ("strcasecmp cannot work correctly on character strings " \
 129                       "in multibyte locales - " \
 130                       "use mbscasecmp if you care about " \
 131                       "internationalization, or use c_strcasecmp (from " \
 132                       "gnulib module c-strcase) if you want a locale " \
 133                       "independent function"), \
 134      strcasecmp (a, b))
 135 #endif
 136
 137 /* Compare no more than N bytes of strings S1 and S2, ignoring case,
 138    returning less than, equal to or greater than zero if S1 is
 139    lexicographically less than, equal to or greater than S2.
 140    Note: This function cannot work correctly in multibyte locales.  */
 141 #if ! @HAVE_DECL_STRNCASECMP@
 142 extern int strncasecmp (char const *s1, char const *s2, size_t n);
 143 #endif
 144 #if defined GNULIB_POSIXCHECK
 145 /* strncasecmp() does not work with multibyte strings:
 146    POSIX says that it operates on "strings", and "string" in POSIX is defined
 147    as a sequence of bytes, not of characters.  */
 148 # undef strncasecmp
 149 # define strncasecmp(a,b,n) \
 150     (GL_LINK_WARNING ("strncasecmp cannot work correctly on character " \
 151                       "strings in multibyte locales - " \
 152                       "use mbsncasecmp or mbspcasecmp if you care about " \
 153                       "internationalization, or use c_strncasecmp (from " \
 154                       "gnulib module c-strcase) if you want a locale " \
 155                       "independent function"), \
 156      strncasecmp (a, b, n))
 157 #endif
 158
 159 #if defined GNULIB_POSIXCHECK
 160 /* strchr() does not work with multibyte strings if the locale encoding is
 161    GB18030 and the character to be searched is a digit.  */
 162 # undef strchr
 163 # define strchr(s,c) \
 164     (GL_LINK_WARNING ("strchr cannot work correctly on character strings " \
 165                       "in some multibyte locales - " \
 166                       "use mbschr if you care about internationalization"), \
 167      strchr (s, c))
 168 #endif
 169
 170 /* Find the first occurrence of C in S or the final NUL byte.  */
 171 #if @GNULIB_STRCHRNUL@
 172 # if ! @HAVE_STRCHRNUL@
 173 extern char *strchrnul (char const *__s, int __c_in);
 174 # endif
 175 #elif defined GNULIB_POSIXCHECK
 176 # undef strchrnul
 177 # define strchrnul(a,b) \
 178     (GL_LINK_WARNING ("strchrnul is unportable - " \
 179                       "use gnulib module strchrnul for portability"), \
 180      strchrnul (a, b))
 181 #endif
 182
 183 /* Duplicate S, returning an identical malloc'd string.  */
 184 #if @GNULIB_STRDUP@
 185 # if ! @HAVE_DECL_STRDUP@ && ! defined strdup
 186 extern char *strdup (char const *__s);
 187 # endif
 188 #elif defined GNULIB_POSIXCHECK
 189 # undef strdup
 190 # define strdup(a) \
 191     (GL_LINK_WARNING ("strdup is unportable - " \
 192                       "use gnulib module strdup for portability"), \
 193      strdup (a))
 194 #endif
 195
 196 /* Return a newly allocated copy of at most N bytes of STRING.  */
 197 #if @GNULIB_STRNDUP@
 198 # if ! @HAVE_STRNDUP@
 199 #  undef strndup
 200 #  define strndup rpl_strndup
 201 # endif
 202 # if ! @HAVE_STRNDUP@ || ! @HAVE_DECL_STRNDUP@
 203 extern char *strndup (char const *__string, size_t __n);
 204 # endif
 205 #elif defined GNULIB_POSIXCHECK
 206 # undef strndup
 207 # define strndup(a,n) \
 208     (GL_LINK_WARNING ("strndup is unportable - " \
 209                       "use gnulib module strndup for portability"), \
 210      strndup (a, n))
 211 #endif
 212
 213 /* Find the length (number of bytes) of STRING, but scan at most
 214    MAXLEN bytes.  If no '\0' terminator is found in that many bytes,
 215    return MAXLEN.  */
 216 #if @GNULIB_STRNLEN@
 217 # if ! @HAVE_DECL_STRNLEN@
 218 extern size_t strnlen (char const *__string, size_t __maxlen);
 219 # endif
 220 #elif defined GNULIB_POSIXCHECK
 221 # undef strnlen
 222 # define strnlen(a,n) \
 223     (GL_LINK_WARNING ("strnlen is unportable - " \
 224                       "use gnulib module strnlen for portability"), \
 225      strnlen (a, n))
 226 #endif
 227
 228 #if defined GNULIB_POSIXCHECK
 229 /* strcspn() assumes the second argument is a list of single-byte characters.
 230    Even in this simple case, it does not work with multibyte strings if the
 231    locale encoding is GB18030 and one of the characters to be searched is a
 232    digit.  */
 233 # undef strcspn
 234 # define strcspn(s,a) \
 235     (GL_LINK_WARNING ("strcspn cannot work correctly on character strings " \
 236                       "in multibyte locales - " \
 237                       "use mbscspn if you care about internationalization"), \
 238      strcspn (s, a))
 239 #endif
 240
 241 /* Find the first occurrence in S of any character in ACCEPT.  */
 242 #if @GNULIB_STRPBRK@
 243 # if ! @HAVE_STRPBRK@
 244 extern char *strpbrk (char const *__s, char const *__accept);
 245 # endif
 246 # if defined GNULIB_POSIXCHECK
 247 /* strpbrk() assumes the second argument is a list of single-byte characters.
 248    Even in this simple case, it does not work with multibyte strings if the
 249    locale encoding is GB18030 and one of the characters to be searched is a
 250    digit.  */
 251 #  undef strpbrk
 252 #  define strpbrk(s,a) \
 253      (GL_LINK_WARNING ("strpbrk cannot work correctly on character strings " \
 254                        "in multibyte locales - " \
 255                        "use mbspbrk if you care about internationalization"), \
 256       strpbrk (s, a))
 257 # endif
 258 #elif defined GNULIB_POSIXCHECK
 259 # undef strpbrk
 260 # define strpbrk(s,a) \
 261     (GL_LINK_WARNING ("strpbrk is unportable - " \
 262                       "use gnulib module strpbrk for portability"), \
 263      strpbrk (s, a))
 264 #endif
 265
 266 #if defined GNULIB_POSIXCHECK
 267 /* strspn() assumes the second argument is a list of single-byte characters.
 268    Even in this simple case, it cannot work with multibyte strings.  */
 269 # undef strspn
 270 # define strspn(s,a) \
 271     (GL_LINK_WARNING ("strspn cannot work correctly on character strings " \
 272                       "in multibyte locales - " \
 273                       "use mbsspn if you care about internationalization"), \
 274      strspn (s, a))
 275 #endif
 276
 277 #if defined GNULIB_POSIXCHECK
 278 /* strrchr() does not work with multibyte strings if the locale encoding is
 279    GB18030 and the character to be searched is a digit.  */
 280 # undef strrchr
 281 # define strrchr(s,c) \
 282     (GL_LINK_WARNING ("strrchr cannot work correctly on character strings " \
 283                       "in some multibyte locales - " \
 284                       "use mbsrchr if you care about internationalization"), \
 285      strrchr (s, c))
 286 #endif
 287
 288 /* Search the next delimiter (char listed in DELIM) starting at *STRINGP.
 289    If one is found, overwrite it with a NUL, and advance *STRINGP
 290    to point to the next char after it.  Otherwise, set *STRINGP to NULL.
 291    If *STRINGP was already NULL, nothing happens.
 292    Return the old value of *STRINGP.
 293
 294    This is a variant of strtok() that is multithread-safe and supports
 295    empty fields.
 296
 297    Caveat: It modifies the original string.
 298    Caveat: These functions cannot be used on constant strings.
 299    Caveat: The identity of the delimiting character is lost.
 300    Caveat: It doesn't work with multibyte strings unless all of the delimiter
 301            characters are ASCII characters < 0x30.
 302
 303    See also strtok_r().  */
 304 #if @GNULIB_STRSEP@
 305 # if ! @HAVE_STRSEP@
 306 extern char *strsep (char **restrict __stringp, char const *restrict __delim);
 307 # endif
 308 # if defined GNULIB_POSIXCHECK
 309 #  undef strsep
 310 #  define strsep(s,d) \
 311      (GL_LINK_WARNING ("strsep cannot work correctly on character strings " \
 312                        "in multibyte locales - " \
 313                        "use mbssep if you care about internationalization"), \
 314       strsep (s, d))
 315 # endif
 316 #elif defined GNULIB_POSIXCHECK
 317 # undef strsep
 318 # define strsep(s,d) \
 319     (GL_LINK_WARNING ("strsep is unportable - " \
 320                       "use gnulib module strsep for portability"), \
 321      strsep (s, d))
 322 #endif
 323
 324 #if defined GNULIB_POSIXCHECK
 325 /* strstr() does not work with multibyte strings if the locale encoding is
 326    different from UTF-8:
 327    POSIX says that it operates on "strings", and "string" in POSIX is defined
 328    as a sequence of bytes, not of characters.  */
 329 # undef strstr
 330 # define strstr(a,b) \
 331     (GL_LINK_WARNING ("strstr cannot work correctly on character strings " \
 332                       "in most multibyte locales - " \
 333                       "use mbsstr if you care about internationalization"), \
 334      strstr (a, b))
 335 #endif
 336
 337 /* Find the first occurrence of NEEDLE in HAYSTACK, using case-insensitive
 338    comparison.  */
 339 #if ! @HAVE_STRCASESTR@
 340 extern char *strcasestr (const char *haystack, const char *needle);
 341 #endif
 342 #if defined GNULIB_POSIXCHECK
 343 /* strcasestr() does not work with multibyte strings:
 344    It is a glibc extension, and glibc implements it only for unibyte
 345    locales.  */
 346 # undef strcasestr
 347 # define strcasestr(a,b) \
 348     (GL_LINK_WARNING ("strcasestr does work correctly on character strings " \
 349                       "in multibyte locales - " \
 350                       "use mbscasestr if you care about " \
 351                       "internationalization, or use c-strcasestr if you want " \
 352                       "a locale independent function"), \
 353      strcasestr (a, b))
 354 #endif
 355
 356 /* Parse S into tokens separated by characters in DELIM.
 357    If S is NULL, the saved pointer in SAVE_PTR is used as
 358    the next starting point.  For example:
 359         char s[] = "-abc-=-def";
 360         char *sp;
 361         x = strtok_r(s, "-", &sp);      // x = "abc", sp = "=-def"
 362         x = strtok_r(NULL, "-=", &sp);  // x = "def", sp = NULL
 363         x = strtok_r(NULL, "=", &sp);   // x = NULL
 364                 // s = "abc\0-def\0"
 365
 366    This is a variant of strtok() that is multithread-safe.
 367
 368    For the POSIX documentation for this function, see:
 369    http://www.opengroup.org/susv3xsh/strtok.html
 370
 371    Caveat: It modifies the original string.
 372    Caveat: These functions cannot be used on constant strings.
 373    Caveat: The identity of the delimiting character is lost.
 374    Caveat: It doesn't work with multibyte strings unless all of the delimiter
 375            characters are ASCII characters < 0x30.
 376
 377    See also strsep().  */
 378 #if @GNULIB_STRTOK_R@
 379 # if ! @HAVE_DECL_STRTOK_R@
 380 extern char *strtok_r (char *restrict s, char const *restrict delim,
 381                        char **restrict save_ptr);
 382 # endif
 383 # if defined GNULIB_POSIXCHECK
 384 #  undef strtok_r
 385 #  define strtok_r(s,d,p) \
 386      (GL_LINK_WARNING ("strtok_r cannot work correctly on character strings " \
 387                        "in multibyte locales - " \
 388                        "use mbstok_r if you care about internationalization"), \
 389       strtok_r (s, d, p))
 390 # endif
 391 #elif defined GNULIB_POSIXCHECK
 392 # undef strtok_r
 393 # define strtok_r(s,d,p) \
 394     (GL_LINK_WARNING ("strtok_r is unportable - " \
 395                       "use gnulib module strtok_r for portability"), \
 396      strtok_r (s, d, p))
 397 #endif
 398
 399
 400 /* The following functions are not specified by POSIX.  They are gnulib
 401    extensions.  */
 402
 403 #if @GNULIB_MBSLEN@
 404 /* Return the number of multibyte characters in the character string STRING.
 405    This considers multibyte characters, unlike strlen, which counts bytes.  */
 406 extern size_t mbslen (const char *string);
 407 #endif
 408
 409 #if @GNULIB_MBSCHR@
 410 /* Locate the first single-byte character C in the character string STRING,
 411    and return a pointer to it.  Return NULL if C is not found in STRING.
 412    Unlike strchr(), this function works correctly in multibyte locales with
 413    encodings such as GB18030.  */
 414 # define mbschr rpl_mbschr /* avoid collision with HP-UX function */
 415 extern char * mbschr (const char *string, int c);
 416 #endif
 417
 418 #if @GNULIB_MBSRCHR@
 419 /* Locate the last single-byte character C in the character string STRING,
 420    and return a pointer to it.  Return NULL if C is not found in STRING.
 421    Unlike strrchr(), this function works correctly in multibyte locales with
 422    encodings such as GB18030.  */
 423 # define mbsrchr rpl_mbsrchr /* avoid collision with HP-UX function */
 424 extern char * mbsrchr (const char *string, int c);
 425 #endif
 426
 427 #if @GNULIB_MBSSTR@
 428 /* Find the first occurrence of the character string NEEDLE in the character
 429    string HAYSTACK.  Return NULL if NEEDLE is not found in HAYSTACK.
 430    Unlike strstr(), this function works correctly in multibyte locales with
 431    encodings different from UTF-8.  */
 432 extern char * mbsstr (const char *haystack, const char *needle);
 433 #endif
 434
 435 #if @GNULIB_MBSCASECMP@
 436 /* Compare the character strings S1 and S2, ignoring case, returning less than,
 437    equal to or greater than zero if S1 is lexicographically less than, equal to
 438    or greater than S2.
 439    Note: This function may, in multibyte locales, return 0 for strings of
 440    different lengths!
 441    Unlike strcasecmp(), this function works correctly in multibyte locales.  */
 442 extern int mbscasecmp (const char *s1, const char *s2);
 443 #endif
 444
 445 #if @GNULIB_MBSNCASECMP@
 446 /* Compare the initial segment of the character string S1 consisting of at most
 447    N characters with the initial segment of the character string S2 consisting
 448    of at most N characters, ignoring case, returning less than, equal to or
 449    greater than zero if the initial segment of S1 is lexicographically less
 450    than, equal to or greater than the initial segment of S2.
 451    Note: This function may, in multibyte locales, return 0 for initial segments
 452    of different lengths!
 453    Unlike strncasecmp(), this function works correctly in multibyte locales.
 454    But beware that N is not a byte count but a character count!  */
 455 extern int mbsncasecmp (const char *s1, const char *s2, size_t n);
 456 #endif
 457
 458 #if @GNULIB_MBSPCASECMP@
 459 /* Compare the initial segment of the character string STRING consisting of
 460    at most mbslen (PREFIX) characters with the character string PREFIX,
 461    ignoring case, returning less than, equal to or greater than zero if this
 462    initial segment is lexicographically less than, equal to or greater than
 463    PREFIX.
 464    Note: This function may, in multibyte locales, return 0 if STRING is of
 465    smaller length than PREFIX!
 466    Unlike strncasecmp(), this function works correctly in multibyte
 467    locales.  */
 468 extern char * mbspcasecmp (const char *string, const char *prefix);
 469 #endif
 470
 471 #if @GNULIB_MBSCASESTR@
 472 /* Find the first occurrence of the character string NEEDLE in the character
 473    string HAYSTACK, using case-insensitive comparison.
 474    Note: This function may, in multibyte locales, return success even if
 475    strlen (haystack) < strlen (needle) !
 476    Unlike strcasestr(), this function works correctly in multibyte locales.  */
 477 extern char * mbscasestr (const char *haystack, const char *needle);
 478 #endif
 479
 480 #if @GNULIB_MBSCSPN@
 481 /* Find the first occurrence in the character string STRING of any character
 482    in the character string ACCEPT.  Return the number of bytes from the
 483    beginning of the string to this occurrence, or to the end of the string
 484    if none exists.
 485    Unlike strcspn(), this function works correctly in multibyte locales.  */
 486 extern size_t mbscspn (const char *string, const char *accept);
 487 #endif
 488
 489 #if @GNULIB_MBSPBRK@
 490 /* Find the first occurrence in the character string STRING of any character
 491    in the character string ACCEPT.  Return the pointer to it, or NULL if none
 492    exists.
 493    Unlike strpbrk(), this function works correctly in multibyte locales.  */
 494 # define mbspbrk rpl_mbspbrk /* avoid collision with HP-UX function */
 495 extern char * mbspbrk (const char *string, const char *accept);
 496 #endif
 497
 498 #if @GNULIB_MBSSPN@
 499 /* Find the first occurrence in the character string STRING of any character
 500    not in the character string REJECT.  Return the number of bytes from the
 501    beginning of the string to this occurrence, or to the end of the string
 502    if none exists.
 503    Unlike strspn(), this function works correctly in multibyte locales.  */
 504 extern size_t mbsspn (const char *string, const char *reject);
 505 #endif
 506
 507 #if @GNULIB_MBSSEP@
 508 /* Search the next delimiter (multibyte character listed in the character
 509    string DELIM) starting at the character string *STRINGP.
 510    If one is found, overwrite it with a NUL, and advance *STRINGP to point
 511    to the next multibyte character after it.  Otherwise, set *STRINGP to NULL.
 512    If *STRINGP was already NULL, nothing happens.
 513    Return the old value of *STRINGP.
 514
 515    This is a variant of mbstok_r() that supports empty fields.
 516
 517    Caveat: It modifies the original string.
 518    Caveat: These functions cannot be used on constant strings.
 519    Caveat: The identity of the delimiting character is lost.
 520
 521    See also mbstok_r().  */
 522 extern char * mbssep (char **stringp, const char *delim);
 523 #endif
 524
 525 #if @GNULIB_MBSTOK_R@
 526 /* Parse the character string STRING into tokens separated by characters in
 527    the character string DELIM.
 528    If STRING is NULL, the saved pointer in SAVE_PTR is used as
 529    the next starting point.  For example:
 530         char s[] = "-abc-=-def";
 531         char *sp;
 532         x = mbstok_r(s, "-", &sp);      // x = "abc", sp = "=-def"
 533         x = mbstok_r(NULL, "-=", &sp);  // x = "def", sp = NULL
 534         x = mbstok_r(NULL, "=", &sp);   // x = NULL
 535                 // s = "abc\0-def\0"
 536
 537    Caveat: It modifies the original string.
 538    Caveat: These functions cannot be used on constant strings.
 539    Caveat: The identity of the delimiting character is lost.
 540
 541    See also mbssep().  */
 542 extern char * mbstok_r (char *string, const char *delim, char **save_ptr);
 543 #endif
 544
 545
 546 #ifdef __cplusplus
 547 }
 548 #endif
 549
 550 #endif