[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_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;

    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)
	      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)
    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: t
 * End:
 *
 * ex: set ts=8 sts=4 sw=4 noet:
 */
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 TC	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	* [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
	31	#ifdef I_LOCALE
	32	# include <locale.h>
	33	#endif
	34
b310b053 JH	35	#ifdef I_LANGINFO
	36	# include <langinfo.h>
	37	#endif
	38
a4af207c JH	39	#include "reentr.h"
a4af207c JH	40
27da23d5	41	#if defined(USE_LOCALE_NUMERIC) \|\| defined(USE_LOCALE_COLLATE)
98994639 HS	42	/*
	43	* Standardize the locale name from a string returned by 'setlocale'.
	44	*
	45	* The standard return value of setlocale() is either
	46	* (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
	47	* (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
	48	* (the space-separated values represent the various sublocales,
	49	* in some unspecificed order)
	50	*
	51	* In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
	52	* which is harmful for further use of the string in setlocale().
	53	*
	54	*/
	55	STATIC char *
	56	S_stdize_locale(pTHX_ char *locs)
	57	{
7452cf6a	58	const char * const s = strchr(locs, '=');
98994639 HS	59	bool okay = TRUE;
98994639 HS	60
7918f24d NC	61	PERL_ARGS_ASSERT_STDIZE_LOCALE;
7918f24d NC	62
8772537c AL	63	if (s) {
8772537c AL	64	const char * const t = strchr(s, '.');
98994639	65	okay = FALSE;
8772537c AL	66	if (t) {
	67	const char * const u = strchr(t, '\n');
	68	if (u && (u[1] == 0)) {
	69	const STRLEN len = u - s;
	70	Move(s + 1, locs, len, char);
	71	locs[len] = 0;
	72	okay = TRUE;
98994639 HS	73	}
	74	}
	75	}
	76
	77	if (!okay)
	78	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);
	79
	80	return locs;
	81	}
27da23d5	82	#endif
98994639 HS	83
	84	void
	85	Perl_set_numeric_radix(pTHX)
	86	{
	87	#ifdef USE_LOCALE_NUMERIC
97aff369	88	dVAR;
98994639	89	# ifdef HAS_LOCALECONV
7452cf6a	90	const struct lconv* const lc = localeconv();
98994639	91
98994639 HS	92	if (lc && lc->decimal_point) {
	93	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	94	SvREFCNT_dec(PL_numeric_radix_sv);
a0714e2c	95	PL_numeric_radix_sv = NULL;
98994639 HS	96	}
	97	else {
	98	if (PL_numeric_radix_sv)
	99	sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	100	else
	101	PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	102	}
	103	}
	104	else
a0714e2c	105	PL_numeric_radix_sv = NULL;
98994639 HS	106	# endif /* HAS_LOCALECONV */
	107	#endif /* USE_LOCALE_NUMERIC */
	108	}
	109
	110	/*
	111	* Set up for a new numeric locale.
	112	*/
	113	void
8772537c	114	Perl_new_numeric(pTHX_ const char *newnum)
98994639 HS	115	{
98994639 HS	116	#ifdef USE_LOCALE_NUMERIC
97aff369	117	dVAR;
98994639 HS	118
98994639 HS	119	if (! newnum) {
43c5f42d NC	120	Safefree(PL_numeric_name);
43c5f42d NC	121	PL_numeric_name = NULL;
98994639 HS	122	PL_numeric_standard = TRUE;
	123	PL_numeric_local = TRUE;
	124	return;
	125	}
	126
	127	if (! PL_numeric_name \|\| strNE(PL_numeric_name, newnum)) {
	128	Safefree(PL_numeric_name);
	129	PL_numeric_name = stdize_locale(savepv(newnum));
770526c1 NC	130	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
770526c1 NC	131	\|\| strEQ(newnum, "POSIX"));
98994639 HS	132	PL_numeric_local = TRUE;
	133	set_numeric_radix();
	134	}
	135
	136	#endif /* USE_LOCALE_NUMERIC */
	137	}
	138
	139	void
	140	Perl_set_numeric_standard(pTHX)
	141	{
	142	#ifdef USE_LOCALE_NUMERIC
97aff369	143	dVAR;
98994639 HS	144
	145	if (! PL_numeric_standard) {
	146	setlocale(LC_NUMERIC, "C");
	147	PL_numeric_standard = TRUE;
	148	PL_numeric_local = FALSE;
	149	set_numeric_radix();
	150	}
	151
	152	#endif /* USE_LOCALE_NUMERIC */
	153	}
	154
	155	void
	156	Perl_set_numeric_local(pTHX)
	157	{
	158	#ifdef USE_LOCALE_NUMERIC
97aff369	159	dVAR;
98994639 HS	160
	161	if (! PL_numeric_local) {
	162	setlocale(LC_NUMERIC, PL_numeric_name);
	163	PL_numeric_standard = FALSE;
	164	PL_numeric_local = TRUE;
	165	set_numeric_radix();
	166	}
	167
	168	#endif /* USE_LOCALE_NUMERIC */
	169	}
	170
	171	/*
	172	* Set up for a new ctype locale.
	173	*/
	174	void
8772537c	175	Perl_new_ctype(pTHX_ const char *newctype)
98994639 HS	176	{
98994639 HS	177	#ifdef USE_LOCALE_CTYPE
27da23d5	178	dVAR;
98994639 HS	179	int i;
98994639 HS	180
7918f24d NC	181	PERL_ARGS_ASSERT_NEW_CTYPE;
7918f24d NC	182
98994639 HS	183	for (i = 0; i < 256; i++) {
	184	if (isUPPER_LC(i))
	185	PL_fold_locale[i] = toLOWER_LC(i);
	186	else if (isLOWER_LC(i))
	187	PL_fold_locale[i] = toUPPER_LC(i);
	188	else
	189	PL_fold_locale[i] = i;
	190	}
	191
	192	#endif /* USE_LOCALE_CTYPE */
7918f24d	193	PERL_ARGS_ASSERT_NEW_CTYPE;
8772537c	194	PERL_UNUSED_ARG(newctype);
96a5add6	195	PERL_UNUSED_CONTEXT;
98994639 HS	196	}
	197
	198	/*
	199	* Set up for a new collation locale.
	200	*/
	201	void
8772537c	202	Perl_new_collate(pTHX_ const char *newcoll)
98994639 HS	203	{
98994639 HS	204	#ifdef USE_LOCALE_COLLATE
97aff369	205	dVAR;
98994639 HS	206
	207	if (! newcoll) {
	208	if (PL_collation_name) {
	209	++PL_collation_ix;
	210	Safefree(PL_collation_name);
	211	PL_collation_name = NULL;
	212	}
	213	PL_collation_standard = TRUE;
	214	PL_collxfrm_base = 0;
	215	PL_collxfrm_mult = 2;
	216	return;
	217	}
	218
	219	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
	220	++PL_collation_ix;
	221	Safefree(PL_collation_name);
	222	PL_collation_name = stdize_locale(savepv(newcoll));
770526c1 NC	223	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
770526c1 NC	224	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	225
	226	{
	227	/* 2: at most so many chars ('a', 'b'). */
	228	/* 50: surely no system expands a char more. */
	229	#define XFRMBUFSIZE (2 * 50)
	230	char xbuf[XFRMBUFSIZE];
8772537c AL	231	const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	232	const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	233	const SSize_t mult = fb - fa;
98994639 HS	234	if (mult < 1)
98994639 HS	235	Perl_croak(aTHX_ "strxfrm() gets absurd");
eb160463	236	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	237	PL_collxfrm_mult = mult;
	238	}
	239	}
	240
	241	#endif /* USE_LOCALE_COLLATE */
	242	}
	243
	244	/*
	245	* Initialize locale awareness.
	246	*/
	247	int
	248	Perl_init_i18nl10n(pTHX_ int printwarn)
	249	{
	250	int ok = 1;
	251	/* returns
	252	* 1 = set ok or not applicable,
	253	* 0 = fallback to C locale,
	254	* -1 = fallback to C locale failed
	255	*/
	256
	257	#if defined(USE_LOCALE)
97aff369	258	dVAR;
98994639 HS	259
	260	#ifdef USE_LOCALE_CTYPE
	261	char *curctype = NULL;
	262	#endif /* USE_LOCALE_CTYPE */
	263	#ifdef USE_LOCALE_COLLATE
	264	char *curcoll = NULL;
	265	#endif /* USE_LOCALE_COLLATE */
	266	#ifdef USE_LOCALE_NUMERIC
	267	char *curnum = NULL;
	268	#endif /* USE_LOCALE_NUMERIC */
	269	#ifdef __GLIBC__
7452cf6a	270	char * const language = PerlEnv_getenv("LANGUAGE");
98994639	271	#endif
7452cf6a AL	272	char * const lc_all = PerlEnv_getenv("LC_ALL");
7452cf6a AL	273	char * const lang = PerlEnv_getenv("LANG");
98994639 HS	274	bool setlocale_failure = FALSE;
	275
	276	#ifdef LOCALE_ENVIRON_REQUIRED
	277
	278	/*
	279	* Ultrix setlocale(..., "") fails if there are no environment
	280	* variables from which to get a locale name.
	281	*/
	282
	283	bool done = FALSE;
	284
	285	#ifdef LC_ALL
	286	if (lang) {
	287	if (setlocale(LC_ALL, ""))
	288	done = TRUE;
	289	else
	290	setlocale_failure = TRUE;
	291	}
	292	if (!setlocale_failure) {
	293	#ifdef USE_LOCALE_CTYPE
56279a21	294	Safefree(curctype);
98994639 HS	295	if (! (curctype =
	296	setlocale(LC_CTYPE,
	297	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
bd61b366	298	? "" : NULL)))
98994639 HS	299	setlocale_failure = TRUE;
	300	else
	301	curctype = savepv(curctype);
	302	#endif /* USE_LOCALE_CTYPE */
	303	#ifdef USE_LOCALE_COLLATE
56279a21	304	Safefree(curcoll);
98994639 HS	305	if (! (curcoll =
	306	setlocale(LC_COLLATE,
	307	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
bd61b366	308	? "" : NULL)))
98994639 HS	309	setlocale_failure = TRUE;
	310	else
	311	curcoll = savepv(curcoll);
	312	#endif /* USE_LOCALE_COLLATE */
	313	#ifdef USE_LOCALE_NUMERIC
56279a21	314	Safefree(curnum);
98994639 HS	315	if (! (curnum =
	316	setlocale(LC_NUMERIC,
	317	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
bd61b366	318	? "" : NULL)))
98994639 HS	319	setlocale_failure = TRUE;
	320	else
	321	curnum = savepv(curnum);
	322	#endif /* USE_LOCALE_NUMERIC */
	323	}
	324
	325	#endif /* LC_ALL */
	326
	327	#endif /* !LOCALE_ENVIRON_REQUIRED */
	328
	329	#ifdef LC_ALL
	330	if (! setlocale(LC_ALL, ""))
	331	setlocale_failure = TRUE;
	332	#endif /* LC_ALL */
	333
	334	if (!setlocale_failure) {
	335	#ifdef USE_LOCALE_CTYPE
6cb43dbf	336	Safefree(curctype);
98994639 HS	337	if (! (curctype = setlocale(LC_CTYPE, "")))
	338	setlocale_failure = TRUE;
	339	else
	340	curctype = savepv(curctype);
	341	#endif /* USE_LOCALE_CTYPE */
	342	#ifdef USE_LOCALE_COLLATE
6cb43dbf	343	Safefree(curcoll);
98994639 HS	344	if (! (curcoll = setlocale(LC_COLLATE, "")))
	345	setlocale_failure = TRUE;
	346	else
	347	curcoll = savepv(curcoll);
	348	#endif /* USE_LOCALE_COLLATE */
	349	#ifdef USE_LOCALE_NUMERIC
6cb43dbf	350	Safefree(curnum);
98994639 HS	351	if (! (curnum = setlocale(LC_NUMERIC, "")))
	352	setlocale_failure = TRUE;
	353	else
	354	curnum = savepv(curnum);
	355	#endif /* USE_LOCALE_NUMERIC */
	356	}
	357
	358	if (setlocale_failure) {
	359	char *p;
0bd48802	360	const bool locwarn = (printwarn > 1 \|\|
98994639 HS	361	(printwarn &&
	362	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
	363
	364	if (locwarn) {
	365	#ifdef LC_ALL
	366
	367	PerlIO_printf(Perl_error_log,
	368	"perl: warning: Setting locale failed.\n");
	369
	370	#else /* !LC_ALL */
	371
	372	PerlIO_printf(Perl_error_log,
	373	"perl: warning: Setting locale failed for the categories:\n\t");
	374	#ifdef USE_LOCALE_CTYPE
	375	if (! curctype)
	376	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
	377	#endif /* USE_LOCALE_CTYPE */
	378	#ifdef USE_LOCALE_COLLATE
	379	if (! curcoll)
	380	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
	381	#endif /* USE_LOCALE_COLLATE */
	382	#ifdef USE_LOCALE_NUMERIC
	383	if (! curnum)
	384	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
	385	#endif /* USE_LOCALE_NUMERIC */
	386	PerlIO_printf(Perl_error_log, "\n");
	387
	388	#endif /* LC_ALL */
	389
	390	PerlIO_printf(Perl_error_log,
	391	"perl: warning: Please check that your locale settings:\n");
	392
	393	#ifdef __GLIBC__
	394	PerlIO_printf(Perl_error_log,
	395	"\tLANGUAGE = %c%s%c,\n",
	396	language ? '"' : '(',
	397	language ? language : "unset",
	398	language ? '"' : ')');
	399	#endif
	400
	401	PerlIO_printf(Perl_error_log,
	402	"\tLC_ALL = %c%s%c,\n",
	403	lc_all ? '"' : '(',
	404	lc_all ? lc_all : "unset",
	405	lc_all ? '"' : ')');
	406
	407	#if defined(USE_ENVIRON_ARRAY)
	408	{
	409	char **e;
	410	for (e = environ; *e; e++) {
	411	if (strnEQ(*e, "LC_", 3)
	412	&& strnNE(*e, "LC_ALL=", 7)
	413	&& (p = strchr(*e, '=')))
	414	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
	415	(int)(p - e), e, p + 1);
	416	}
	417	}
	418	#else
	419	PerlIO_printf(Perl_error_log,
	420	"\t(possibly more locale environment variables)\n");
	421	#endif
	422
	423	PerlIO_printf(Perl_error_log,
	424	"\tLANG = %c%s%c\n",
425	lang ? '"' : '(',
426	lang ? lang : "unset",
427	lang ? '"' : ')');
428
429	PerlIO_printf(Perl_error_log,
430	" are supported and installed on your system.\n");
431	}
432
433	#ifdef LC_ALL
434
435	if (setlocale(LC_ALL, "C")) {
436	if (locwarn)
437	PerlIO_printf(Perl_error_log,
438	"perl: warning: Falling back to the standard locale (\"C\").\n");
439	ok = 0;
440	}
441	else {
442	if (locwarn)
443	PerlIO_printf(Perl_error_log,
444	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
445	ok = -1;
446	}
447
448	#else /* ! LC_ALL */
449
450	if (0
451	#ifdef USE_LOCALE_CTYPE
452	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
453	#endif /* USE_LOCALE_CTYPE */
454	#ifdef USE_LOCALE_COLLATE
455	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
456	#endif /* USE_LOCALE_COLLATE */
457	#ifdef USE_LOCALE_NUMERIC
458	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
459	#endif /* USE_LOCALE_NUMERIC */
460	)
461	{
462	if (locwarn)
463	PerlIO_printf(Perl_error_log,
464	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
465	ok = -1;
466	}
467
468	#endif /* ! LC_ALL */
469
470	#ifdef USE_LOCALE_CTYPE
56279a21	471	Safefree(curctype);
bd61b366	472	curctype = savepv(setlocale(LC_CTYPE, NULL));
98994639 HS	473	#endif /* USE_LOCALE_CTYPE */
98994639 HS	474	#ifdef USE_LOCALE_COLLATE
56279a21	475	Safefree(curcoll);
bd61b366	476	curcoll = savepv(setlocale(LC_COLLATE, NULL));
98994639 HS	477	#endif /* USE_LOCALE_COLLATE */
98994639 HS	478	#ifdef USE_LOCALE_NUMERIC
56279a21	479	Safefree(curnum);
bd61b366	480	curnum = savepv(setlocale(LC_NUMERIC, NULL));
98994639 HS	481	#endif /* USE_LOCALE_NUMERIC */
	482	}
	483	else {
	484
	485	#ifdef USE_LOCALE_CTYPE
	486	new_ctype(curctype);
	487	#endif /* USE_LOCALE_CTYPE */
	488
	489	#ifdef USE_LOCALE_COLLATE
	490	new_collate(curcoll);
	491	#endif /* USE_LOCALE_COLLATE */
	492
	493	#ifdef USE_LOCALE_NUMERIC
	494	new_numeric(curnum);
	495	#endif /* USE_LOCALE_NUMERIC */
b310b053	496
98994639 HS	497	}
	498
	499	#endif /* USE_LOCALE */
	500
ec71e770	501	#ifdef USE_PERLIO
b310b053	502	{
fde18df1	503	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	504	any of the following are true:
085a54d9	505	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	506	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	507	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	508	- $ENV{LANG} contains /^utf-?8/i
	509	The LC_ALL, LC_CTYPE, LANG obey the usual override
	510	hierarchy of locale environment variables. (LANGUAGE
	511	affects only LC_MESSAGES only under glibc.) (If present,
	512	it overrides LC_MESSAGES for GNU gettext, and it also
	513	can have more than one locale, separated by spaces,
	514	in case you need to know.)
8aa8f774	515	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	516	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	517	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	518	*/
fde18df1	519	bool utf8locale = FALSE;
b310b053 JH	520	char *codeset = NULL;
	521	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	522	codeset = nl_langinfo(CODESET);
	523	#endif
61de9fb5	524	if (codeset)
e6226b18 KW	525	utf8locale = (foldEQ(codeset, STR_WITH_LEN("UTF-8"))
e6226b18 KW	526	\|\| foldEQ(codeset, STR_WITH_LEN("UTF8") ));
077a72a9	527	#if defined(USE_LOCALE)
61de9fb5 JH	528	else { /* nl_langinfo(CODESET) is supposed to correctly
	529	* interpret the locale environment variables,
	530	* but just in case it fails, let's do this manually. */
	531	if (lang)
e6226b18 KW	532	utf8locale = (foldEQ(lang, STR_WITH_LEN("UTF-8"))
e6226b18 KW	533	\|\| foldEQ(lang, STR_WITH_LEN("UTF8") ));
b310b053	534	#ifdef USE_LOCALE_CTYPE
61de9fb5	535	if (curctype)
e6226b18 KW	536	utf8locale = (foldEQ(curctype, STR_WITH_LEN("UTF-8"))
e6226b18 KW	537	\|\| foldEQ(curctype, STR_WITH_LEN("UTF8") ));
b310b053	538	#endif
61de9fb5	539	if (lc_all)
e6226b18 KW	540	utf8locale = (foldEQ(lc_all, STR_WITH_LEN("UTF-8"))
e6226b18 KW	541	\|\| foldEQ(lc_all, STR_WITH_LEN("UTF8") ));
61de9fb5	542	}
fde18df1 JH	543	#endif /* USE_LOCALE */
	544	if (utf8locale)
	545	PL_utf8locale = TRUE;
	546	}
a05d7ebb	547	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	548	This is an alternative to using the -C command line switch
	549	(the -C if present will override this). */
	550	{
dd374669	551	const char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb	552	PL_unicode = p ? parse_unicode_opts(&p) : 0;
5a22a2bb NC	553	if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
5a22a2bb NC	554	PL_utf8cache = -1;
b310b053	555	}
ec71e770	556	#endif
b310b053	557
98994639	558	#ifdef USE_LOCALE_CTYPE
43c5f42d	559	Safefree(curctype);
98994639 HS	560	#endif /* USE_LOCALE_CTYPE */
98994639 HS	561	#ifdef USE_LOCALE_COLLATE
43c5f42d	562	Safefree(curcoll);
98994639 HS	563	#endif /* USE_LOCALE_COLLATE */
98994639 HS	564	#ifdef USE_LOCALE_NUMERIC
43c5f42d	565	Safefree(curnum);
98994639 HS	566	#endif /* USE_LOCALE_NUMERIC */
	567	return ok;
	568	}
	569
98994639 HS	570	#ifdef USE_LOCALE_COLLATE
	571
	572	/*
	573	* mem_collxfrm() is a bit like strxfrm() but with two important
	574	* differences. First, it handles embedded NULs. Second, it allocates
	575	* a bit more memory than needed for the transformed data itself.
	576	* The real transformed data begins at offset sizeof(collationix).
	577	* Please see sv_collxfrm() to see how this is used.
	578	*/
166f8a29	579
98994639 HS	580	char *
	581	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	582	{
97aff369	583	dVAR;
98994639 HS	584	char *xbuf;
	585	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	586
7918f24d NC	587	PERL_ARGS_ASSERT_MEM_COLLXFRM;
7918f24d NC	588
98994639 HS	589	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	590	/* the +1 is for the terminating NUL. */
	591
	592	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
a02a5408	593	Newx(xbuf, xAlloc, char);
98994639 HS	594	if (! xbuf)
	595	goto bad;
	596
	597	(U32)xbuf = PL_collation_ix;
	598	xout = sizeof(PL_collation_ix);
	599	for (xin = 0; xin < len; ) {
224e8ef5	600	Size_t xused;
98994639 HS	601
	602	for (;;) {
	603	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
224e8ef5	604	if (xused >= PERL_INT_MAX)
98994639	605	goto bad;
eb160463	606	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	607	break;
	608	xAlloc = (2 * xAlloc) + 1;
	609	Renew(xbuf, xAlloc, char);
	610	if (! xbuf)
	611	goto bad;
	612	}
	613
	614	xin += strlen(s + xin) + 1;
	615	xout += xused;
	616
	617	/* Embedded NULs are understood but silently skipped
	618	* because they make no sense in locale collation. */
	619	}
	620
	621	xbuf[xout] = '\0';
	622	*xlen = xout - sizeof(PL_collation_ix);
	623	return xbuf;
	624
	625	bad:
	626	Safefree(xbuf);
	627	*xlen = 0;
	628	return NULL;
	629	}
	630
	631	#endif /* USE_LOCALE_COLLATE */
	632
66610fdd RGS	633	/*
	634	* Local variables:
	635	* c-indentation-style: bsd
	636	* c-basic-offset: 4
	637	* indent-tabs-mode: t
	638	* End:
	639	*
37442d52 RGS	640	* ex: set ts=8 sts=4 sw=4 noet:
37442d52 RGS	641	*/