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