[perl5.git] / locale.c

/*    locale.c
 *
 *    Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
 *    2002, 2003, 2005, 2006, 2007, 2008 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!
 *
 *     [p.238 of _The Lord of the Rings_, II/i: "Many Meetings"]
 */

/* 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_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 unspecified 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;

    PERL_ARGS_ASSERT_STDIZE_LOCALE;

    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
    dVAR;
# 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_ const char *newnum)
{
#ifdef USE_LOCALE_NUMERIC
    dVAR;

    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
    dVAR;

    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
    dVAR;

    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_ const char *newctype)
{
#ifdef USE_LOCALE_CTYPE
    dVAR;
    int i;

    PERL_ARGS_ASSERT_NEW_CTYPE;

    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_ARGS_ASSERT_NEW_CTYPE;
    PERL_UNUSED_ARG(newctype);
    PERL_UNUSED_CONTEXT;
}

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

    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 && !(fa == 0 && fb == 0))
	      Perl_croak(aTHX_ "panic: strxfrm() gets absurd - a => %"UVuf", ab => %"UVuf,
			 (UV) fa, (UV) fb);
	  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)
    dVAR;

#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
	Safefree(curctype);
	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
	Safefree(curcoll);
	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
	Safefree(curnum);
	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
	Safefree(curctype);
	if (! (curctype = setlocale(LC_CTYPE, "")))
	    setlocale_failure = TRUE;
	else
	    curctype = savepv(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	Safefree(curcoll);
	if (! (curcoll = setlocale(LC_COLLATE, "")))
	    setlocale_failure = TRUE;
	else
	    curcoll = savepv(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	Safefree(curnum);
	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
	Safefree(curctype);
	curctype = savepv(setlocale(LC_CTYPE, NULL));
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	Safefree(curcoll);
	curcoll = savepv(setlocale(LC_COLLATE, NULL));
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	Safefree(curnum);
	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 = (foldEQ(codeset, STR_WITH_LEN("UTF-8"))
			    || foldEQ(codeset, STR_WITH_LEN("UTF8") ));
#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 = (foldEQ(lang, STR_WITH_LEN("UTF-8"))
				 || foldEQ(lang, STR_WITH_LEN("UTF8") ));
#ifdef USE_LOCALE_CTYPE
	      if (curctype)
		   utf8locale = (foldEQ(curctype, STR_WITH_LEN("UTF-8"))
				 || foldEQ(curctype, STR_WITH_LEN("UTF8") ));
#endif
	      if (lc_all)
		   utf8locale = (foldEQ(lc_all, STR_WITH_LEN("UTF-8"))
				 || foldEQ(lc_all, STR_WITH_LEN("UTF8") ));
	 }
#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). */
    {
	 const 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)
{
    dVAR;
    char *xbuf;
    STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */

    PERL_ARGS_ASSERT_MEM_COLLXFRM;

    /* 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; ) {
	Size_t xused;

	for (;;) {
	    xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	    if (xused >= PERL_INT_MAX)
		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: nil
 * End:
 *
 * ex: set ts=8 sts=4 sw=4 et:
 */
Commit	Line	Data
98994639 HS	1	/* locale.c
98994639 HS	2	*
1129b882 NC	3	* Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
1129b882 NC	4	* 2002, 2003, 2005, 2006, 2007, 2008 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	/*
4ac71550	12	* A Elbereth Gilthoniel,
cdad3b53	13	* silivren penna míriel
4ac71550	14	* o menel aglar elenath!
cdad3b53	15	* Na-chaered palan-díriel
4ac71550 TC	16	* o galadhremmin ennorath,
	17	* Fanuilos, le linnathon
	18	* nef aear, si nef aearon!
	19	*
	20	* [p.238 of _The Lord of the Rings_, II/i: "Many Meetings"]
98994639 HS	21	*/
98994639 HS	22
166f8a29 DM	23	/* utility functions for handling locale-specific stuff like what
	24	* character represents the decimal point.
	25	*/
	26
98994639 HS	27	#include "EXTERN.h"
	28	#define PERL_IN_LOCALE_C
	29	#include "perl.h"
	30
b310b053 JH	31	#ifdef I_LANGINFO
	32	# include <langinfo.h>
	33	#endif
	34
a4af207c JH	35	#include "reentr.h"
a4af207c JH	36
27da23d5	37	#if defined(USE_LOCALE_NUMERIC) \|\| defined(USE_LOCALE_COLLATE)
98994639 HS	38	/*
	39	* Standardize the locale name from a string returned by 'setlocale'.
	40	*
	41	* The standard return value of setlocale() is either
	42	* (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
	43	* (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
	44	* (the space-separated values represent the various sublocales,
486ec47a	45	* in some unspecified order)
98994639 HS	46	*
	47	* In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
	48	* which is harmful for further use of the string in setlocale().
	49	*
	50	*/
	51	STATIC char *
	52	S_stdize_locale(pTHX_ char *locs)
	53	{
7452cf6a	54	const char * const s = strchr(locs, '=');
98994639 HS	55	bool okay = TRUE;
98994639 HS	56
7918f24d NC	57	PERL_ARGS_ASSERT_STDIZE_LOCALE;
7918f24d NC	58
8772537c AL	59	if (s) {
8772537c AL	60	const char * const t = strchr(s, '.');
98994639	61	okay = FALSE;
8772537c 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	}
27da23d5	78	#endif
98994639 HS	79
	80	void
	81	Perl_set_numeric_radix(pTHX)
	82	{
	83	#ifdef USE_LOCALE_NUMERIC
97aff369	84	dVAR;
98994639	85	# ifdef HAS_LOCALECONV
7452cf6a	86	const struct lconv* const lc = localeconv();
98994639	87
98994639 HS	88	if (lc && lc->decimal_point) {
	89	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	90	SvREFCNT_dec(PL_numeric_radix_sv);
a0714e2c	91	PL_numeric_radix_sv = NULL;
98994639 HS	92	}
	93	else {
	94	if (PL_numeric_radix_sv)
	95	sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	96	else
	97	PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	98	}
	99	}
	100	else
a0714e2c	101	PL_numeric_radix_sv = NULL;
98994639 HS	102	# endif /* HAS_LOCALECONV */
	103	#endif /* USE_LOCALE_NUMERIC */
	104	}
	105
	106	/*
	107	* Set up for a new numeric locale.
	108	*/
	109	void
8772537c	110	Perl_new_numeric(pTHX_ const char *newnum)
98994639 HS	111	{
98994639 HS	112	#ifdef USE_LOCALE_NUMERIC
97aff369	113	dVAR;
98994639 HS	114
98994639 HS	115	if (! newnum) {
43c5f42d NC	116	Safefree(PL_numeric_name);
43c5f42d NC	117	PL_numeric_name = NULL;
98994639 HS	118	PL_numeric_standard = TRUE;
	119	PL_numeric_local = TRUE;
	120	return;
	121	}
	122
	123	if (! PL_numeric_name \|\| strNE(PL_numeric_name, newnum)) {
	124	Safefree(PL_numeric_name);
	125	PL_numeric_name = stdize_locale(savepv(newnum));
770526c1 NC	126	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
770526c1 NC	127	\|\| strEQ(newnum, "POSIX"));
98994639 HS	128	PL_numeric_local = TRUE;
	129	set_numeric_radix();
	130	}
	131
	132	#endif /* USE_LOCALE_NUMERIC */
	133	}
	134
	135	void
	136	Perl_set_numeric_standard(pTHX)
	137	{
	138	#ifdef USE_LOCALE_NUMERIC
97aff369	139	dVAR;
98994639 HS	140
	141	if (! PL_numeric_standard) {
	142	setlocale(LC_NUMERIC, "C");
	143	PL_numeric_standard = TRUE;
	144	PL_numeric_local = FALSE;
	145	set_numeric_radix();
	146	}
	147
	148	#endif /* USE_LOCALE_NUMERIC */
	149	}
	150
	151	void
	152	Perl_set_numeric_local(pTHX)
	153	{
	154	#ifdef USE_LOCALE_NUMERIC
97aff369	155	dVAR;
98994639 HS	156
	157	if (! PL_numeric_local) {
	158	setlocale(LC_NUMERIC, PL_numeric_name);
	159	PL_numeric_standard = FALSE;
	160	PL_numeric_local = TRUE;
	161	set_numeric_radix();
	162	}
	163
	164	#endif /* USE_LOCALE_NUMERIC */
	165	}
	166
	167	/*
	168	* Set up for a new ctype locale.
	169	*/
	170	void
8772537c	171	Perl_new_ctype(pTHX_ const char *newctype)
98994639 HS	172	{
98994639 HS	173	#ifdef USE_LOCALE_CTYPE
27da23d5	174	dVAR;
98994639 HS	175	int i;
98994639 HS	176
7918f24d NC	177	PERL_ARGS_ASSERT_NEW_CTYPE;
7918f24d NC	178
98994639 HS	179	for (i = 0; i < 256; i++) {
	180	if (isUPPER_LC(i))
	181	PL_fold_locale[i] = toLOWER_LC(i);
	182	else if (isLOWER_LC(i))
	183	PL_fold_locale[i] = toUPPER_LC(i);
	184	else
	185	PL_fold_locale[i] = i;
	186	}
	187
	188	#endif /* USE_LOCALE_CTYPE */
7918f24d	189	PERL_ARGS_ASSERT_NEW_CTYPE;
8772537c	190	PERL_UNUSED_ARG(newctype);
96a5add6	191	PERL_UNUSED_CONTEXT;
98994639 HS	192	}
	193
	194	/*
	195	* Set up for a new collation locale.
	196	*/
	197	void
8772537c	198	Perl_new_collate(pTHX_ const char *newcoll)
98994639 HS	199	{
98994639 HS	200	#ifdef USE_LOCALE_COLLATE
97aff369	201	dVAR;
98994639 HS	202
	203	if (! newcoll) {
	204	if (PL_collation_name) {
	205	++PL_collation_ix;
	206	Safefree(PL_collation_name);
	207	PL_collation_name = NULL;
	208	}
	209	PL_collation_standard = TRUE;
	210	PL_collxfrm_base = 0;
	211	PL_collxfrm_mult = 2;
	212	return;
	213	}
	214
	215	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
	216	++PL_collation_ix;
	217	Safefree(PL_collation_name);
	218	PL_collation_name = stdize_locale(savepv(newcoll));
770526c1 NC	219	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
770526c1 NC	220	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	221
	222	{
	223	/* 2: at most so many chars ('a', 'b'). */
	224	/* 50: surely no system expands a char more. */
	225	#define XFRMBUFSIZE (2 * 50)
	226	char xbuf[XFRMBUFSIZE];
8772537c AL	227	const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	228	const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	229	const SSize_t mult = fb - fa;
e0601d3c	230	if (mult < 1 && !(fa == 0 && fb == 0))
ad49ad39 NC	231	Perl_croak(aTHX_ "panic: strxfrm() gets absurd - a => %"UVuf", ab => %"UVuf,
ad49ad39 NC	232	(UV) fa, (UV) fb);
eb160463	233	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	234	PL_collxfrm_mult = mult;
	235	}
	236	}
	237
	238	#endif /* USE_LOCALE_COLLATE */
	239	}
	240
	241	/*
	242	* Initialize locale awareness.
	243	*/
	244	int
	245	Perl_init_i18nl10n(pTHX_ int printwarn)
	246	{
	247	int ok = 1;
	248	/* returns
	249	* 1 = set ok or not applicable,
	250	* 0 = fallback to C locale,
	251	* -1 = fallback to C locale failed
	252	*/
	253
	254	#if defined(USE_LOCALE)
97aff369	255	dVAR;
98994639 HS	256
	257	#ifdef USE_LOCALE_CTYPE
	258	char *curctype = NULL;
	259	#endif /* USE_LOCALE_CTYPE */
	260	#ifdef USE_LOCALE_COLLATE
	261	char *curcoll = NULL;
	262	#endif /* USE_LOCALE_COLLATE */
	263	#ifdef USE_LOCALE_NUMERIC
	264	char *curnum = NULL;
	265	#endif /* USE_LOCALE_NUMERIC */
	266	#ifdef __GLIBC__
7452cf6a	267	char * const language = PerlEnv_getenv("LANGUAGE");
98994639	268	#endif
7452cf6a AL	269	char * const lc_all = PerlEnv_getenv("LC_ALL");
7452cf6a AL	270	char * const lang = PerlEnv_getenv("LANG");
98994639 HS	271	bool setlocale_failure = FALSE;
	272
	273	#ifdef LOCALE_ENVIRON_REQUIRED
	274
	275	/*
	276	* Ultrix setlocale(..., "") fails if there are no environment
	277	* variables from which to get a locale name.
	278	*/
	279
	280	bool done = FALSE;
	281
	282	#ifdef LC_ALL
	283	if (lang) {
	284	if (setlocale(LC_ALL, ""))
	285	done = TRUE;
	286	else
	287	setlocale_failure = TRUE;
	288	}
	289	if (!setlocale_failure) {
	290	#ifdef USE_LOCALE_CTYPE
56279a21	291	Safefree(curctype);
98994639 HS	292	if (! (curctype =
	293	setlocale(LC_CTYPE,
	294	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
bd61b366	295	? "" : NULL)))
98994639 HS	296	setlocale_failure = TRUE;
	297	else
	298	curctype = savepv(curctype);
	299	#endif /* USE_LOCALE_CTYPE */
	300	#ifdef USE_LOCALE_COLLATE
56279a21	301	Safefree(curcoll);
98994639 HS	302	if (! (curcoll =
	303	setlocale(LC_COLLATE,
	304	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
bd61b366	305	? "" : NULL)))
98994639 HS	306	setlocale_failure = TRUE;
	307	else
	308	curcoll = savepv(curcoll);
	309	#endif /* USE_LOCALE_COLLATE */
	310	#ifdef USE_LOCALE_NUMERIC
56279a21	311	Safefree(curnum);
98994639 HS	312	if (! (curnum =
	313	setlocale(LC_NUMERIC,
	314	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
bd61b366	315	? "" : NULL)))
98994639 HS	316	setlocale_failure = TRUE;
	317	else
	318	curnum = savepv(curnum);
	319	#endif /* USE_LOCALE_NUMERIC */
	320	}
	321
	322	#endif /* LC_ALL */
	323
	324	#endif /* !LOCALE_ENVIRON_REQUIRED */
	325
	326	#ifdef LC_ALL
	327	if (! setlocale(LC_ALL, ""))
	328	setlocale_failure = TRUE;
	329	#endif /* LC_ALL */
	330
	331	if (!setlocale_failure) {
	332	#ifdef USE_LOCALE_CTYPE
6cb43dbf	333	Safefree(curctype);
98994639 HS	334	if (! (curctype = setlocale(LC_CTYPE, "")))
	335	setlocale_failure = TRUE;
	336	else
	337	curctype = savepv(curctype);
	338	#endif /* USE_LOCALE_CTYPE */
	339	#ifdef USE_LOCALE_COLLATE
6cb43dbf	340	Safefree(curcoll);
98994639 HS	341	if (! (curcoll = setlocale(LC_COLLATE, "")))
	342	setlocale_failure = TRUE;
	343	else
	344	curcoll = savepv(curcoll);
	345	#endif /* USE_LOCALE_COLLATE */
	346	#ifdef USE_LOCALE_NUMERIC
6cb43dbf	347	Safefree(curnum);
98994639 HS	348	if (! (curnum = setlocale(LC_NUMERIC, "")))
	349	setlocale_failure = TRUE;
	350	else
	351	curnum = savepv(curnum);
	352	#endif /* USE_LOCALE_NUMERIC */
	353	}
	354
	355	if (setlocale_failure) {
	356	char *p;
0bd48802	357	const bool locwarn = (printwarn > 1 \|\|
98994639 HS	358	(printwarn &&
	359	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
	360
	361	if (locwarn) {
	362	#ifdef LC_ALL
	363
	364	PerlIO_printf(Perl_error_log,
	365	"perl: warning: Setting locale failed.\n");
	366
	367	#else /* !LC_ALL */
	368
	369	PerlIO_printf(Perl_error_log,
	370	"perl: warning: Setting locale failed for the categories:\n\t");
	371	#ifdef USE_LOCALE_CTYPE
	372	if (! curctype)
	373	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
	374	#endif /* USE_LOCALE_CTYPE */
	375	#ifdef USE_LOCALE_COLLATE
	376	if (! curcoll)
	377	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
	378	#endif /* USE_LOCALE_COLLATE */
	379	#ifdef USE_LOCALE_NUMERIC
	380	if (! curnum)
	381	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
	382	#endif /* USE_LOCALE_NUMERIC */
	383	PerlIO_printf(Perl_error_log, "\n");
	384
	385	#endif /* LC_ALL */
	386
	387	PerlIO_printf(Perl_error_log,
	388	"perl: warning: Please check that your locale settings:\n");
	389
	390	#ifdef __GLIBC__
	391	PerlIO_printf(Perl_error_log,
	392	"\tLANGUAGE = %c%s%c,\n",
	393	language ? '"' : '(',
	394	language ? language : "unset",
	395	language ? '"' : ')');
	396	#endif
	397
	398	PerlIO_printf(Perl_error_log,
	399	"\tLC_ALL = %c%s%c,\n",
	400	lc_all ? '"' : '(',
	401	lc_all ? lc_all : "unset",
	402	lc_all ? '"' : ')');
	403
	404	#if defined(USE_ENVIRON_ARRAY)
	405	{
	406	char **e;
	407	for (e = environ; *e; e++) {
	408	if (strnEQ(*e, "LC_", 3)
	409	&& strnNE(*e, "LC_ALL=", 7)
	410	&& (p = strchr(*e, '=')))
	411	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
	412	(int)(p - e), e, p + 1);
	413	}
	414	}
	415	#else
	416	PerlIO_printf(Perl_error_log,
	417	"\t(possibly more locale environment variables)\n");
	418	#endif
	419
	420	PerlIO_printf(Perl_error_log,
	421	"\tLANG = %c%s%c\n",
422	lang ? '"' : '(',
423	lang ? lang : "unset",
424	lang ? '"' : ')');
425
426	PerlIO_printf(Perl_error_log,
427	" are supported and installed on your system.\n");
428	}
429
430	#ifdef LC_ALL
431
432	if (setlocale(LC_ALL, "C")) {
433	if (locwarn)
434	PerlIO_printf(Perl_error_log,
435	"perl: warning: Falling back to the standard locale (\"C\").\n");
436	ok = 0;
437	}
438	else {
439	if (locwarn)
440	PerlIO_printf(Perl_error_log,
441	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
442	ok = -1;
443	}
444
445	#else /* ! LC_ALL */
446
447	if (0
448	#ifdef USE_LOCALE_CTYPE
449	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
450	#endif /* USE_LOCALE_CTYPE */
451	#ifdef USE_LOCALE_COLLATE
452	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
453	#endif /* USE_LOCALE_COLLATE */
454	#ifdef USE_LOCALE_NUMERIC
455	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
456	#endif /* USE_LOCALE_NUMERIC */
457	)
458	{
459	if (locwarn)
460	PerlIO_printf(Perl_error_log,
461	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
462	ok = -1;
463	}
464
465	#endif /* ! LC_ALL */
466
467	#ifdef USE_LOCALE_CTYPE
56279a21	468	Safefree(curctype);
bd61b366	469	curctype = savepv(setlocale(LC_CTYPE, NULL));
98994639 HS	470	#endif /* USE_LOCALE_CTYPE */
98994639 HS	471	#ifdef USE_LOCALE_COLLATE
56279a21	472	Safefree(curcoll);
bd61b366	473	curcoll = savepv(setlocale(LC_COLLATE, NULL));
98994639 HS	474	#endif /* USE_LOCALE_COLLATE */
98994639 HS	475	#ifdef USE_LOCALE_NUMERIC
56279a21	476	Safefree(curnum);
bd61b366	477	curnum = savepv(setlocale(LC_NUMERIC, NULL));
98994639 HS	478	#endif /* USE_LOCALE_NUMERIC */
	479	}
	480	else {
	481
	482	#ifdef USE_LOCALE_CTYPE
	483	new_ctype(curctype);
	484	#endif /* USE_LOCALE_CTYPE */
	485
	486	#ifdef USE_LOCALE_COLLATE
	487	new_collate(curcoll);
	488	#endif /* USE_LOCALE_COLLATE */
	489
	490	#ifdef USE_LOCALE_NUMERIC
	491	new_numeric(curnum);
	492	#endif /* USE_LOCALE_NUMERIC */
b310b053	493
98994639 HS	494	}
	495
	496	#endif /* USE_LOCALE */
	497
ec71e770	498	#ifdef USE_PERLIO
b310b053	499	{
fde18df1	500	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	501	any of the following are true:
085a54d9	502	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	503	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	504	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	505	- $ENV{LANG} contains /^utf-?8/i
	506	The LC_ALL, LC_CTYPE, LANG obey the usual override
	507	hierarchy of locale environment variables. (LANGUAGE
	508	affects only LC_MESSAGES only under glibc.) (If present,
	509	it overrides LC_MESSAGES for GNU gettext, and it also
	510	can have more than one locale, separated by spaces,
	511	in case you need to know.)
8aa8f774	512	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	513	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	514	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	515	*/
fde18df1	516	bool utf8locale = FALSE;
b310b053 JH	517	char *codeset = NULL;
	518	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	519	codeset = nl_langinfo(CODESET);
	520	#endif
61de9fb5	521	if (codeset)
e6226b18 KW	522	utf8locale = (foldEQ(codeset, STR_WITH_LEN("UTF-8"))
e6226b18 KW	523	\|\| foldEQ(codeset, STR_WITH_LEN("UTF8") ));
077a72a9	524	#if defined(USE_LOCALE)
61de9fb5 JH	525	else { /* nl_langinfo(CODESET) is supposed to correctly
	526	* interpret the locale environment variables,
	527	* but just in case it fails, let's do this manually. */
	528	if (lang)
e6226b18 KW	529	utf8locale = (foldEQ(lang, STR_WITH_LEN("UTF-8"))
e6226b18 KW	530	\|\| foldEQ(lang, STR_WITH_LEN("UTF8") ));
b310b053	531	#ifdef USE_LOCALE_CTYPE
61de9fb5	532	if (curctype)
e6226b18 KW	533	utf8locale = (foldEQ(curctype, STR_WITH_LEN("UTF-8"))
e6226b18 KW	534	\|\| foldEQ(curctype, STR_WITH_LEN("UTF8") ));
b310b053	535	#endif
61de9fb5	536	if (lc_all)
e6226b18 KW	537	utf8locale = (foldEQ(lc_all, STR_WITH_LEN("UTF-8"))
e6226b18 KW	538	\|\| foldEQ(lc_all, STR_WITH_LEN("UTF8") ));
61de9fb5	539	}
fde18df1 JH	540	#endif /* USE_LOCALE */
	541	if (utf8locale)
	542	PL_utf8locale = TRUE;
	543	}
a05d7ebb	544	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	545	This is an alternative to using the -C command line switch
	546	(the -C if present will override this). */
	547	{
dd374669	548	const char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb	549	PL_unicode = p ? parse_unicode_opts(&p) : 0;
5a22a2bb NC	550	if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
5a22a2bb NC	551	PL_utf8cache = -1;
b310b053	552	}
ec71e770	553	#endif
b310b053	554
98994639	555	#ifdef USE_LOCALE_CTYPE
43c5f42d	556	Safefree(curctype);
98994639 HS	557	#endif /* USE_LOCALE_CTYPE */
98994639 HS	558	#ifdef USE_LOCALE_COLLATE
43c5f42d	559	Safefree(curcoll);
98994639 HS	560	#endif /* USE_LOCALE_COLLATE */
98994639 HS	561	#ifdef USE_LOCALE_NUMERIC
43c5f42d	562	Safefree(curnum);
98994639 HS	563	#endif /* USE_LOCALE_NUMERIC */
	564	return ok;
	565	}
	566
98994639 HS	567	#ifdef USE_LOCALE_COLLATE
	568
	569	/*
	570	* mem_collxfrm() is a bit like strxfrm() but with two important
	571	* differences. First, it handles embedded NULs. Second, it allocates
	572	* a bit more memory than needed for the transformed data itself.
	573	* The real transformed data begins at offset sizeof(collationix).
	574	* Please see sv_collxfrm() to see how this is used.
	575	*/
166f8a29	576
98994639 HS	577	char *
	578	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	579	{
97aff369	580	dVAR;
98994639 HS	581	char *xbuf;
	582	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	583
7918f24d NC	584	PERL_ARGS_ASSERT_MEM_COLLXFRM;
7918f24d NC	585
98994639 HS	586	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	587	/* the +1 is for the terminating NUL. */
	588
	589	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
a02a5408	590	Newx(xbuf, xAlloc, char);
98994639 HS	591	if (! xbuf)
	592	goto bad;
	593
	594	(U32)xbuf = PL_collation_ix;
	595	xout = sizeof(PL_collation_ix);
	596	for (xin = 0; xin < len; ) {
224e8ef5	597	Size_t xused;
98994639 HS	598
	599	for (;;) {
	600	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
224e8ef5	601	if (xused >= PERL_INT_MAX)
98994639	602	goto bad;
eb160463	603	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	604	break;
	605	xAlloc = (2 * xAlloc) + 1;
	606	Renew(xbuf, xAlloc, char);
	607	if (! xbuf)
	608	goto bad;
	609	}
	610
	611	xin += strlen(s + xin) + 1;
	612	xout += xused;
	613
	614	/* Embedded NULs are understood but silently skipped
	615	* because they make no sense in locale collation. */
	616	}
	617
	618	xbuf[xout] = '\0';
	619	*xlen = xout - sizeof(PL_collation_ix);
	620	return xbuf;
	621
	622	bad:
	623	Safefree(xbuf);
	624	*xlen = 0;
	625	return NULL;
	626	}
	627
	628	#endif /* USE_LOCALE_COLLATE */
	629
66610fdd RGS	630	/*
	631	* Local variables:
	632	* c-indentation-style: bsd
	633	* c-basic-offset: 4
14d04a33	634	* indent-tabs-mode: nil
66610fdd RGS	635	* End:
66610fdd RGS	636	*
14d04a33	637	* ex: set ts=8 sts=4 sw=4 et:
37442d52	638	*/