3 * Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
4 * 2000, 2001, 2002, 2003, 2005, by Larry Wall and others
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
12 * A Elbereth Gilthoniel,
13 * silivren penna míriel
14 * o menel aglar elenath!
15 * Na-chaered palan-díriel
16 * o galadhremmin ennorath,
17 * Fanuilos, le linnathon
18 * nef aear, si nef aearon!
21 /* utility functions for handling locale-specific stuff like what
22 * character represents the decimal point.
26 #define PERL_IN_LOCALE_C
34 # include <langinfo.h>
39 #if defined(USE_LOCALE_NUMERIC) || defined(USE_LOCALE_COLLATE)
41 * Standardize the locale name from a string returned by 'setlocale'.
43 * The standard return value of setlocale() is either
44 * (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
45 * (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
46 * (the space-separated values represent the various sublocales,
47 * in some unspecificed order)
49 * In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
50 * which is harmful for further use of the string in setlocale().
54 S_stdize_locale(pTHX_ char *locs)
56 const char *s = strchr(locs, '=');
60 const char * const t = strchr(s, '.');
63 const char * const u = strchr(t, '\n');
64 if (u && (u[1] == 0)) {
65 const STRLEN len = u - s;
66 Move(s + 1, locs, len, char);
74 Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);
81 Perl_set_numeric_radix(pTHX)
83 #ifdef USE_LOCALE_NUMERIC
84 # ifdef HAS_LOCALECONV
88 if (lc && lc->decimal_point) {
89 if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
90 SvREFCNT_dec(PL_numeric_radix_sv);
91 PL_numeric_radix_sv = Nullsv;
94 if (PL_numeric_radix_sv)
95 sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
97 PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
101 PL_numeric_radix_sv = Nullsv;
102 # endif /* HAS_LOCALECONV */
103 #endif /* USE_LOCALE_NUMERIC */
107 * Set up for a new numeric locale.
110 Perl_new_numeric(pTHX_ const char *newnum)
112 #ifdef USE_LOCALE_NUMERIC
115 Safefree(PL_numeric_name);
116 PL_numeric_name = NULL;
117 PL_numeric_standard = TRUE;
118 PL_numeric_local = TRUE;
122 if (! PL_numeric_name || strNE(PL_numeric_name, newnum)) {
123 Safefree(PL_numeric_name);
124 PL_numeric_name = stdize_locale(savepv(newnum));
125 PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
126 || strEQ(newnum, "POSIX"));
127 PL_numeric_local = TRUE;
131 #endif /* USE_LOCALE_NUMERIC */
135 Perl_set_numeric_standard(pTHX)
137 #ifdef USE_LOCALE_NUMERIC
139 if (! PL_numeric_standard) {
140 setlocale(LC_NUMERIC, "C");
141 PL_numeric_standard = TRUE;
142 PL_numeric_local = FALSE;
146 #endif /* USE_LOCALE_NUMERIC */
150 Perl_set_numeric_local(pTHX)
152 #ifdef USE_LOCALE_NUMERIC
154 if (! PL_numeric_local) {
155 setlocale(LC_NUMERIC, PL_numeric_name);
156 PL_numeric_standard = FALSE;
157 PL_numeric_local = TRUE;
161 #endif /* USE_LOCALE_NUMERIC */
165 * Set up for a new ctype locale.
168 Perl_new_ctype(pTHX_ const char *newctype)
170 #ifdef USE_LOCALE_CTYPE
174 for (i = 0; i < 256; i++) {
176 PL_fold_locale[i] = toLOWER_LC(i);
177 else if (isLOWER_LC(i))
178 PL_fold_locale[i] = toUPPER_LC(i);
180 PL_fold_locale[i] = i;
183 #endif /* USE_LOCALE_CTYPE */
184 PERL_UNUSED_ARG(newctype);
188 * Set up for a new collation locale.
191 Perl_new_collate(pTHX_ const char *newcoll)
193 #ifdef USE_LOCALE_COLLATE
196 if (PL_collation_name) {
198 Safefree(PL_collation_name);
199 PL_collation_name = NULL;
201 PL_collation_standard = TRUE;
202 PL_collxfrm_base = 0;
203 PL_collxfrm_mult = 2;
207 if (! PL_collation_name || strNE(PL_collation_name, newcoll)) {
209 Safefree(PL_collation_name);
210 PL_collation_name = stdize_locale(savepv(newcoll));
211 PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
212 || strEQ(newcoll, "POSIX"));
215 /* 2: at most so many chars ('a', 'b'). */
216 /* 50: surely no system expands a char more. */
217 #define XFRMBUFSIZE (2 * 50)
218 char xbuf[XFRMBUFSIZE];
219 const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
220 const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
221 const SSize_t mult = fb - fa;
223 Perl_croak(aTHX_ "strxfrm() gets absurd");
224 PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
225 PL_collxfrm_mult = mult;
229 #endif /* USE_LOCALE_COLLATE */
233 * Initialize locale awareness.
236 Perl_init_i18nl10n(pTHX_ int printwarn)
240 * 1 = set ok or not applicable,
241 * 0 = fallback to C locale,
242 * -1 = fallback to C locale failed
245 #if defined(USE_LOCALE)
247 #ifdef USE_LOCALE_CTYPE
248 char *curctype = NULL;
249 #endif /* USE_LOCALE_CTYPE */
250 #ifdef USE_LOCALE_COLLATE
251 char *curcoll = NULL;
252 #endif /* USE_LOCALE_COLLATE */
253 #ifdef USE_LOCALE_NUMERIC
255 #endif /* USE_LOCALE_NUMERIC */
257 char *language = PerlEnv_getenv("LANGUAGE");
259 char *lc_all = PerlEnv_getenv("LC_ALL");
260 char *lang = PerlEnv_getenv("LANG");
261 bool setlocale_failure = FALSE;
263 #ifdef LOCALE_ENVIRON_REQUIRED
266 * Ultrix setlocale(..., "") fails if there are no environment
267 * variables from which to get a locale name.
274 if (setlocale(LC_ALL, ""))
277 setlocale_failure = TRUE;
279 if (!setlocale_failure) {
280 #ifdef USE_LOCALE_CTYPE
283 (!done && (lang || PerlEnv_getenv("LC_CTYPE")))
285 setlocale_failure = TRUE;
287 curctype = savepv(curctype);
288 #endif /* USE_LOCALE_CTYPE */
289 #ifdef USE_LOCALE_COLLATE
291 setlocale(LC_COLLATE,
292 (!done && (lang || PerlEnv_getenv("LC_COLLATE")))
294 setlocale_failure = TRUE;
296 curcoll = savepv(curcoll);
297 #endif /* USE_LOCALE_COLLATE */
298 #ifdef USE_LOCALE_NUMERIC
300 setlocale(LC_NUMERIC,
301 (!done && (lang || PerlEnv_getenv("LC_NUMERIC")))
303 setlocale_failure = TRUE;
305 curnum = savepv(curnum);
306 #endif /* USE_LOCALE_NUMERIC */
311 #endif /* !LOCALE_ENVIRON_REQUIRED */
314 if (! setlocale(LC_ALL, ""))
315 setlocale_failure = TRUE;
318 if (!setlocale_failure) {
319 #ifdef USE_LOCALE_CTYPE
320 if (! (curctype = setlocale(LC_CTYPE, "")))
321 setlocale_failure = TRUE;
323 curctype = savepv(curctype);
324 #endif /* USE_LOCALE_CTYPE */
325 #ifdef USE_LOCALE_COLLATE
326 if (! (curcoll = setlocale(LC_COLLATE, "")))
327 setlocale_failure = TRUE;
329 curcoll = savepv(curcoll);
330 #endif /* USE_LOCALE_COLLATE */
331 #ifdef USE_LOCALE_NUMERIC
332 if (! (curnum = setlocale(LC_NUMERIC, "")))
333 setlocale_failure = TRUE;
335 curnum = savepv(curnum);
336 #endif /* USE_LOCALE_NUMERIC */
339 if (setlocale_failure) {
341 bool locwarn = (printwarn > 1 ||
343 (!(p = PerlEnv_getenv("PERL_BADLANG")) || atoi(p))));
348 PerlIO_printf(Perl_error_log,
349 "perl: warning: Setting locale failed.\n");
353 PerlIO_printf(Perl_error_log,
354 "perl: warning: Setting locale failed for the categories:\n\t");
355 #ifdef USE_LOCALE_CTYPE
357 PerlIO_printf(Perl_error_log, "LC_CTYPE ");
358 #endif /* USE_LOCALE_CTYPE */
359 #ifdef USE_LOCALE_COLLATE
361 PerlIO_printf(Perl_error_log, "LC_COLLATE ");
362 #endif /* USE_LOCALE_COLLATE */
363 #ifdef USE_LOCALE_NUMERIC
365 PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
366 #endif /* USE_LOCALE_NUMERIC */
367 PerlIO_printf(Perl_error_log, "\n");
371 PerlIO_printf(Perl_error_log,
372 "perl: warning: Please check that your locale settings:\n");
375 PerlIO_printf(Perl_error_log,
376 "\tLANGUAGE = %c%s%c,\n",
377 language ? '"' : '(',
378 language ? language : "unset",
379 language ? '"' : ')');
382 PerlIO_printf(Perl_error_log,
383 "\tLC_ALL = %c%s%c,\n",
385 lc_all ? lc_all : "unset",
388 #if defined(USE_ENVIRON_ARRAY)
391 for (e = environ; *e; e++) {
392 if (strnEQ(*e, "LC_", 3)
393 && strnNE(*e, "LC_ALL=", 7)
394 && (p = strchr(*e, '=')))
395 PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
396 (int)(p - *e), *e, p + 1);
400 PerlIO_printf(Perl_error_log,
401 "\t(possibly more locale environment variables)\n");
404 PerlIO_printf(Perl_error_log,
407 lang ? lang : "unset",
410 PerlIO_printf(Perl_error_log,
411 " are supported and installed on your system.\n");
416 if (setlocale(LC_ALL, "C")) {
418 PerlIO_printf(Perl_error_log,
419 "perl: warning: Falling back to the standard locale (\"C\").\n");
424 PerlIO_printf(Perl_error_log,
425 "perl: warning: Failed to fall back to the standard locale (\"C\").\n");
432 #ifdef USE_LOCALE_CTYPE
433 || !(curctype || setlocale(LC_CTYPE, "C"))
434 #endif /* USE_LOCALE_CTYPE */
435 #ifdef USE_LOCALE_COLLATE
436 || !(curcoll || setlocale(LC_COLLATE, "C"))
437 #endif /* USE_LOCALE_COLLATE */
438 #ifdef USE_LOCALE_NUMERIC
439 || !(curnum || setlocale(LC_NUMERIC, "C"))
440 #endif /* USE_LOCALE_NUMERIC */
444 PerlIO_printf(Perl_error_log,
445 "perl: warning: Cannot fall back to the standard locale (\"C\").\n");
449 #endif /* ! LC_ALL */
451 #ifdef USE_LOCALE_CTYPE
452 curctype = savepv(setlocale(LC_CTYPE, Nullch));
453 #endif /* USE_LOCALE_CTYPE */
454 #ifdef USE_LOCALE_COLLATE
455 curcoll = savepv(setlocale(LC_COLLATE, Nullch));
456 #endif /* USE_LOCALE_COLLATE */
457 #ifdef USE_LOCALE_NUMERIC
458 curnum = savepv(setlocale(LC_NUMERIC, Nullch));
459 #endif /* USE_LOCALE_NUMERIC */
463 #ifdef USE_LOCALE_CTYPE
465 #endif /* USE_LOCALE_CTYPE */
467 #ifdef USE_LOCALE_COLLATE
468 new_collate(curcoll);
469 #endif /* USE_LOCALE_COLLATE */
471 #ifdef USE_LOCALE_NUMERIC
473 #endif /* USE_LOCALE_NUMERIC */
477 #endif /* USE_LOCALE */
481 /* Set PL_utf8locale to TRUE if using PerlIO _and_
482 any of the following are true:
483 - nl_langinfo(CODESET) contains /^utf-?8/i
484 - $ENV{LC_ALL} contains /^utf-?8/i
485 - $ENV{LC_CTYPE} contains /^utf-?8/i
486 - $ENV{LANG} contains /^utf-?8/i
487 The LC_ALL, LC_CTYPE, LANG obey the usual override
488 hierarchy of locale environment variables. (LANGUAGE
489 affects only LC_MESSAGES only under glibc.) (If present,
490 it overrides LC_MESSAGES for GNU gettext, and it also
491 can have more than one locale, separated by spaces,
492 in case you need to know.)
493 If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
494 are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
495 on STDIN, STDOUT, STDERR, _and_ the default open discipline.
497 bool utf8locale = FALSE;
498 char *codeset = NULL;
499 #if defined(HAS_NL_LANGINFO) && defined(CODESET)
500 codeset = nl_langinfo(CODESET);
503 utf8locale = (ibcmp(codeset, "UTF-8", 5) == 0 ||
504 ibcmp(codeset, "UTF8", 4) == 0);
505 #if defined(USE_LOCALE)
506 else { /* nl_langinfo(CODESET) is supposed to correctly
507 * interpret the locale environment variables,
508 * but just in case it fails, let's do this manually. */
510 utf8locale = (ibcmp(lang, "UTF-8", 5) == 0 ||
511 ibcmp(lang, "UTF8", 4) == 0);
512 #ifdef USE_LOCALE_CTYPE
514 utf8locale = (ibcmp(curctype, "UTF-8", 5) == 0 ||
515 ibcmp(curctype, "UTF8", 4) == 0);
518 utf8locale = (ibcmp(lc_all, "UTF-8", 5) == 0 ||
519 ibcmp(lc_all, "UTF8", 4) == 0);
521 #endif /* USE_LOCALE */
523 PL_utf8locale = TRUE;
525 /* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
526 This is an alternative to using the -C command line switch
527 (the -C if present will override this). */
529 const char *p = PerlEnv_getenv("PERL_UNICODE");
530 PL_unicode = p ? parse_unicode_opts(&p) : 0;
534 #ifdef USE_LOCALE_CTYPE
536 #endif /* USE_LOCALE_CTYPE */
537 #ifdef USE_LOCALE_COLLATE
539 #endif /* USE_LOCALE_COLLATE */
540 #ifdef USE_LOCALE_NUMERIC
542 #endif /* USE_LOCALE_NUMERIC */
546 /* Backwards compatibility. */
548 Perl_init_i18nl14n(pTHX_ int printwarn)
550 return init_i18nl10n(printwarn);
553 #ifdef USE_LOCALE_COLLATE
556 * mem_collxfrm() is a bit like strxfrm() but with two important
557 * differences. First, it handles embedded NULs. Second, it allocates
558 * a bit more memory than needed for the transformed data itself.
559 * The real transformed data begins at offset sizeof(collationix).
560 * Please see sv_collxfrm() to see how this is used.
564 Perl_mem_collxfrm(pTHX_ const char *s, STRLEN len, STRLEN *xlen)
567 STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
569 /* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
570 /* the +1 is for the terminating NUL. */
572 xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
573 Newx(xbuf, xAlloc, char);
577 *(U32*)xbuf = PL_collation_ix;
578 xout = sizeof(PL_collation_ix);
579 for (xin = 0; xin < len; ) {
583 xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
586 if ((STRLEN)xused < xAlloc - xout)
588 xAlloc = (2 * xAlloc) + 1;
589 Renew(xbuf, xAlloc, char);
594 xin += strlen(s + xin) + 1;
597 /* Embedded NULs are understood but silently skipped
598 * because they make no sense in locale collation. */
602 *xlen = xout - sizeof(PL_collation_ix);
611 #endif /* USE_LOCALE_COLLATE */
615 * c-indentation-style: bsd
617 * indent-tabs-mode: t
620 * ex: set ts=8 sts=4 sw=4 noet: