[perl5.git] / locale.c

/*    locale.c
 *
 *    Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
 *    2000, 2001, 2002, 2003, 2005, by Larry Wall and others
 *
 *    You may distribute under the terms of either the GNU General Public
 *    License or the Artistic License, as specified in the README file.
 *
 */

/*
 * A Elbereth Gilthoniel,
 * silivren penna míriel
 * o menel aglar elenath!
 * Na-chaered palan-díriel
 * o galadhremmin ennorath,
 * Fanuilos, le linnathon
 * nef aear, si nef aearon!
 */

/* utility functions for handling locale-specific stuff like what
 * character represents the decimal point.
 */

#include "EXTERN.h"
#define PERL_IN_LOCALE_C
#include "perl.h"

#ifdef I_LOCALE
#  include <locale.h>
#endif

#ifdef I_LANGINFO
#   include <langinfo.h>
#endif

#include "reentr.h"

#if defined(USE_LOCALE_NUMERIC) || defined(USE_LOCALE_COLLATE)
/*
 * Standardize the locale name from a string returned by 'setlocale'.
 *
 * The standard return value of setlocale() is either
 * (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
 * (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
 *     (the space-separated values represent the various sublocales,
 *      in some unspecificed order)
 *
 * In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
 * which is harmful for further use of the string in setlocale().
 *
 */
STATIC char *
S_stdize_locale(pTHX_ char *locs)
{
    const char * const s = strchr(locs, '=');
    bool okay = TRUE;

    if (s) {
	const char * const t = strchr(s, '.');
	okay = FALSE;
	if (t) {
	    const char * const u = strchr(t, '\n');
	    if (u && (u[1] == 0)) {
		const STRLEN len = u - s;
		Move(s + 1, locs, len, char);
		locs[len] = 0;
		okay = TRUE;
	    }
	}
    }

    if (!okay)
	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);

    return locs;
}
#endif

void
Perl_set_numeric_radix(pTHX)
{
#ifdef USE_LOCALE_NUMERIC
# ifdef HAS_LOCALECONV
    const struct lconv* const lc = localeconv();

    if (lc && lc->decimal_point) {
	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	    SvREFCNT_dec(PL_numeric_radix_sv);
	    PL_numeric_radix_sv = NULL;
	}
	else {
	    if (PL_numeric_radix_sv)
		sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	    else
		PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	}
    }
    else
	PL_numeric_radix_sv = NULL;
# endif /* HAS_LOCALECONV */
#endif /* USE_LOCALE_NUMERIC */
}

/*
 * Set up for a new numeric locale.
 */
void
Perl_new_numeric(pTHX_ char *newnum)
{
#ifdef USE_LOCALE_NUMERIC

    if (! newnum) {
	Safefree(PL_numeric_name);
	PL_numeric_name = NULL;
	PL_numeric_standard = TRUE;
	PL_numeric_local = TRUE;
	return;
    }

    if (! PL_numeric_name || strNE(PL_numeric_name, newnum)) {
	Safefree(PL_numeric_name);
	PL_numeric_name = stdize_locale(savepv(newnum));
	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
			       || strEQ(newnum, "POSIX"));
	PL_numeric_local = TRUE;
	set_numeric_radix();
    }

#endif /* USE_LOCALE_NUMERIC */
}

void
Perl_set_numeric_standard(pTHX)
{
#ifdef USE_LOCALE_NUMERIC

    if (! PL_numeric_standard) {
	setlocale(LC_NUMERIC, "C");
	PL_numeric_standard = TRUE;
	PL_numeric_local = FALSE;
	set_numeric_radix();
    }

#endif /* USE_LOCALE_NUMERIC */
}

void
Perl_set_numeric_local(pTHX)
{
#ifdef USE_LOCALE_NUMERIC

    if (! PL_numeric_local) {
	setlocale(LC_NUMERIC, PL_numeric_name);
	PL_numeric_standard = FALSE;
	PL_numeric_local = TRUE;
	set_numeric_radix();
    }

#endif /* USE_LOCALE_NUMERIC */
}

/*
 * Set up for a new ctype locale.
 */
void
Perl_new_ctype(pTHX_ char *newctype)
{
#ifdef USE_LOCALE_CTYPE
    int i;

    for (i = 0; i < 256; i++) {
	if (isUPPER_LC(i))
	    PL_fold_locale[i] = toLOWER_LC(i);
	else if (isLOWER_LC(i))
	    PL_fold_locale[i] = toUPPER_LC(i);
	else
	    PL_fold_locale[i] = i;
    }

#endif /* USE_LOCALE_CTYPE */
    PERL_UNUSED_ARG(newctype);
}

/*
 * Set up for a new collation locale.
 */
void
Perl_new_collate(pTHX_ char *newcoll)
{
#ifdef USE_LOCALE_COLLATE

    if (! newcoll) {
	if (PL_collation_name) {
	    ++PL_collation_ix;
	    Safefree(PL_collation_name);
	    PL_collation_name = NULL;
	}
	PL_collation_standard = TRUE;
	PL_collxfrm_base = 0;
	PL_collxfrm_mult = 2;
	return;
    }

    if (! PL_collation_name || strNE(PL_collation_name, newcoll)) {
	++PL_collation_ix;
	Safefree(PL_collation_name);
	PL_collation_name = stdize_locale(savepv(newcoll));
	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
				 || strEQ(newcoll, "POSIX"));

	{
	  /*  2: at most so many chars ('a', 'b'). */
	  /* 50: surely no system expands a char more. */
#define XFRMBUFSIZE  (2 * 50)
	  char xbuf[XFRMBUFSIZE];
	  const Size_t fa = strxfrm(xbuf, "a",  XFRMBUFSIZE);
	  const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	  const SSize_t mult = fb - fa;
	  if (mult < 1)
	      Perl_croak(aTHX_ "strxfrm() gets absurd");
	  PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
	  PL_collxfrm_mult = mult;
	}
    }

#endif /* USE_LOCALE_COLLATE */
}

/*
 * Initialize locale awareness.
 */
int
Perl_init_i18nl10n(pTHX_ int printwarn)
{
    int ok = 1;
    /* returns
     *    1 = set ok or not applicable,
     *    0 = fallback to C locale,
     *   -1 = fallback to C locale failed
     */

#if defined(USE_LOCALE)

#ifdef USE_LOCALE_CTYPE
    char *curctype   = NULL;
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
    char *curcoll    = NULL;
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
    char *curnum     = NULL;
#endif /* USE_LOCALE_NUMERIC */
#ifdef __GLIBC__
    char * const language   = PerlEnv_getenv("LANGUAGE");
#endif
    char * const lc_all     = PerlEnv_getenv("LC_ALL");
    char * const lang       = PerlEnv_getenv("LANG");
    bool setlocale_failure = FALSE;

#ifdef LOCALE_ENVIRON_REQUIRED

    /*
     * Ultrix setlocale(..., "") fails if there are no environment
     * variables from which to get a locale name.
     */

    bool done = FALSE;

#ifdef LC_ALL
    if (lang) {
	if (setlocale(LC_ALL, ""))
	    done = TRUE;
	else
	    setlocale_failure = TRUE;
    }
    if (!setlocale_failure) {
#ifdef USE_LOCALE_CTYPE
	if (! (curctype =
	       setlocale(LC_CTYPE,
			 (!done && (lang || PerlEnv_getenv("LC_CTYPE")))
				    ? "" : NULL)))
	    setlocale_failure = TRUE;
	else
	    curctype = savepv(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	if (! (curcoll =
	       setlocale(LC_COLLATE,
			 (!done && (lang || PerlEnv_getenv("LC_COLLATE")))
				   ? "" : NULL)))
	    setlocale_failure = TRUE;
	else
	    curcoll = savepv(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	if (! (curnum =
	       setlocale(LC_NUMERIC,
			 (!done && (lang || PerlEnv_getenv("LC_NUMERIC")))
				  ? "" : NULL)))
	    setlocale_failure = TRUE;
	else
	    curnum = savepv(curnum);
#endif /* USE_LOCALE_NUMERIC */
    }

#endif /* LC_ALL */

#endif /* !LOCALE_ENVIRON_REQUIRED */

#ifdef LC_ALL
    if (! setlocale(LC_ALL, ""))
	setlocale_failure = TRUE;
#endif /* LC_ALL */

    if (!setlocale_failure) {
#ifdef USE_LOCALE_CTYPE
	if (! (curctype = setlocale(LC_CTYPE, "")))
	    setlocale_failure = TRUE;
	else
	    curctype = savepv(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	if (! (curcoll = setlocale(LC_COLLATE, "")))
	    setlocale_failure = TRUE;
	else
	    curcoll = savepv(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	if (! (curnum = setlocale(LC_NUMERIC, "")))
	    setlocale_failure = TRUE;
	else
	    curnum = savepv(curnum);
#endif /* USE_LOCALE_NUMERIC */
    }

    if (setlocale_failure) {
	char *p;
	const bool locwarn = (printwarn > 1 ||
			(printwarn &&
			 (!(p = PerlEnv_getenv("PERL_BADLANG")) || atoi(p))));

	if (locwarn) {
#ifdef LC_ALL

	    PerlIO_printf(Perl_error_log,
	       "perl: warning: Setting locale failed.\n");

#else /* !LC_ALL */

	    PerlIO_printf(Perl_error_log,
	       "perl: warning: Setting locale failed for the categories:\n\t");
#ifdef USE_LOCALE_CTYPE
	    if (! curctype)
		PerlIO_printf(Perl_error_log, "LC_CTYPE ");
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	    if (! curcoll)
		PerlIO_printf(Perl_error_log, "LC_COLLATE ");
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	    if (! curnum)
		PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
#endif /* USE_LOCALE_NUMERIC */
	    PerlIO_printf(Perl_error_log, "\n");

#endif /* LC_ALL */

	    PerlIO_printf(Perl_error_log,
		"perl: warning: Please check that your locale settings:\n");

#ifdef __GLIBC__
	    PerlIO_printf(Perl_error_log,
			  "\tLANGUAGE = %c%s%c,\n",
			  language ? '"' : '(',
			  language ? language : "unset",
			  language ? '"' : ')');
#endif

	    PerlIO_printf(Perl_error_log,
			  "\tLC_ALL = %c%s%c,\n",
			  lc_all ? '"' : '(',
			  lc_all ? lc_all : "unset",
			  lc_all ? '"' : ')');

#if defined(USE_ENVIRON_ARRAY)
	    {
	      char **e;
	      for (e = environ; *e; e++) {
		  if (strnEQ(*e, "LC_", 3)
			&& strnNE(*e, "LC_ALL=", 7)
			&& (p = strchr(*e, '=')))
		      PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
				    (int)(p - *e), *e, p + 1);
	      }
	    }
#else
	    PerlIO_printf(Perl_error_log,
			  "\t(possibly more locale environment variables)\n");
#endif

	    PerlIO_printf(Perl_error_log,
			  "\tLANG = %c%s%c\n",
			  lang ? '"' : '(',
			  lang ? lang : "unset",
			  lang ? '"' : ')');

	    PerlIO_printf(Perl_error_log,
			  "    are supported and installed on your system.\n");
	}

#ifdef LC_ALL

	if (setlocale(LC_ALL, "C")) {
	    if (locwarn)
		PerlIO_printf(Perl_error_log,
      "perl: warning: Falling back to the standard locale (\"C\").\n");
	    ok = 0;
	}
	else {
	    if (locwarn)
		PerlIO_printf(Perl_error_log,
      "perl: warning: Failed to fall back to the standard locale (\"C\").\n");
	    ok = -1;
	}

#else /* ! LC_ALL */

	if (0
#ifdef USE_LOCALE_CTYPE
	    || !(curctype || setlocale(LC_CTYPE, "C"))
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	    || !(curcoll || setlocale(LC_COLLATE, "C"))
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	    || !(curnum || setlocale(LC_NUMERIC, "C"))
#endif /* USE_LOCALE_NUMERIC */
	    )
	{
	    if (locwarn)
		PerlIO_printf(Perl_error_log,
      "perl: warning: Cannot fall back to the standard locale (\"C\").\n");
	    ok = -1;
	}

#endif /* ! LC_ALL */

#ifdef USE_LOCALE_CTYPE
	curctype = savepv(setlocale(LC_CTYPE, NULL));
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	curcoll = savepv(setlocale(LC_COLLATE, NULL));
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	curnum = savepv(setlocale(LC_NUMERIC, NULL));
#endif /* USE_LOCALE_NUMERIC */
    }
    else {

#ifdef USE_LOCALE_CTYPE
    new_ctype(curctype);
#endif /* USE_LOCALE_CTYPE */

#ifdef USE_LOCALE_COLLATE
    new_collate(curcoll);
#endif /* USE_LOCALE_COLLATE */

#ifdef USE_LOCALE_NUMERIC
    new_numeric(curnum);
#endif /* USE_LOCALE_NUMERIC */

    }

#endif /* USE_LOCALE */

#ifdef USE_PERLIO
    {
      /* Set PL_utf8locale to TRUE if using PerlIO _and_
	 any of the following are true:
	 - nl_langinfo(CODESET) contains /^utf-?8/i
	 - $ENV{LC_ALL}   contains /^utf-?8/i
	 - $ENV{LC_CTYPE} contains /^utf-?8/i
	 - $ENV{LANG}     contains /^utf-?8/i
	 The LC_ALL, LC_CTYPE, LANG obey the usual override
	 hierarchy of locale environment variables.  (LANGUAGE
	 affects only LC_MESSAGES only under glibc.) (If present,
	 it overrides LC_MESSAGES for GNU gettext, and it also
	 can have more than one locale, separated by spaces,
	 in case you need to know.)
	 If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
         are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
	 on STDIN, STDOUT, STDERR, _and_ the default open discipline.
      */
	 bool utf8locale = FALSE;
	 char *codeset = NULL;
#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	 codeset = nl_langinfo(CODESET);
#endif
	 if (codeset)
	      utf8locale = (ibcmp(codeset,  "UTF-8", 5) == 0 ||
 			    ibcmp(codeset,  "UTF8",  4) == 0);
#if defined(USE_LOCALE)
	 else { /* nl_langinfo(CODESET) is supposed to correctly
		 * interpret the locale environment variables,
		 * but just in case it fails, let's do this manually. */ 
	      if (lang)
		   utf8locale = (ibcmp(lang,     "UTF-8", 5) == 0 ||
			         ibcmp(lang,     "UTF8",  4) == 0);
#ifdef USE_LOCALE_CTYPE
	      if (curctype)
		   utf8locale = (ibcmp(curctype,     "UTF-8", 5) == 0 ||
			         ibcmp(curctype,     "UTF8",  4) == 0);
#endif
	      if (lc_all)
		   utf8locale = (ibcmp(lc_all,   "UTF-8", 5) == 0 ||
			         ibcmp(lc_all,   "UTF8",  4) == 0);
	 }
#endif /* USE_LOCALE */
	 if (utf8locale)
	      PL_utf8locale = TRUE;
    }
    /* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
       This is an alternative to using the -C command line switch
       (the -C if present will override this). */
    {
	 char *p = PerlEnv_getenv("PERL_UNICODE");
	 PL_unicode = p ? parse_unicode_opts(&p) : 0;
	 if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
	     PL_utf8cache = -1;
    }
#endif

#ifdef USE_LOCALE_CTYPE
    Safefree(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
    Safefree(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
    Safefree(curnum);
#endif /* USE_LOCALE_NUMERIC */
    return ok;
}

#ifdef USE_LOCALE_COLLATE

/*
 * mem_collxfrm() is a bit like strxfrm() but with two important
 * differences. First, it handles embedded NULs. Second, it allocates
 * a bit more memory than needed for the transformed data itself.
 * The real transformed data begins at offset sizeof(collationix).
 * Please see sv_collxfrm() to see how this is used.
 */

char *
Perl_mem_collxfrm(pTHX_ const char *s, STRLEN len, STRLEN *xlen)
{
    char *xbuf;
    STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */

    /* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
    /* the +1 is for the terminating NUL. */

    xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
    Newx(xbuf, xAlloc, char);
    if (! xbuf)
	goto bad;

    *(U32*)xbuf = PL_collation_ix;
    xout = sizeof(PL_collation_ix);
    for (xin = 0; xin < len; ) {
	SSize_t xused;

	for (;;) {
	    xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	    if (xused == -1)
		goto bad;
	    if ((STRLEN)xused < xAlloc - xout)
		break;
	    xAlloc = (2 * xAlloc) + 1;
	    Renew(xbuf, xAlloc, char);
	    if (! xbuf)
		goto bad;
	}

	xin += strlen(s + xin) + 1;
	xout += xused;

	/* Embedded NULs are understood but silently skipped
	 * because they make no sense in locale collation. */
    }

    xbuf[xout] = '\0';
    *xlen = xout - sizeof(PL_collation_ix);
    return xbuf;

  bad:
    Safefree(xbuf);
    *xlen = 0;
    return NULL;
}

#endif /* USE_LOCALE_COLLATE */

/*
 * Local variables:
 * c-indentation-style: bsd
 * c-basic-offset: 4
 * indent-tabs-mode: t
 * End:
 *
 * ex: set ts=8 sts=4 sw=4 noet:
 */
Commit	Line	Data
98994639 HS	1	/* locale.c
98994639 HS	2	*
e6906430	3	* Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
29652248	4	* 2000, 2001, 2002, 2003, 2005, by Larry Wall and others
98994639 HS	5	*
	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.
	8	*
	9	*/
	10
	11	/*
	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!
	19	*/
	20
40d34c0d SB	21	/* utility functions for handling locale-specific stuff like what
	22	* character represents the decimal point.
	23	*/
	24
98994639 HS	25	#include "EXTERN.h"
	26	#define PERL_IN_LOCALE_C
	27	#include "perl.h"
	28
	29	#ifdef I_LOCALE
	30	# include <locale.h>
	31	#endif
	32
b310b053 JH	33	#ifdef I_LANGINFO
	34	# include <langinfo.h>
	35	#endif
	36
a4af207c JH	37	#include "reentr.h"
a4af207c JH	38
fe20fd30	39	#if defined(USE_LOCALE_NUMERIC) \|\| defined(USE_LOCALE_COLLATE)
98994639 HS	40	/*
	41	* Standardize the locale name from a string returned by 'setlocale'.
	42	*
	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)
	48	*
	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().
	51	*
	52	*/
	53	STATIC char *
	54	S_stdize_locale(pTHX_ char *locs)
	55	{
6d29369a	56	const char * const s = strchr(locs, '=');
98994639 HS	57	bool okay = TRUE;
98994639 HS	58
4f1e9d25 AL	59	if (s) {
4f1e9d25 AL	60	const char * const t = strchr(s, '.');
98994639	61	okay = FALSE;
4f1e9d25 AL	62	if (t) {
	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);
	67	locs[len] = 0;
	68	okay = TRUE;
98994639 HS	69	}
	70	}
	71	}
	72
	73	if (!okay)
	74	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);
	75
	76	return locs;
	77	}
fe20fd30	78	#endif
98994639 HS	79
	80	void
	81	Perl_set_numeric_radix(pTHX)
	82	{
	83	#ifdef USE_LOCALE_NUMERIC
	84	# ifdef HAS_LOCALECONV
6d29369a	85	const struct lconv* const lc = localeconv();
98994639	86
98994639 HS	87	if (lc && lc->decimal_point) {
	88	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	89	SvREFCNT_dec(PL_numeric_radix_sv);
0e2d6244	90	PL_numeric_radix_sv = NULL;
98994639 HS	91	}
	92	else {
	93	if (PL_numeric_radix_sv)
	94	sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	95	else
	96	PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	97	}
	98	}
	99	else
0e2d6244	100	PL_numeric_radix_sv = NULL;
98994639 HS	101	# endif /* HAS_LOCALECONV */
	102	#endif /* USE_LOCALE_NUMERIC */
	103	}
	104
	105	/*
	106	* Set up for a new numeric locale.
	107	*/
	108	void
	109	Perl_new_numeric(pTHX_ char *newnum)
	110	{
	111	#ifdef USE_LOCALE_NUMERIC
	112
	113	if (! newnum) {
4c58c75a NC	114	Safefree(PL_numeric_name);
4c58c75a NC	115	PL_numeric_name = NULL;
98994639 HS	116	PL_numeric_standard = TRUE;
	117	PL_numeric_local = TRUE;
	118	return;
	119	}
	120
	121	if (! PL_numeric_name \|\| strNE(PL_numeric_name, newnum)) {
	122	Safefree(PL_numeric_name);
	123	PL_numeric_name = stdize_locale(savepv(newnum));
29652248 NC	124	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
29652248 NC	125	\|\| strEQ(newnum, "POSIX"));
98994639 HS	126	PL_numeric_local = TRUE;
	127	set_numeric_radix();
	128	}
	129
	130	#endif /* USE_LOCALE_NUMERIC */
	131	}
	132
	133	void
	134	Perl_set_numeric_standard(pTHX)
	135	{
	136	#ifdef USE_LOCALE_NUMERIC
	137
	138	if (! PL_numeric_standard) {
	139	setlocale(LC_NUMERIC, "C");
	140	PL_numeric_standard = TRUE;
	141	PL_numeric_local = FALSE;
	142	set_numeric_radix();
	143	}
	144
	145	#endif /* USE_LOCALE_NUMERIC */
	146	}
	147
	148	void
	149	Perl_set_numeric_local(pTHX)
	150	{
	151	#ifdef USE_LOCALE_NUMERIC
	152
	153	if (! PL_numeric_local) {
	154	setlocale(LC_NUMERIC, PL_numeric_name);
	155	PL_numeric_standard = FALSE;
	156	PL_numeric_local = TRUE;
	157	set_numeric_radix();
	158	}
	159
	160	#endif /* USE_LOCALE_NUMERIC */
	161	}
	162
	163	/*
	164	* Set up for a new ctype locale.
	165	*/
	166	void
	167	Perl_new_ctype(pTHX_ char *newctype)
	168	{
	169	#ifdef USE_LOCALE_CTYPE
98994639 HS	170	int i;
	171
	172	for (i = 0; i < 256; i++) {
	173	if (isUPPER_LC(i))
	174	PL_fold_locale[i] = toLOWER_LC(i);
	175	else if (isLOWER_LC(i))
	176	PL_fold_locale[i] = toUPPER_LC(i);
	177	else
	178	PL_fold_locale[i] = i;
	179	}
	180
	181	#endif /* USE_LOCALE_CTYPE */
4f1e9d25	182	PERL_UNUSED_ARG(newctype);
98994639 HS	183	}
	184
	185	/*
	186	* Set up for a new collation locale.
	187	*/
	188	void
	189	Perl_new_collate(pTHX_ char *newcoll)
	190	{
	191	#ifdef USE_LOCALE_COLLATE
	192
	193	if (! newcoll) {
	194	if (PL_collation_name) {
	195	++PL_collation_ix;
	196	Safefree(PL_collation_name);
	197	PL_collation_name = NULL;
	198	}
	199	PL_collation_standard = TRUE;
	200	PL_collxfrm_base = 0;
	201	PL_collxfrm_mult = 2;
	202	return;
	203	}
	204
	205	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
	206	++PL_collation_ix;
	207	Safefree(PL_collation_name);
	208	PL_collation_name = stdize_locale(savepv(newcoll));
29652248 NC	209	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
29652248 NC	210	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	211
	212	{
	213	/* 2: at most so many chars ('a', 'b'). */
	214	/* 50: surely no system expands a char more. */
	215	#define XFRMBUFSIZE (2 * 50)
	216	char xbuf[XFRMBUFSIZE];
4f1e9d25 AL	217	const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	218	const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	219	const SSize_t mult = fb - fa;
98994639 HS	220	if (mult < 1)
98994639 HS	221	Perl_croak(aTHX_ "strxfrm() gets absurd");
eb160463	222	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	223	PL_collxfrm_mult = mult;
	224	}
	225	}
	226
	227	#endif /* USE_LOCALE_COLLATE */
	228	}
	229
	230	/*
	231	* Initialize locale awareness.
	232	*/
	233	int
	234	Perl_init_i18nl10n(pTHX_ int printwarn)
	235	{
	236	int ok = 1;
	237	/* returns
	238	* 1 = set ok or not applicable,
	239	* 0 = fallback to C locale,
	240	* -1 = fallback to C locale failed
	241	*/
	242
	243	#if defined(USE_LOCALE)
	244
	245	#ifdef USE_LOCALE_CTYPE
	246	char *curctype = NULL;
	247	#endif /* USE_LOCALE_CTYPE */
	248	#ifdef USE_LOCALE_COLLATE
	249	char *curcoll = NULL;
	250	#endif /* USE_LOCALE_COLLATE */
	251	#ifdef USE_LOCALE_NUMERIC
	252	char *curnum = NULL;
	253	#endif /* USE_LOCALE_NUMERIC */
	254	#ifdef __GLIBC__
6d29369a	255	char * const language = PerlEnv_getenv("LANGUAGE");
98994639	256	#endif
6d29369a AL	257	char * const lc_all = PerlEnv_getenv("LC_ALL");
6d29369a AL	258	char * const lang = PerlEnv_getenv("LANG");
98994639 HS	259	bool setlocale_failure = FALSE;
	260
	261	#ifdef LOCALE_ENVIRON_REQUIRED
	262
	263	/*
	264	* Ultrix setlocale(..., "") fails if there are no environment
	265	* variables from which to get a locale name.
	266	*/
	267
	268	bool done = FALSE;
	269
	270	#ifdef LC_ALL
	271	if (lang) {
	272	if (setlocale(LC_ALL, ""))
	273	done = TRUE;
	274	else
	275	setlocale_failure = TRUE;
	276	}
	277	if (!setlocale_failure) {
	278	#ifdef USE_LOCALE_CTYPE
	279	if (! (curctype =
	280	setlocale(LC_CTYPE,
	281	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
0e2d6244	282	? "" : NULL)))
98994639 HS	283	setlocale_failure = TRUE;
	284	else
	285	curctype = savepv(curctype);
	286	#endif /* USE_LOCALE_CTYPE */
	287	#ifdef USE_LOCALE_COLLATE
	288	if (! (curcoll =
	289	setlocale(LC_COLLATE,
	290	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
0e2d6244	291	? "" : NULL)))
98994639 HS	292	setlocale_failure = TRUE;
	293	else
	294	curcoll = savepv(curcoll);
	295	#endif /* USE_LOCALE_COLLATE */
	296	#ifdef USE_LOCALE_NUMERIC
	297	if (! (curnum =
	298	setlocale(LC_NUMERIC,
	299	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
0e2d6244	300	? "" : NULL)))
98994639 HS	301	setlocale_failure = TRUE;
	302	else
	303	curnum = savepv(curnum);
	304	#endif /* USE_LOCALE_NUMERIC */
	305	}
	306
	307	#endif /* LC_ALL */
	308
	309	#endif /* !LOCALE_ENVIRON_REQUIRED */
	310
	311	#ifdef LC_ALL
	312	if (! setlocale(LC_ALL, ""))
	313	setlocale_failure = TRUE;
	314	#endif /* LC_ALL */
	315
	316	if (!setlocale_failure) {
	317	#ifdef USE_LOCALE_CTYPE
	318	if (! (curctype = setlocale(LC_CTYPE, "")))
	319	setlocale_failure = TRUE;
	320	else
	321	curctype = savepv(curctype);
	322	#endif /* USE_LOCALE_CTYPE */
	323	#ifdef USE_LOCALE_COLLATE
	324	if (! (curcoll = setlocale(LC_COLLATE, "")))
	325	setlocale_failure = TRUE;
	326	else
	327	curcoll = savepv(curcoll);
	328	#endif /* USE_LOCALE_COLLATE */
	329	#ifdef USE_LOCALE_NUMERIC
	330	if (! (curnum = setlocale(LC_NUMERIC, "")))
	331	setlocale_failure = TRUE;
	332	else
	333	curnum = savepv(curnum);
	334	#endif /* USE_LOCALE_NUMERIC */
	335	}
	336
	337	if (setlocale_failure) {
	338	char *p;
8c89da26	339	const bool locwarn = (printwarn > 1 \|\|
98994639 HS	340	(printwarn &&
	341	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
	342
	343	if (locwarn) {
	344	#ifdef LC_ALL
	345
	346	PerlIO_printf(Perl_error_log,
	347	"perl: warning: Setting locale failed.\n");
	348
	349	#else /* !LC_ALL */
	350
	351	PerlIO_printf(Perl_error_log,
	352	"perl: warning: Setting locale failed for the categories:\n\t");
	353	#ifdef USE_LOCALE_CTYPE
	354	if (! curctype)
	355	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
	356	#endif /* USE_LOCALE_CTYPE */
	357	#ifdef USE_LOCALE_COLLATE
	358	if (! curcoll)
	359	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
	360	#endif /* USE_LOCALE_COLLATE */
	361	#ifdef USE_LOCALE_NUMERIC
	362	if (! curnum)
	363	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
	364	#endif /* USE_LOCALE_NUMERIC */
	365	PerlIO_printf(Perl_error_log, "\n");
	366
	367	#endif /* LC_ALL */
	368
	369	PerlIO_printf(Perl_error_log,
	370	"perl: warning: Please check that your locale settings:\n");
	371
	372	#ifdef __GLIBC__
	373	PerlIO_printf(Perl_error_log,
	374	"\tLANGUAGE = %c%s%c,\n",
	375	language ? '"' : '(',
	376	language ? language : "unset",
	377	language ? '"' : ')');
	378	#endif
	379
	380	PerlIO_printf(Perl_error_log,
	381	"\tLC_ALL = %c%s%c,\n",
	382	lc_all ? '"' : '(',
	383	lc_all ? lc_all : "unset",
	384	lc_all ? '"' : ')');
	385
	386	#if defined(USE_ENVIRON_ARRAY)
	387	{
	388	char **e;
	389	for (e = environ; *e; e++) {
	390	if (strnEQ(*e, "LC_", 3)
	391	&& strnNE(*e, "LC_ALL=", 7)
	392	&& (p = strchr(*e, '=')))
	393	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
	394	(int)(p - e), e, p + 1);
	395	}
	396	}
	397	#else
	398	PerlIO_printf(Perl_error_log,
	399	"\t(possibly more locale environment variables)\n");
	400	#endif
	401
	402	PerlIO_printf(Perl_error_log,
	403	"\tLANG = %c%s%c\n",
404	lang ? '"' : '(',
405	lang ? lang : "unset",
406	lang ? '"' : ')');
407
408	PerlIO_printf(Perl_error_log,
409	" are supported and installed on your system.\n");
410	}
411
412	#ifdef LC_ALL
413
414	if (setlocale(LC_ALL, "C")) {
415	if (locwarn)
416	PerlIO_printf(Perl_error_log,
417	"perl: warning: Falling back to the standard locale (\"C\").\n");
418	ok = 0;
419	}
420	else {
421	if (locwarn)
422	PerlIO_printf(Perl_error_log,
423	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
424	ok = -1;
425	}
426
427	#else /* ! LC_ALL */
428
429	if (0
430	#ifdef USE_LOCALE_CTYPE
431	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
432	#endif /* USE_LOCALE_CTYPE */
433	#ifdef USE_LOCALE_COLLATE
434	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
435	#endif /* USE_LOCALE_COLLATE */
436	#ifdef USE_LOCALE_NUMERIC
437	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
438	#endif /* USE_LOCALE_NUMERIC */
439	)
440	{
441	if (locwarn)
442	PerlIO_printf(Perl_error_log,
443	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
444	ok = -1;
445	}
446
447	#endif /* ! LC_ALL */
448
449	#ifdef USE_LOCALE_CTYPE
0e2d6244	450	curctype = savepv(setlocale(LC_CTYPE, NULL));
98994639 HS	451	#endif /* USE_LOCALE_CTYPE */
98994639 HS	452	#ifdef USE_LOCALE_COLLATE
0e2d6244	453	curcoll = savepv(setlocale(LC_COLLATE, NULL));
98994639 HS	454	#endif /* USE_LOCALE_COLLATE */
98994639 HS	455	#ifdef USE_LOCALE_NUMERIC
0e2d6244	456	curnum = savepv(setlocale(LC_NUMERIC, NULL));
98994639 HS	457	#endif /* USE_LOCALE_NUMERIC */
	458	}
	459	else {
	460
	461	#ifdef USE_LOCALE_CTYPE
	462	new_ctype(curctype);
	463	#endif /* USE_LOCALE_CTYPE */
	464
	465	#ifdef USE_LOCALE_COLLATE
	466	new_collate(curcoll);
	467	#endif /* USE_LOCALE_COLLATE */
	468
	469	#ifdef USE_LOCALE_NUMERIC
	470	new_numeric(curnum);
	471	#endif /* USE_LOCALE_NUMERIC */
b310b053	472
98994639 HS	473	}
	474
	475	#endif /* USE_LOCALE */
	476
ec71e770	477	#ifdef USE_PERLIO
b310b053	478	{
6563884d	479	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	480	any of the following are true:
085a54d9	481	- nl_langinfo(CODESET) contains /^utf-?8/i
c240c76d	482	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	483	- $ENV{LC_CTYPE} contains /^utf-?8/i
c240c76d JH	484	- $ENV{LANG} contains /^utf-?8/i
	485	The LC_ALL, LC_CTYPE, LANG obey the usual override
	486	hierarchy of locale environment variables. (LANGUAGE
	487	affects only LC_MESSAGES only under glibc.) (If present,
	488	it overrides LC_MESSAGES for GNU gettext, and it also
	489	can have more than one locale, separated by spaces,
	490	in case you need to know.)
f8bb70a6 JH	491	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
f8bb70a6 JH	492	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
6563884d	493	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	494	*/
6563884d	495	bool utf8locale = FALSE;
b310b053 JH	496	char *codeset = NULL;
	497	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	498	codeset = nl_langinfo(CODESET);
	499	#endif
c240c76d	500	if (codeset)
6563884d JH	501	utf8locale = (ibcmp(codeset, "UTF-8", 5) == 0 \|\|
6563884d JH	502	ibcmp(codeset, "UTF8", 4) == 0);
077a72a9	503	#if defined(USE_LOCALE)
c240c76d JH	504	else { /* nl_langinfo(CODESET) is supposed to correctly
	505	* interpret the locale environment variables,
	506	* but just in case it fails, let's do this manually. */
	507	if (lang)
6563884d JH	508	utf8locale = (ibcmp(lang, "UTF-8", 5) == 0 \|\|
6563884d JH	509	ibcmp(lang, "UTF8", 4) == 0);
b310b053	510	#ifdef USE_LOCALE_CTYPE
c240c76d	511	if (curctype)
6563884d JH	512	utf8locale = (ibcmp(curctype, "UTF-8", 5) == 0 \|\|
6563884d JH	513	ibcmp(curctype, "UTF8", 4) == 0);
b310b053	514	#endif
c240c76d	515	if (lc_all)
6563884d JH	516	utf8locale = (ibcmp(lc_all, "UTF-8", 5) == 0 \|\|
6563884d JH	517	ibcmp(lc_all, "UTF8", 4) == 0);
c240c76d	518	}
6563884d JH	519	#endif /* USE_LOCALE */
	520	if (utf8locale)
	521	PL_utf8locale = TRUE;
	522	}
f8bb70a6	523	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
6563884d JH	524	This is an alternative to using the -C command line switch
	525	(the -C if present will override this). */
	526	{
f8b14789	527	char *p = PerlEnv_getenv("PERL_UNICODE");
f8bb70a6	528	PL_unicode = p ? parse_unicode_opts(&p) : 0;
2f583b3c NC	529	if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
2f583b3c NC	530	PL_utf8cache = -1;
b310b053	531	}
ec71e770	532	#endif
b310b053	533
98994639	534	#ifdef USE_LOCALE_CTYPE
4c58c75a	535	Safefree(curctype);
98994639 HS	536	#endif /* USE_LOCALE_CTYPE */
98994639 HS	537	#ifdef USE_LOCALE_COLLATE
4c58c75a	538	Safefree(curcoll);
98994639 HS	539	#endif /* USE_LOCALE_COLLATE */
98994639 HS	540	#ifdef USE_LOCALE_NUMERIC
4c58c75a	541	Safefree(curnum);
98994639 HS	542	#endif /* USE_LOCALE_NUMERIC */
	543	return ok;
	544	}
	545
98994639 HS	546	#ifdef USE_LOCALE_COLLATE
	547
	548	/*
	549	* mem_collxfrm() is a bit like strxfrm() but with two important
	550	* differences. First, it handles embedded NULs. Second, it allocates
	551	* a bit more memory than needed for the transformed data itself.
	552	* The real transformed data begins at offset sizeof(collationix).
	553	* Please see sv_collxfrm() to see how this is used.
	554	*/
40d34c0d	555
98994639 HS	556	char *
	557	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	558	{
	559	char *xbuf;
	560	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	561
	562	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	563	/* the +1 is for the terminating NUL. */
	564
	565	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
cd7a8267	566	Newx(xbuf, xAlloc, char);
98994639 HS	567	if (! xbuf)
	568	goto bad;
	569
	570	(U32)xbuf = PL_collation_ix;
	571	xout = sizeof(PL_collation_ix);
	572	for (xin = 0; xin < len; ) {
	573	SSize_t xused;
	574
	575	for (;;) {
	576	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	577	if (xused == -1)
	578	goto bad;
eb160463	579	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	580	break;
	581	xAlloc = (2 * xAlloc) + 1;
	582	Renew(xbuf, xAlloc, char);
	583	if (! xbuf)
	584	goto bad;
	585	}
	586
	587	xin += strlen(s + xin) + 1;
	588	xout += xused;
	589
	590	/* Embedded NULs are understood but silently skipped
	591	* because they make no sense in locale collation. */
	592	}
	593
	594	xbuf[xout] = '\0';
	595	*xlen = xout - sizeof(PL_collation_ix);
	596	return xbuf;
	597
	598	bad:
	599	Safefree(xbuf);
	600	*xlen = 0;
	601	return NULL;
	602	}
	603
	604	#endif /* USE_LOCALE_COLLATE */
	605
d8294a4d NC	606	/*
	607	* Local variables:
	608	* c-indentation-style: bsd
	609	* c-basic-offset: 4
	610	* indent-tabs-mode: t
	611	* End:
	612	*
	613	* ex: set ts=8 sts=4 sw=4 noet:
	614	*/